DEV Community: Rupesh Tiwari

Azure Availability Set Vs Availability Zone

Rupesh Tiwari — Sat, 29 Jan 2022 00:00:00 +0000

Did you know how can you protect your business against software or hardware disaster within Azure data-center? Did you also know how can you protect your services against data-center disasters? Azure guaranty to provide 99.99% SLA while using virtual machines within Availability Set and Zones. Read this article to understand the concept of Availability Set and Availability Zones. When you want to create VMs and manage them in Azure then you must know this concepts.

Availability Set

An availability set is a logical grouping of 2 or more than 2 VMs that help keep your application available during planned or unplanned maintenance. You can not put only 1 VM in Availability Set. You must need 2 or more VMs. There is no cost for the Availability Set itself, you only pay for each VM instance that you create. Running a VM with one or more replicated copies on separate hardware within the same Availability Set , providing resiliency against machine failure.

Note: No Infrastructure redundancy required means: If this machine fails then no other machine will serve user request.

Creating 1st VM Lets create 1st VM and create new Availability Set.

Step 1: Create New Virtual Machine

Step2: Select your Availability & Update domains.

Create 2nd VM Create 2nd VM in the same Resource Group & same Availability Set.

Both virtual machines under same Availability Set. The first VM is in Fault Domain-0 & Update Domain-0. The 2nd VM is in Fault Domain-1 & Update Domain-1

How VMs are spreadNotice our 2 VMs are in 2 different Fault domains and 2 different Update domains. Which gives Hardware fault tolerance.

So when Planned update is going on Update Domain 0 all the request to application will be navigated to the Update domain 1 this way you get 99.95% SLA.

How many update domains are assigned to an availability set?

Each availability set can be configured with up to three fault domains and twenty update domains. Update domains indicate groups of virtual machines and underlying physical hardware that can be rebooted at the same time.

Key Points to remember

You need to create virtual machines in the same resource group as the availability set.
One virtual machine can only be in one availability set.
You can assign virtual machines to the availability set only during the creation of virtual machines.
You should create Azure storage account per VM to get the diagnostic logs and Metrics per VM.

What is Availability Zones?

Availability Zones also has multiple Fault Domains and Update Domains. So in case Availability Zones, each Availability Zone is consider as a single Fault Domain and Update Domain.

Availability Zone is a unique physical location within an Azure Region. Each Availability Zone has multiple data-centers with independent Power supply, Network switch and Cooling system. All Region has Availability Zone.

Each Region has minimum 3 Availability Zones. At the Region level you can think of Availability Zone as 3 Fault Domains and 3 Update Domains. Suppose if you keep 3 VMs in 3 different Availability Zones. Then if 1st Availability Zone is down then your 1st VM will be not working. However, your business will not be affected since other 2 Availability Zones are still up and running. Azure will route your client request to next available Availability Zone.

Creating VMs in Availability Zone Steps

Location plays major role for Availability Zone

Availability Zones are available on selected region only.

Like for South India you don’t get any Availability Zone.

But for East USA you get Availability Zone.

While creating 1st Virtual Machine You select Availability zone as 1. When you create 2nd VM then give 2 for availability zone.

Always click on create new IP in Networking tab.

Make your IP address Zone-redundant

For creating 2nd VM you must select same resource group, same region and give 2 as Availability Zone.

Availability Zone gives 99.99% SLA.

You get more than > 99.99% of SLA in below services:

Azure Container
Service Fabrics
Azure Web Apps
Azure Batch

Azure Availability SLA

{: .notice–danger} \

🔔 IMPORTANT : The Availability set requires at least 2 running virtual machine instances to get the actual SLA.

If I create single VM you get 99.9% up time with premium storage.

If you create 2 machine with different Availability Set means they are in different Fault Domains then you get 99.95%. This provides resiliency against hardware failure within Data Center.

If you create 2 machine inside 2 different Availability Zones that is they are across different data centers inside different zones then you will get 99.99% up time. This protects from Data Center failure.

If you want to protect your Virtual Machines for disaster recovery across the region then place them in across different regions.

Single VM	Availability Sets	Availability Zones	Region Pairs
99.9% with Premium Storage only	99.95% up time( 4:50 down time)	99.99% up time (0:53 down time)	-
Single VM easier for lift and shift scenario	Protecting against hardware failures, network outages or power interruptions	Protecting against Data Center Failure	Protecting against region disaster
Single VM	2 or More VMs	2 or more VMs	2 or more VMs

Thanks for reading my article till end. I hope you learned something special today. If you enjoyed this article then please share to your friends and if you have suggestions or thoughts to share with me then please write in the comment box.

Become full stack developer 💻

I teach at Fullstack Master. If you want to become Software Developer and grow your carrier as new Software Engineer or Lead Developer/Architect. Consider subscribing to our full stack development training programs. You will learn Angular, RxJS, JavaScript, System Architecture and much more with lots of hands on coding. We have All-Access Monthly membership plans and you will get unlimited access to all of our video courses, slides , download source code & Monthly video calls.

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

Your bright future is awaiting for you so visit today FullstackMaster and allow me to help you to board on your dream software company as a new Software Developer, Architect or Lead Engineer role.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Azure Update Domain vs Fault Domain

Rupesh Tiwari — Sat, 22 Jan 2022 00:00:00 +0000

Do you know in Azure Data Center you can protect your server against regular server software maintenance outages? You can also protect your VMs on Azure against Network breakdown or power interruptions. Read this article to learn what is Update and Fault Domains in Data-Center and how they protect against software and hardware disasters.

In the Data Center the servers are divided into multiple physical and logical groups. The physical grouping is called Fault Domain and logical grouping is called Update Domain. Each Rack of servers with its own power supply and network switch as one fault domain. If there are 10 Racks in the Data Center means you have 10 different Fault Domains.

Update Domain

An update domain is a logical group of underlying hardware that can undergo maintenance or be rebooted at the same time. As you create VMs within an availability set, the Azure platform automatically distributes your VMs across these update domains.

Update domains comprises of groups of VMs and underlying physical hardware that can be rebooted at the same time.
Update domains are a logical part of each data center and are implemented with software and logic.
The Update Domains tells us how it is going to be patched when Azure does update to those machines.

With Respect to IaaS Services Your VMs on Azure are called as Guest VMs. With IaaS Microsoft will not going to patch Guest VMs. So if you spin up an IaaS VM with Windows on it and you do not turn on automatic updates or any of those upgrade features. Microsoft Azure is not going to patch your Virtual Machine OS at any point. But Azure will be patching the Hyper Visor as planned maintenance work. The Virtual Machine is running through some Hypervisor in the Azure. And that HyperVisor must be patched regularly.

So if you think your virtual machine will never be updated and never be switched off. And you control your VM you know when it will be down. And you think you don’t need multiple VMs. Then your are in-correct. You do need to have multiple VMs because Microsoft Azure is going to patch that Rack or that Hypervisor and it is going to bring your single VM down (or Azure Fabric will move it to other VM to give SLA) So you need to make sure that you have 2 VMs to get the defined SLA.

With Respect to PaaS ServicesSuppose you spin up an application in the PaaS services. So PaaS does patch the guest operating system (OS). When you bring the CS Package to do update. Then Azure will patch Update Domain 0 first then Update Domain 1 next Update Domain 2. So it will walk the stack to do the updates. That way client does not know and does not get the downtime.

Suppose you have a Web App running in Azure VM as PaaS, then Microsoft Azure will turn off the VM and when the next request comes from the client then Azure will send that request to other available VM to maintain the SLA.

You start with 5 update domains and you can take up to 20 Update Domains. Suppose you had 7 instances, So it will distribute your instances in 5 update domains then once it hit those first 5 update domains. It will layer the other 2 instances.

Fault Domain

It is kind of clustering like windows cluster. Its their to protect the physical assets. The Fault Domain primary use case is If the Rack went down, something bad happened in that Rack in Azure. Because, of the hardware broke, something broke in switch or something bad happened in the server of that Rack. This Rack can be repaired. However, Azure Fabric will detect that Rack went down and it will move the workloads to some other Rack as long as you have another virtual machine in another fault domain. On the top across the Racks Azure has VIPs or those virtual IPs across the top of our availability set so that it will continue to let the client write or read the servers. The servers are identical.

You can have maximum of 3 Fault Domains

It is a IaaS offering from the Azure
A fault domain is essentially a rack of servers. It provides the physical separation of your workload across different power, cooling, and network hardware that support the physical servers in the data center server racks.
The group of virtual machines that share common hardware are in the same fault domain.
Set of hardware component that share the single point of failure.
For fault tolerance, you need multiple Fault Domain
fault Domain distributes the loads across the physical Racks.

Learn More About Fault and Update Domains

Become full stack developer 💻

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Roadmap for mastering Coding & Competitive Programming interview

Rupesh Tiwari — Wed, 19 Jan 2022 00:00:00 +0000

Are you non-computer background student? I am a Electronics Engineering student and have been in software company and worked more than 16 years from junior developer to as full stack developer with angular and c# asp.net and various web development technologies. How should I plan to learn competitive programming in order to crack Big 4 (Google, Facebook, Amazon & Microsoft) companies coding-interviews?

Nowadays, even consulting companies like Infosys, TCS etc., are asking some coding problems in their interview. Therefore, if you are not preparing for FAANGM ( facebook, apple, amazon, netflix, google, microsoft) still you must learn how to code and practice coding. Solve easy problems from each Data Structure and sharpen your Logic skills.

If you are a beginner like me, never start directly solving competitive programming problems. First solve Data Structure and Algorithm problems. Read this to know what is the difference between Data Structure and Algorithm problems and Competitive Programming problems. So basically, you have start learning coding from basic.

Start solving problems in below sequence to practice:

Solve Data Structure and Algorithm problems especially from GeekOfGeeks site.
Solve Competitive Programming problems especially solve it from Leetcode. Here is the list of problems from Leetcode that you must try solving them.

I myself did very hard work and trying to solve problems given by Love Babbar 450 questions and from many other places.

You can view my solutions in JavaScript in this codepen link.
I also solved couple of problems related to basic data structure and algorithm here.

You can always fork my codepen and do the changes. I have added Jasmine Test in my code pen so that you can write function and run your specs.

Here is one example how you can solve problems all related to binary search and run specs to proove that they all work.

See the Pen Finding first or last occurrence of a number by Rupesh Tiwari (@rupeshtiwari) on CodePen.

Which Language should I use in a coding interview?

Many companies don’t prefer Python in coding interviews. Because, it has an in-built library for stack or so on. If you are doing Python then learn one more language in backup ( C#, C, C++ or Java ). Python is mostly used in Data Science and Machine Learning since these topics are very complex and you must write less code to solve these big problems. I personally use JavaScript language.

Where should I practice for Data Structure and Algorithm problems?

Use geeksforgeeks.com to practice data structure and algorithm problems.

Solve Problems for DataStructure & Algorithms
Read company interview experiences that will explain you how companies take interview what are their patterns.

Amazon, Meta (FaceBook), Google, Microsoft etc. they ask the questions from the Geek for Geeks sometime. So it is a good idea to solve them. Geeks For Geeks is the best site to start learning coding for any beginners.

Here is the steps to start coding practice.

Go to https://practice.geeksforgeeks.org/
Select Data Structure like start from Array.
Select the difficulty level easy ( if you are beginner I would say try School and Basic level first )
Daily do minimum 2 questions one from easy one from medium level difficulty
For each Data Structure target yourself to solve 50 questions
Do more than 50 questions for Array and String related problems especially

In Geek of Geeks you can filter difficulty level and you can also check which problem is asked in which company. Select Array then filter by Basic, Medium or Hard then work on them.

If you want to dedicate more time on coding practice and finish early then every day you do 2 easy, 2 medium and 1 hard problems for each data structure.

Which Data Structure Problem Should I Focus and solve?

Do 50 questions on below data structures. Do 70 questions on Array and Linked List. First solve the problems which are described in the data structure in hindi videos. Then solve others on your own. If you don’t watch videos and try to solve your own in Geek of Geeks you will solve 5 problems. However, if you watch videos and solve them then you should be able to solve 15 problems on your own. Do below Data Structures:

I have gone through Academic Booster channels videos from data structure in hindi videos and created my own GitHub repository. There I added his questions in JavaScript and put his solutions also. This will help you to get started. I have not finished all of his questions. If you are able to finish the playlist and want to contribute to this playlist then please create pull request. I will be happy to integrate.

Which Algorithm Problem Should I Do?

Most of the companies ask Dynamic Programming problems, they are very important.

Searching - do 50 problems
Sorting - do 50 problems
Greedy - do 70 problems
Dynamic Programming - do 70 problems

If you finish Dynamic programming then whatever Data Structure concepts that you have first finished. There you can optimize your solution using Dynamic Programming.

Where should I practice Competitive Programming Problems?

LeetCode.com is best for practicing for competitive programing.
Codechef.com/ is also good.

Participate in competitive problem competitions like:

Long Challenge: Within 10 days you have to solve problems
Cook Off
Lunch Time: Within 2-3 hours solve problems

Do not attempt solving Hard or Medium Problems when you start

When you are a beginner do not start directly with hard or medium problems otherwise you will be disheartened and you may lose your interest to learn coding. Therefore, I would advise not to waste time solving hard problems or mediums. Rather start with easy problems to build some confidence.

I keep forgetting previous solved problems when I do more problems. What to do?

When you will target yourself to solve 50 Array questions and daily you are solving lets say 5 problems. Then you will encounter an issue that you will start forgetting 1st problem solution, when you are attempting 6th one. This is a normal issue and it happens with every student. So do not dis-hearten. I will give you my solution to what I did.

There are some rules that I used to follow to keep myself updated with the problems that I was working on. Let me explain that it may or may not work for you but give it a try since it worked for me.

Before solving problem

Create one Excel sheet or Google Sheet called “Coding Problems”. Add columns like Problem Statement, Logic, Algorithm Name, Need more practice, category and difficulty level.

Coding Problem Cheat Sheet

Problem Statement	Logic	Algorithm Name	Need more practice?	Category	Difficulty Level
Write the problem clearly	Explain the logic in brief	Greedy	Yes	Array	Easy

During solving problem

Read the question 2-3 times until you kind or memorize or internalize the question. Sometimes, if you write down the question on a paper or white board and read it out loud, that will help you to understand and remember the question. So the point I am making is to understand and remember the question, not the solution.

While solving problem try to correlate your logic with the question, by reading the question again and again since the question has hints for the solution.

After solving a problem

Once you have solved the problem. Open coding problems sheet enter following:

Problem statement: Write down the problem, remember the very important read each word carefully!
Logic: Write solution logic in brief ( write as much as you need to remind yourself later time)
Algorithm Name: Write any algorithm names you have used like Greedy, Binary Search etc.
Need more practice: While reading this sheet you can mark this yes or no based on how much you can remember problem and its logic
Category: Write what Data Structure you are working or which Algorithm you are working
Difficulty Level: Which level you are working on Easy, Medium or Hard

Again before solving a problem

Next time when you start coding then first open this sheet, hide the logic column and read the problems and try to recollect the logic of the solution. If you are not able to recollect logic then unhide the logic column and read it. Next Mark the column needs more practice as YES. Next continue reading the next problems and keep updating the need for more practice columns. Once you finish this sheet then only start solving new problems.

Once you do this entry then every time you start coding first read this excel and try to re-collect the problem and its logic on how to solve it. Also if you think you can not remember then solve the problem again, change the logic entry in the sheet such that you can recollect when you read it again. Ideally you should re-collect your solution after reading the entry from this sheet till then keep changing the entry of this sheet to make it very much helpful for you to remind you of your solution. This is your diary for your work.

This is what I used to do to remember the logic of each problem that I solve. You may ask how can I remember 400 questions? So do not worry once you start working on this sheet your brain will be trained such that you will be able to solve many problems without even remembering. However, you have to start doing these notes and read them every time when you start coding practice.

Should I solve all Array problems before starting Linked List problems? What is my strategy for coding practice?

This is an interesting and very important question, comes in everyone mind. It depends where you are. If you are a beginner and you have a good amount of time if not working individually then 6 months and if you are working professional then you have 1 year time or so then you should do each problem from each Data Structure. I personally prefer working with all Data Structures at the same time rather finishing one Data Structure completely before starting another. Example when you start coding for the first time, do one easy problem from each Data Structure ( Array, Linked List, Stack, Queue, Tree, Graph). Next day you pick 2 easy, 2 medium and 2 hard from each Data Structure. Or you keep doing one easy question from each Data Structure. This way you should combine your Data Structure while practicing.

If you are an expert and you have already solved many problems and you want to just sharpen your skills with only one targeted Data Structure then pick that one and work on it. However, this one is only for the experts who had already done 50 problems from each Data Structure and they want to sharpen their skill on targeted one.

If you have any futher question please write in comment box.

Reference

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

How to start preparing for coding interview

Rupesh Tiwari — Mon, 17 Jan 2022 00:00:00 +0000

When I started learning coding for a coding-interview for Facebook, Amazon or Google. I faced problems when I saw programmers were using terminologies like back tracking, dp, 2 pointer concept, etc. I was really upset and could not continue my journey to prepare coding and did give up my learning process. Therefore, I think if you are also in the same boat then read this article to know how you can start coding and what you should do when you become a competitive programmer.

There is a subtle difference between software development engineer vs web development engineer. If you are already in the company 3-4 years of experience and coding in angular, react, node.js or any front-end technology then remember you are just a web development side.

Most of the Big 4 FAANGM (Facebook, Amazon, Apple, Netflix, Google, Microsoft) companies look for Software Development Engineers and there they will ask you first Data Structure and Algorithms related competitive questions. They may not ask you about angular reactions. So therefore, if you want to go for any product-based company like ( paypal, uber, linkedin, and FAANGM ) etc. You have to be an expert on competitive programming questions.

You need to practice 2 things below in order to crack your next coding interview that is for sure.

Data Structure & Algorithms (DS & ALGO)
Competitive Programming ( CP )

You must start learning coding from Data Structure and Algorithms and then solve competitive problems.

Do Not start form Competitive Programming (CP)

If you go directly to Competitive Programming (CP) you may be able to solve a couple of problems. You may find it difficult to solve further problems and will be disheartened and come back. Finally, you will drop the plan to learn coding and back out. So Do not start from CP.

Example: like in 5th grade students get word problems. Which finally boils down to do either sum, subtract or multiply. If you give the same word problems to 3rd grade students they can’t solve it. However, they can solve addition, subtraction and multiplication If you give them directly.

Lets learn what is the basic difference between Data Structure & Algorithms VS Competitive problems.

What is Data Structure & Algorithms Programming?

Data Structure (DS) is something where you learn the basics of programming and structuring your data in the correct fashion. Data structures are like ( Array, String, Hash Table, Linked List, Tree, Graph etc. ). Normally, in IIT Gate interview or in computer science semester questions are pure data structure and algorithm (DS & ALGO) questions like reverse link list, write quick sort algo, tree navigations, find loop in Linked List etc. So in Big 4 or product development companies they do not ask direct DS & ALGO questions they rather create some fancy questions which combines many DS, Algorithms and best-practices that will enforce you to think deep and identify the solution.

So therefore, DS & ALGO is foundation and first thing to start for your coding interview practice.

What is Competitive Programming?

Here you will get a paragraph of 10-15 lines of problems to read. You have to read that problem and think which DS and Algo you need to apply to solve the given problem.

Normally, In that problem statement, it asks you to solve the same DS Algo questions that you might have done earlier. However, you need to identify which one you need to apply to solve the given problem.

In Competitive Programming you may be applying below knowledges

DS & Algo Concepts
Mathematics Concepts
Bit Mask Concepts

This is the major difference between competitive programming and DS Algo. Therefore, you must always start your learning from Data Structure Programming then you should jump to Competitive Programming.

Competitive programming wants optimization from you.

My Experience with Facebook coding interview

In my Facebook coding interview, I have been asked to create an Iterator class which will have 2 methods one hasAny another method next.

Basically, if you read this question you just need to apply the find next in-order successor algorithm of Binary Tree data structure.

So data structure algorithm (DS ALGO) basic questions are wrapped in a fancy style and being asked to you in competitive programming. Therefore, you must be strong in DS & ALGO.

Summary

I used to ask my-self question why should I read a book on algorithms and data structure? So the answer is you must read them if you enjoy reading books. Or watch videos if you prefer video lectures to learn DS and ALGO. Therefore, the underline statement is you must learn and solve data structure & algorithm problems and then solve competitive programming problems to become a competitive programmer.

Reference

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Introduction of Azure Data Center

Rupesh Tiwari — Sat, 15 Jan 2022 00:00:00 +0000

Did you know how your servers are placed in Azure Data-Center? Do you understand how Azure Data Center protects your servers against software or hardware disasters? Read this article to understand Azure Data Center in-depth.

What is Azure Data Center?

Azure Data Center has Servers , Networks Switch, Power Supply & Hardware’s. Servers are placed on a Server Rack. Each Server Rack with it’s own Power Supply and Network Switch is called as Fault Domain. Each Server Rack has multiple Update Domains. Data-center normally comprises of lots of Fault Domains.

What is Planned Maintenance?

A planned maintenance event is when the underlying Azure fabric that hosts VMs is updated by Microsoft.

Security Patch, Performance patch, add and update features are called as planed maintenance.
After planned maintenance you must reboot your VMs.

What is Unplanned Maintenance?

Unplanned maintenance events involve a hardware failure in the data center, such as:

power outage
disk failure

Why do we separate servers in different Racks?

In order to provide single point of failures like you spread your money in different banks and mutual funds 😃.

Example lets say you have 100 servers in data center and they are all connected by single power supply and network switch.

What if the single power supply or network will fail it will bring down to all of the servers in the data-center.

Therefore, in a data-center we spread the servers in the different Racks with their own Power and Network Switch called as Fault Domain.

So if one Rack with 10 servers are affected then rest of the other Racks means remaining 90 servers are still up and running.

By grouping servers in multiple Racks or Fault domains we are eliminating single point of failure.

Why group servers in Update Domains?

Similar to Fault Domain your servers will be also grouped logically to an Update Domain. Very often planned maintenance requires server patches and software updates need to be applied on the Azure Hypervisor which is actually hosting your VMs. So your VM on that Azure Hypervisor can go down while patch is being applied. And we don’t all the servers to be rebooted at the same time. Therefore, all the servers within same update domain gets rebooted at one time and they take 30 minutes time to recover back. Meanwhile the machines which are in different update domain can serve the client requests. Therefore, update domain is a group of servers that can be updated and rebooted at the same time.

What is Availability Set in Data-Center?

Availability Set is a concept within a data-center. The servers in a data center are divided into multiple physical groups called Fault Domains and logical groups called Update Domains.

Availability Set in data-center is made up of multiple Fault Domains and Update Domains.

When you create an Availability Set you can decide how many Fault Domains and Update domains you want. A group of servers physically grouped are called Fault Domain and a logical group of server is called as Update Domain.

What is Availability Zones?

Availability Zones also has multiple Fault Domains and Update Domains. So in case Availability Zones, each Availability Zone is consider as a single Fault Domain and Update Domain.

Each Region has minimum 3 Availability Zones. At the Region level you can think of Availability Zone as 3 Fault Domains and 3 Update Domains.

Suppose if you keep 3 VMs in 3 different Availability Zones. Then if 1st Availability Zone is down then your 1st VM will be not working. However, your business will not be affected since other 2 Availability Zones are still up and running. Azure will route your client request to next available Availability Zone.

Azure Availability SLA

With Availability Zone Microsoft offers industries best SLA 99.99% up time.

Single VM	Availability Sets	Availability Zones	Region Pairs
99.9% with Premium Storage only	99.95% up time( 4:50 down time)	99.99% up time (0:53 down time)	-
Single VM easier for lift and shift scenario	Protecting against hardware failures, network outages or power interruptions	Protecting against Data Center Failure	Protecting against region disaster
Single VM	2 or More VMs	2 or more VMs	2 or more VMs

Reference

https://www.youtube.com/watch?v=On3cjV4VaKU

Become full stack developer 💻

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Cloud Solution Architect Soft Skills Questions

Rupesh Tiwari — Fri, 14 Jan 2022 00:00:00 +0000

Are you preparing for cloud solution architect interview then you must aware of soft skill questions. In this article I will give you questions that you must prepare for your next up coming interview.

This is part 2 of my cloud solution architect interview questions series. Read the part 1 where you will get the basic technical questions most frequently asked in cloud solution architect interview.

Tell me about yourself?

Firstly Thank you for the opportunity to be an interviewee in AWS sr. A cloud solution architect role is a privilege. Hi everyone, My name is Rupesh Tiwari, I have been working in IT for the last 15 years. I have started my career as a junior software engineer. I am currently working as Sr. Software Architect.

You can also describe your previous domain that you worked on and current company detail.

What qualities make you a good fit for the job?

Read your job description and prepare your answer accordingly.

I like helping customers and solving their business problems with the latest innovative cloud solutions & technical expertise. I have been developing, maintaining, and optimizing cloud platform solutions and services.

I am providing thought leadership both on technical design and business development. I have been owning the technical engagement of many initiatives and helped customers to arrive at their goal. By earning customer trust, building relationships and I have become a SME for Driver and cloud related developments.

In my past and current organizations, I have been part of the hiring and training department therefore, I have been sharing my knowledge, coaching our team on cloud architecture and related technologies. Being part of the core team, I have been helping various teams by learning their framework and solving their technical issues by designing and improving their architecture.

Personally, I have been writing content from the last 10 years on cloud and related technologies, posting cloud and related technology contents on my YouTube channel. In our local community I started a cloud learning community and started speaking on cloud fundamentals and socializing concepts.

I am highly energetic, enthusiastic and innovative about technologies and I believe I can look at your challenges and help you to solve them and make your team perform more effectively.

Why do you want to join us (your company name) ?

Sit down and close your eyes and think at least 3 points why you want to join this company.

Why should I hire you?

It is same as the question that I already asked you earlier.

What is the personality type of your hiring manager?

Your hirring manager could be of type:

Analytical

If your interviewer is asking for specifics, and some numbers and details.
Show them statistic, fact, details, show correlation on chart
Highly Logical : use detail, fact, figure

Intuitive : They want to get to the point very fast. Don’t give them detail. You talk about Big Picture of your architecture don’t go details.
Personal : they are good on communication, great listener. You SHARE
Functional : Process-Oriented people. They need process and procedure for everything. They plan everything perfect. They go in details You show them HOW things are done. Be Methodical.

4 PERSONALITY TYPES

ANALYTICAL — FACTS AND FIGURES
INTUlTlVE - BIG PICTURE
FUNCTIONAL - PROCESSES
PERSONAL - EMOTIONAL INTELLIGENCE (EQ)

What is Shut Down interview technique?

You control the interview is called shut down. Hiring manager wants to see my Your below skills:

Technical competency
Verbal communication skills

When he ask you technical question and u know it then answer in real depth. With so much depth so that there is no doubt that you are expert when they see and hear level of expertise and knowledge properly delivered there wont be any more technical questions.

What Major contribution u made to your last employer?

Go understand your current project and find which project you contributed a lot both technically as well as design and conversation with your client.

What Major Challenge at work and what you did to get pass it?

Go understand your current project and find which project was most challenging and list your challenges.

How to Sale yourself how to become a best sale representative?

Find your Hiring manager goal
Find the way to create a solution to solve the customer problem by giving far more benefit with same quality.
Show them your skills and your values are higher than the cost of their salary.

Hiring manager is looking for below skills:

Technical competency
Trusty
Emotional Intelligence
Communication Skills
Team player
Energetic, enthusiastic
Bring up best in other people.

What is building rapport?

By creating Personal and emotional connection between people. If u establish rapport to the hiring manager and make a personal connection then every answer will sound better and everything u don’t know It will be not that matter.

Know how to be solution to their problem. Know about the role each and every details. If you tell them how u can going to improve their business it is connecting rapport.
Research the hiring manager

What does your hiring manager want?

If asked direct question ask the direct answer don’t babble!

Any technology question is asked then answer below questions while describing answer:

What is this ?
How this work?
Why do you use it?

They are hiring you to solve their problem, they are not hirring you to solve your problem.

What Managers Want?

They want someone who’s technically competent
They want someone to help meet or exceed their goals
They want someone they can trust
They want someone who works well with people
They want someone who brings energy
They want someone with passion
They want someone who works well in a
They want someone who goes above and beyond

Tips for your Interview

During Interview

Be early for your interview
Companies don’t like procrastinators
Assume it’s an interview as soon as you walk in the door
Show depth of knowledge when you can
Highlight what you’re good at
Be honest about your weaknesses
Talk about what you’ve done

Be Honest

Be upfront with what you know and what you don’t
Show off your strengths but don’t lie about your weaknesses
This helps gain trust and respect from the interviewer
Point the conversation into the direction of your strengths

Control The Interview

Show expertise in what you know and go into detail when possible
This helps when over the employer
Impress your interviewer upfront

Know What You Don’t Know

Don’t put things on your resume that you don’t know
Be clear about your areas of expertise
Being a specialist isn’t a bad thing
Brush up on topics you’re not sure about

Attitude

Show that you’re motivated and passionate
Being enthusiastic is noticeable right away
Be the person that motivates the team as well
Energize others
Be Focused

Soft skills for solution architect

Read the body language of your hirring team or manager.

55% what we look like when u say something
38% sound like when u say something
7% content of your saying

Body Language

If manager is Leaning that means you should keep talking they are willing to listen you.
If manager is Leaning back then you must stop talking
If manager is Crossing arm then you must change what you are saying.
If manager Feet’s are pointing towards the door then they are ready to finish the interview you should don’t hold them and stop talking immediately.

Additional points to remember during interview:

Don’t say anything negative about your past employer/company
Not getting to the point!
Not reading body language
Listen More than you speak: Listen to the hiring manager if your Hiring manager is babbling, just keep hearing him.
Make your Strength and weakness ahead of time: Prepare your honest weakness and strength

How to Dressed for interview

Make sure you were the business suite, shoes and belts are shining and polished.

Do you have any questions for me?

When hirring manger is asking at the end of interview to ask question. Then your answer could be:

“What is your goal Mr. Hirring Manager (Your manger name)? Because if I was part of your team I would want to know that I could help you solve your goals.”

References

Go Cloud Videos

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Cloud Solution Architect Technical Interview Questions

Rupesh Tiwari — Thu, 13 Jan 2022 00:00:00 +0000

Are you willing to become cloud solution architect? I am going to add a few important technical questions that you must know for your next cloud solution architect interview. My answers would not be perfect so feel free to search for better answers. Important part from this article is what kind of technical questions you should know to become a good cloud solution architect. So focus on the questions that I am asking here and learn from anywhere about their answers.

Most of the time you will be asked various open-ended technical questions. The best strategy to answer them is think of below 4 things:

What is technology?
How does it work?
Why would you use it?
How do you architect them?

You must know what a software Architect does?

Business strategist
Advisor
designer

They have to speak the languages of CEO, CIO, CFO and Engineers.

What is Amazon EC2 (Elastic Compute Cloud) in AWS or Virtual Machine in Azure cloud?

It is virtualization of computers on cloud. Amazon EC2 is a virtualized machine where we can install applications and OS. It is an Infrastructure as a Service offering. We virtualize storage, network, CPU etc. using hypervisors.

What is caching and how does it work?

Caching is to offload servers by storing high frequently added information in the cache. If the user tries to take information from DB. It goes to cache and it will return. If the cache has no data it will go to the server return and cache it. Next person asks if it will return from the cache. It can improve performance, it can reduce latency. If you set up time to live of data in cache to timeout to keep it refreshing.

What is DNS? And what are the key record types?

Domain name System used to map IP addresses to a user friendly name. Phonebook of the internet. You can route traffic Based on geography, latency , weight etc. Key Record Types are as following:

A Record that match name to IPV4 address
AAAA Record that match name to IPV6 address
CNAME record where u map one domain to another domain
MX Record required for email.

What is the MiTM (Man in The Middle) attack?

“A Man-In-The-Middle attack is the type of attack where attackers intrude into an existing connection to intercept the exchanged data and inject false information. It involves eavesdropping on a connection and intercepting messages.” -Toolbox for IT.

Hackers can use Packet sniffing software (WIRESHARK) to intrude the connection. Get the sender and receiver server information. He can get access to the website as Alice or Bob.

Data encryption at rest and transit?

At rest encrypt the disk and hence you get data encryption at rest out of the box. At transit use VPN or SSL/TLS (HTTPS = 443) to secure your data by hashing it during transit.

What is Symmetric and Asymmetric?

Asymmetric: Sender data is encrypted with key1 and the receiver side uses key2 to decrypt the data.

Symmetric: both sender and receiver use the same key to encrypt and decrypt data.

What is NMap?

Used to scan a network, segment or host to find out the open ports. So that you can attack on those ports. Interviewer may ask if there is a linux machine and it has an application that is not working. What will you do. I will answer by saying I will first scan the network by using Nmap or wireshark to find out which ports are closed or opened etc.

What is IAM?

Users are defined in IAM and they get assigned with roles and they have access based on the role to the protected resources. Also it does logging of each activity done by the user. Authentication & Authorization also done by IAM. Who is the user and what he is allowed to do.

What is Social Engineering?

Social Engineer is someone who is a master of asking seemingly non-invasive

or unimportant questions to gather information over time

Gain trust
Reduce defenses

Social Engineer use deception to get sensitive info from someone also s/he can combine number of techniques to gather sensitive information.

How would you secure enterprise?

Follow Multi layer approach:

Policy
Look from User , Encryption, IAM and Tech perspective
Create Firewall protect your network perimeter
IDS and IPS system behind the firewall
DDoS ( Aws shield, cloud flare )
ACLS on the router and subnet
On the VM you install WAF to protect (XSS, XSRF)
Physical system secure ( like MAC address stop unused ports )
Phishing attack secure

How would you secure the cloud?

Follow Defence in Depth and secure physical to data layer.

Steps are as below:

Physical security
IAM (Identity and Access Management)
DDos, Firewall (Perimeter), IDS/IPS or intelligent thread detection
NSG (Network Security Group): Network protection
WAF (Web Application Firewall): Compute, Patches, Malware protection etc.
Application: design/code review of applications installed
Data: User azure PaaS services which is by default data at rest.

What is DDoS ?

Distributed denial of service (DDOS) is when a server is exhausted by serving large numbers malicious requests and could not serve genuine requests is called DDoS. It loses CIA principle ( Confidentiality , Integrity and Availability ). Now it is not available for genuine users. Azure DDoS protection , cloud flare , AWS shield can save you.

What is IDS or IPS?

IDS (IIntrusion Detection System)

Been around for quite awhile, fairly common and easier to set up. Logs alerts and events for later analysis
Allows for reactive response / research

IPS ( Intrusion Prevention System)

Newer platform over the last few years
Enables prevention (such as blocking IP addresses, etc.)
False positives could block legitimate traffic

What is DHCP?

Dynamic host configuration protocol (DHCP) This is a server that assigns automatically/dynamically IP to connected hosts or servers. Suppose you have 300K systems then you use DHCP. How does it work? Whenever a new system comes up. It sends a broadcast (“Hey I am new in the network I need an address” ) called DHCP discover. All the DHCP servers will respond ( called as DHCP offer ). If there are 10 DHCP servers the client will receive 10 IPs then the client has to send a DHCP request to his favorite or desired DHCP server. Then the server sends back a DHCP acknowledgement to the client.

What is a VLAN?

VLAN (virtual local area network) You have 100 computers in one switch and within one subnet. Limit the finance to access the things only they need. Finance, HR, Sales, ITOps team might need access to everything. You can virtualize your switch to chop into multiple virtual logical switches. This way you can enhance security for those logical groups of computers. Cloud Architect Technical Interview (Master The Cloud Architect Interview Questions!)

What is VLAN Trunking?

4 VLAN in single switch and another 4 in other switch u run a single cable between them to establish connection. That is called VLAN Trunking? How do you do it in a private and secure manner? The trunk has a tag for a VLAN. Each VLAN will get its own IEEE 802.1Q tag.

In a cloud computing environment. On-premise to Azure cloud when you do direct connection or express route. You send data over VLAN over a Trunk. Azure takes your information and your VLAN Tag and keeps it separate from other customers’ traffic.

What’s the difference between a hybrid cloud and a full cloud environment ? And what are the advantages of each?

Hybrid Cloud architecturally speaking you have your datacenter that has servers, compute, storage, networking and security and you also have the cloud. And you have both connected to get the benefits of your current infrastructure and connect the cloud to get the agility, scaling, and resilience.

Performance: You get lower latency from your datacenter compared to the cloud. Also you get control in case the cloud is down and you have your datacenter still running.

Native Cloud has No maintenance switch, routing, cooling, heating, physical security etc. If you are a new company and you don’t have money for capex then you should use cloud. So you save money and only spend on OPEX. You get more speed and scaling, reliance.

How can autoscaling help with DDoS protection?

Hackers control multiple computers and they send requests to your network and overwhelm your systems. If your system can handle 10K web requests and DDoS is giving 30K requests per system. Each of the systems had autoscaling enabled. Then automatically when the system will be over utilized (75% or more) cloud will add a new system and load balance them. So now you spawn up 10 diff systems and now you can handle 50K requests. So the DDoS attack was controlled by auto scaling.

What is Classful IP Addressing and problems?

Classful IP addressing host counts:

Class A - 16 million host identifiers
Class B - 65535 host identifiers
Class C - 254 host identifiers

Classful IP addressing IP ranges:

Class A 1.0.0.0-127.255.255.255 with mask 255.0.0.0 or (/8 in CIDR), 16 millions (2^24-2) IPs and 2^7 Networks.
Class B 128.0.0.0-191.255.255.255 with default subnet mask of 255.255.0.0 (/16 in CIDR) 65K (2^16-2) IPs and 2^14 Networks.
Class C 192.0.0.0-223.255.255.255 with default subnet mask of 255.255.255.0 (/24in CIDR) 254 (2^8-2) IPS and 2^21 Networks.
Class D and E are reserved for multicasts. 224.0.0.0 - 239.255.255.255

The problem would commonly occur when an organization required more than 254 host machines and therefore would no longer fall into class C but rather class B. This means that the organization would use a class B license even though they had far less than 65,535 hosts. Therefore if an organization only required 2,500 hosts, they would be wasting about 63,000 hosts by holding a class B license which would greatly decrease the availability of IPv4 addresses unnecessarily

From <https://www.keycdn.com/support/what-is-cidr>

What is CIDR?

Classless Inter-Domain Routing (CIDR) divides the IP address in network ID and host Id. CIDR is also known as supernetting. It is a method of allocating IP address and IP routing. It replaces the old school of classful IP addressing. CIDR is based on variable-length subnet masking (VLSM). This allows it to define prefixes of arbitrary lengths making it much more efficient than the old system. CIDR IP addresses are composed of two sets of numbers. The network address is written as a prefix, like you would see a normal IP address (e.g. 192.255.255.255). The second part is the suffix which indicates how many bits are in the entire address (e.g. /12). Putting it together, a CIDR IP address would look like the following: 192.255.255.255/12. The network prefix is also specified as part of the IP address. This varies depending upon the number of bits required. Therefore, taking the example above, we can say that the first 12 bits are the network part of the address while the last 20 bits are for host addresses.

What is SQS?

Simple Queueing System. Decouple A and B so you can scale both of them independently. It is used to decouple 2 independent components or microservices by giving reliable communication mechanisms.

What is autoscaling?

Buy the biggest server to meet your busiest day. On Christmas they get 10M requests. They can buy VMs and put them in an auto scaling group to scale automatically.

However in the cloud you get an autoscaling group for EC2 instances in AWS and you have Virtual Machine Scale Set in Microsoft Azure cloud.

There you just need to define:

Desired Capacity : 2
Minimum Capacity : 2
Maximum Size: 10 when 75% of both machine is reach

What is a Load Balancer? Why do you use it?

It is a server or a hardware that will determine which server to send traffic. To share the traffic. What if a single server failed because of load. You don’t have redundancy. For designing High Performance, High Available web applications, you need a load balancer. Load balancers increase the availability and performance of your system.

We can divide the load balancer in 2 parts: network and application load balancer.

Network Load Balancers are really fast because they work at layer -4. They are looking at TCP headers and they are pushing at your servers. So if you want high performance you need a network load balancer. Cross-Region or global load balancing.

Application Load Balancers work at layer-7 so it is a web-traffic load balancer so they enable us to manage traffic to your web application. SSL termination, cookie-based session, round robin for load-balancing traffic. If you want application intelligence and routing then use this.

If you want to route between microservices then use it.

What is API Gateway and its benefits?

https://dashbird.io/blog/can-api-gateway-act-load-balancer/

API Gateway offers below which ALB does not:

It integrates with IAM
Authentication & Authorization
API Tokens issuing and managing
Throttling
Caching

What is Firewall?

It blocks traffic at the edge of your network,
it keeps your network secure
It blocks all traffic denied by policies
It lets your outbound traffic back if it’s coming from you.

A firewall protects the perimeter of the edge of your network. It is a security appliance that’s basically standing guard saying you are not allowed in, you are allowed. It looks at traffic based on policy. By default it will block all the traffic from coming in. It will allow your internal traffic out on the internet and it will allow all your return traffic back through the firewall. Because Firewalls are Stateful therefore. Within perimeter if Mike wants to open google.com. Firewall is Stateful it will say okay Mike started the session, let his return traffic come back because, I know it’s destined for mike and I know it’s related to the same session where Mike requested google.com.

How does cloud computing affect an organization’s costs?

Saves capital cost. Costs like Workstation, firewall, load balancer, cooling, networking, power etc.

Operation cost also like : Electric bill, Network, staff

Move to the cloud, nothing to buy. You can do lift and shift and save capex. You only pay for what you use. Your Capex goes down and Opex goes up. Agility benefit, u can spawn up vms fastly.

What are VPC flow Logs ? NSG flow log In Azure?

In networking we have Cisco NetFlow. It will examine the traffic, they would cross cisco routers and you could say, hm the traffic all going from here to end point B and check for security violation. You can find areas of congestion in your network. Check any issue on network traffic and troubleshoot.

Vpc flow log gives info about traffic flow. They are excellent for any network troubleshooting especially in security.

It will tell you below things:

Source
Destination
Protocols
Port number

Something accepted by ACL but rejected by security group. If this traffic was supposed to be a web app hosted in a VM and you want to know why it got rejected by SG. Check the VPC flow logs.

What’s needed for a simple lift and shift of a system to the cloud?

Define the configuration of the server like How many cores, RAM and Storage.

What is the difference between Latency based Routing and Geolocation based routing?

Latency routing is for speed and performance, geo-location routing will also optimize the speed and performance however it will also send traffic to appropriate websites based on their locations.

Latency based routing determines the lowest latency and sends you to the web server with the lowest latency so you get the best experience

Latency based routing done by Traffic Manager/Front door

Geolocation routing looks at the source IP address and determines what is closest to you and sends them to a different destination. Example you can check IP and decide province and route to that website. For Finch province route to the French website. Route the English province to the English website. If you are dealing with a company having Multi Language websites there, use this.

What is the difference between a region, an availability zone, an edge location and local zone?

What is Region?

AWS has the concept of a Region, which is a physical location around the world where we cluster data centers. We call each group of logical data centers an Availability Zone. Each AWS Region consists of multiple, isolated, and physically separate AZs within a geographic area.

What is Availability Zone?

From <https://aws.amazon.com/about-aws/global-infrastructure/regions_az/> An Availability Zone (AZ) is one or more discrete data centers with redundant power, networking, and connectivity in an AWS Region.

What is Local Zones?

From <https://aws.amazon.com/about-aws/global-infrastructure/regions_az/> AWS Local Zones place compute, storage, database, and other select AWS services closer to end-users. This is for critical latency systems like Games.

What is Edge Location?

From <https://aws.amazon.com/about-aws/global-infrastructure/regions_az/> Edge location is Reliable, low latency and high throughput network connectivity. Amazon CloudFront peers with thousands of Tier 1/2/3 telecom carriers globally, is well connected with all major access networks for optimal performance, and has hundreds of terabits of deployed capacity. CloudFront Edge locations are connected to the AWS Regions through the AWS network backbone - fully redundant, multiple 100GbE parallel fiber that circles the globe and links with tens of thousands of networks for improved origin fetches and dynamic content acceleration.

To deliver content to end users with lower latency, Amazon CloudFront uses a global network of 225+ Points of Presence (215+ Edge locations and 13 regional mid-tier caches) in 90 cities across 47 countries. Amazon CloudFront Edge locations are located in: North America, Europe, Asia etc.

From <https://aws.amazon.com/cloudfront/features/?whats-new-cloudfront.sort-by=item.additionalFields.postDateTime&whats-new-cloudfront.sort-order=desc>

What is IAM?

AAA (authentication, authorization and accounting)

Who is allowed in the system, What they can do and Track what they do.

What is an 802.1Q tag?

VLAN is a subnet in the cloud. Take a physical switch and virtualize them here are examples:

Port 10-20 vlan -2
Port 20-30 vlan -3
Port 30-40 vlan-4

Take a single physical connection coming from on-premise to cloud vlan network using 802.1Q tag. With 802.1q tag aws separate your traffic from other customers for direct connection.

What is CDN?

Amazon CloudFront is a fast content delivery network (CDN) service that securely delivers data, videos, applications, and APIs to customers globally with low latency, high transfer speeds, all within a developer-friendly environment.

From <https://aws.amazon.com/cloudfront/>

Speed up content
Reduce network bandwidth to aws network
Closer to user
Guaranteed performance

Instead of going to Amazon.com go to dns and data center. It goes to CDN and you are calling from Miami and serving the site from that location. If cdn has not site cached then it can go fastly to the datacenter and fetch the content since cdn is on aws edge location which is connected to the aws backbone network. Reducing load on web server and Transfer cost on AWS or any cloud network for other cloud service providers.

What is the process to take a web server and make it publicly available on the internet if you are inside of the AWS cloud ?

You will take an external public IP from amazon pool and assign it to EC2 instance. From EC2 now you want the default router to be the internet router. Therefore, you need an internet gateway.

What is the optimal subnet size for a point-to-point WAN link?

Answer is /30 So I got 4 IP 2 reserved and got 2 free.

192.168.0.0/30
192.168.0.1/30
192.168.0.2/30
192.168.0.3/30

0.0.0.0 reserved for internet and 255.255.255.255 is for broadcast

What is the difference between IOPS and throughput?

IOPS (input output per seconds) determines how frequently you can read and write to the disc and how fast. IOPS is related to latency the higher the IOPS, meaning the more read/write operations per second, the lower the latency.

NVME (Non-Volatile Memory Express) is faster than SSD. Both NVME and SSD drives tend to have relatively low latency because the read/write operations are very fast or the IOPS are very fast. Magnetic drives have much higher latency and much lower IOPS. But if you put a bunch of Magnetic drives together in a RAID environment you can still get great THROUGHOUT. IOPS inversely proposal to Latency

The higher IOPS the lower latency and the latency on the network and vice versa. THROUGHPUT is the amount of stuff and data that can be moved at any one period of time.

Example: cars run 55 mph then its throughput would be whatever it could stuff in the trunk, or the boot, But if you had a Freight Train with 55 mph then you could carry a lot more stuff. And the stuff is the THROUGHPUT.

IOPS = LATENCY
THROUGHPUT = The amount of STUFF

Use case of IOPS and Throughput:

Video editors need a drive that needs high throughput because they are working with large video files but they can tolerate the latency.

A database needs extreme speed in terms of read and write operation but it is not moving large amounts of data then it needs higher IOPS.

What is the difference between ACL and NSG ?

Access Control Lists (ACL) is a Networking concept. It keeps traffic out of the subnet and Network Security Group (NSG) keeps traffic out of the Host. ACL protects the subnet and NSG protects the server. ACLs are stateless meaning it is not tracking the state of the connection. The traffic that is allowed in is not being monitored, so therefore, you don’t know a lot of return traffic. So rules in ACL have to be applied in both directions.

NSG is just like a firewall, it is stateful so you have to only enable the inbound direction. So if traffic is coming to the host then the return traffic is allowed from the same host back to the source.

ACL works at layer -3, NSG is stateful They are stateful, meaning that they allow return traffic to flow. In general allow everything in and out and block the specific one.

Stateful vs Stateless?

Stateful services keep track of sessions or transactions and react differently to the same inputs based on that history. Stateless services rely on clients to maintain sessions and center around operations that manipulate resources, rather than the state.

ACL is stateless
Firewall is stateful
NSG is stateful both in AWS and Azure

What is RAID 0, 1, 5 , 10?

Redundant Array of Independent Disks (RAID) is a virtual disk technology that combines multiple physical drives into one unit.

Raid-0 : Stripping, high speed, no FT, 100% capacity, data can not be recovered
Raid-1 : Mirroring , normal speed, FT, 50% capacity, data can be recovered when 1 disk crash, during 1 disk crash performance remains same, minimum 2 disk required. If you have 2 , 10TB drives you get only 1TB of storage.
Raid-5 : Stripping with Parity, High Throughput and normal IOPS, normal speed, FT, 1 disk capacity we lose, minimum 3 disk we need, 1 disk for parity bit, if any 1 disk failed data can be recovered, during disk failure performance become more down because it has to calculate the parity bit, capacity wise 1 disk capacity lost and used for parity saving
Raid-10 : Stripping and Mirroring together, logical drive made of raid 0 and raid 1, first we create 2 raid 0 disks (4) then we combine them to a logical drive called raid 1. finally it becomes raid 10. This gives high performance, FT, and capacity is 50% when 2 disks are lost.

Stripping means distributing bits in different disk and hence faster write because per disk you write very less bit and parallely you write so your write and read both are faster. Here you get speed however you don’t get Fault Tolerance one disk fails then entire data is lost.

Mirroring means duplicating means while writing you write all bits simultaneously in 2 different disks to keep redundant disk and protect from data loss. Mirroring does not slow down read write since its writing parallely in 2 disks together. However, since you are writing all bits in one disk at a time, it is slower than the Stripping technique. Here you do not get stripping technology disk speed however you get fault tolerance so if one disk failed then you still can recover your data since you had other disk available as redundant. You have to pay more for mirroring since you want to store 1 TB then you need 2TB for redundancy.

What is the AWS Shared Responsibility Model?

We take care of the cloud. You take care in the cloud.

AWS can manage below:

Physical network
Hardware
Manage server for SERVERLESS

https://aws.amazon.com/compliance/shared-responsibility-model/

How do you secure VPC?

Follow Defense in Depth model and secure from physical to data layer.

Physical

1- MAC are open for multiple unwanted ports then block them

Perimeter

Firewall
IPS
IDS
DDOS

IAM

Single sign On and Identity Providers

Network

Monitoring
Encryption
Session protection (XSS, XSR, Session Fixation, Session Side Jacking)
Network segmentation
Inside VPC to keep unwanted traffic out of the subnet use ACL

Compute

NSG to secure host
Anti-virus
Single Sign On
Managed Identity
Virtual Patching (this is for legacy apps lift and shift with the bad code: WAS and WAF whenever web application scan finds a threat it creates a automatic rule in the web application firewall )

Applications

SSL/TLS, HTTPS
Single Sign ON
Application integrity ( adhering to policies )
Vulnerability Scans

DATA

Server less -> enable the threat detection -> SEIM , enable data encryption at rest and over network .

Encryption at rest and transit
Masking

What are the options for using the cloud for disaster recovery?

#	Plan	Cost	Speed
1	Manual Backup and Restore, Point in time backups into the DR region.	Cheap($)	RPO = Hours, RTO = Days. Long time to come back to service.
2	Pilot Light (RPO 1min RTO 1hr). VMs are created on the DR site and turned off. However the database in the DR site is in active mode and you write your data in DR site live. WEB and Business VMs will have applications and configuration will be copied to DR sites however they are turned off.	$$	RPO=Minutes, RTO=Hours.
3	WARM standby(RPO 1sec, RTO 1Min). Replicate your environment but use very small instances in the DR site but place them in an auto scaling group. When DR workloads are scaled up fully then it is called HOT standby. The more scale up the RTO you get in the DR site.	$$$	RPO = Seconds, RTO = Minutes
4	Multi Region (Multi Site) Active-Active (RPO and RTO are zero) Your workload is deployed to, and actively serving traffic from, multiple AWS Regions. This strategy requires you to synchronize data across Regions. Use DNS to route the traffic	Most Costly	RPO=0, RTO=0

When VPN and Direct connection?

If you need guaranty consistent latency and guaranty bandwidth use Direct connection since it is connected with direct wire. Costly option. Use the VPN because it is cheaper, works with the internet, and is flexible since everyone has the internet. You can create connections on demand. It was easy to connect to multiple remote sites. The side is you depend on internet bandwidth.

You have site where u have 10 remote campuses. The main site is hosted in the cloud. Everyone wants to talk to everyone. How can you set up your topology?

#	Technique	Description	Speed	Maintenance
1	Full Mesh	Number of edges will be n x (n-1)/2	FAST	High
2	VPN	Connect on-premise with site2site VPN and connect each on-premise.	Normal	Medium
3	Hub & Spoke (Cloud Hub)	Hub and spoke topology on cloud	Normal	Low
4	Transit Gateway	Managed service from cloud provider to provide transit connection among virtual networks in cloud.	FAST	High

What function is achieved with IPSEC?

IPSec provides the ability to authenticate each remote end to prevent man-in-the-middle attacks. And it ensures the integrity of the data because it uses a hashing algorithm to encrypt the payload. IPSec provides Non-Repudiation means Assurance that the sender of information is provided with proof of delivery and the recipient is provided with proof of the sender’s identity, so neither can later deny having processed the information. Also it provides the tunnel for private IP addresses and private traffic and private routing information over public networks.

From <https://csrc.nist.gov/glossary/term/non_repudiation> It is a replacement of internet protocol. IPsec is a suite of protocols. IP protocol and security protocol. It encrypts the payload and keeps the header un-encrypted. It is used in VPN to create a tunnel of encryption from on-premise to cloud.

How can you scale relational databases?

Ask the question what type of traffic is being used by your database?

It’s a combination of read and write traffic.

If you have read traffic more then create read replicas and offload the read work to read replicas. Which will free up resources on the master database. You can further reduce the load on read replicas by adding some caching. In order to reduce write load, so you don’t have peaks and valleys you can use queuing system.

ACID vs BASE ?

Relational DB is ACID by nature:

Atomicity
Consistency
Isolation
Durability

NoSQL are BASE by nature:

Basic Availability: Application must handle partial failures itself
Soft State: DB state can change even without inputs
Eventually Consistency: • DB will “eventually” become consistent

What is virtualization, how does it work and why will you use it?

Take a bare metal server, add a hypervisor and create a virtual machine on the hypervisor.

Block Storage and Object Storage how they work, what are their advantages?

If you need storage on a VM that does not go away with reboot, you are going to use something like block storage. If you have a static website or you want to distribute software, you are going to use something like object storage. If you are going to create data-lake , use object storage. If you need high performance storage you will use block storage. Object storage is cheap and can store large volumes of data.

Object Storage

Object in the object storage is consist of :

Data
Unique Id: 128 bit
Expandable metadata: contextual data

Objects can’t be modified—you have to write the object completely at once. Object storage also doesn’t work well with traditional databases, because writing objects is a slow process and writing an app to use an object storage API isn’t as simple as using file storage.

From <https://www.redhat.com/en/topics/data-storage/file-block-object-storage>

Block Storage

Block storage chops data into blocks—get it?—and stores them as separate pieces. Each block of data is given a unique identifier, which allows a storage system to place the smaller pieces of data wherever is most convenient. That means that some data can be stored in a Linux® environment and some can be stored in a Windows unit. It is usually deployed in storage-area network (SAN) environments and must be tied to a functioning server. Block storage can be expensive. It has limited capability to handle metadata, which means it needs to be dealt with at the application or database level—adding another thing for a developer or systems administrator to worry about. From <https://www.redhat.com/en/topics/data-storage/file-block-object-storage>

What Is NAT and Why would you use it?

Network address translation and its really about translating one IP address to another address.

Now they could be used to connect private addresses to the internet by translating it into a public address.

From <https://docs.defenseorchestrator.com/Configuration_Guides/Security_Policy_Management/Network_Address_Translation/Common_Use_Cases_for_NAT>

What is Stateful Firewall or AWS NSG?

Stateful means watching the state of the connection, they keep the context of the route and remember which request is initiated from which host. If you dis-allowed outbound to port 80 then automatically inbound flights are prohibited.

When you initiate a connection it goes through the firewall on the way out to the internet. The firewall looks at your connection and it tracks what you have done. Because when a response comes it will allow you to back in because it knows you initiated the connections.

When connecting to AWS when would you use direct or VPN connection?

Direct connection is a wired connection from your on-premise route to Cloud provider router. You can use either use your ISP or direct Microsoft to connect direct connections. If you want guaranty bandwidth, performance and latency use direct connect.

VPN is You connect to the internet on both side and then you create a tunnel and you encrypt your data over the internet. So the internet Is not private and that’s why you have to encrypt your data. When you use a VPN it is cheaper because you are not buying a direct connection. You are connecting to the internet. VPN are really simple to setup because you can basically you can create any connection to any place that has internet connectivity, which is wonderful.

Advantages of VPN

Econmical
Flexible
Easy to setup

Disadvantages of VPN

Internet bandwidth and performance is not guarantied. Your ISP may guarantee you access to them but when you are on their network and off their network and on the internet there are no guarantees.
Internet is called best effort delivered if you try to reach webpage it does mean its going to get there. It should, it will try but it will not be guarantied.
So internet speed is high you get good performance if low then you get bad performance.
Latency is not guaranty in direct connect you get 3 ms to reach. In VPN message1 can take 2 ms. Message2 takes 100 ms and then message3 takes 5 ms. So message2 can arrive after message3 and that may be a horrible situation. Message2 takes

In reality, no physical tunnel exists, of course; the data has to travel through the same wires as any other data passing through the public network. Rather, VPN tunneling employs the concepts known as data encapsulation and encryption to safely carry data traffic through the non-secure environment. Encapsulation insulates the packet of data from other data traveling over the same network, while encryption makes the data “invisible” (unreadable) even to surveillance agents and criminals who recognize it as encrypted information. So it’s like the data is traveling inside a tunnel. Figure 1.0 below might help you visualize the process.

You can think of VPN tunneling as similar to the process of moving physical cash from one bank location to another using an armored transport van along public highways. The cash in this instance is your data, the public highway is the non-secure public network, and the armored van is analogous to a VPN tunnel (data encapsulation and encryption).

From <https://www.vpnmentor.com/blog/ultimate-guide-to-vpn-tunneling/>

What is BGP and why is it used?

Border Gateway Protocol (BGP) is layer 4 protocol. It is an exterior gateway protocol and it works with a path vector routing protocol that operates on TCP port 179. When you connect an entity to an external entity, that’s when you use an exterior gateway. That’s why organization use BGP to connect to AWS or GCP as opposed to something like OSPF (Open Shortest Path First, small routing within campus networks) or EIGRP (enhanced interior gateway routing protocol, used in networks of large commercial companies ) https://networkjourney.com/difference-between-eigrp-ospf-vs-bgp-routing-protocols/

BGP is highly Tunable and highly scalable. For example if an internet routing table has basically 3 quarters of a million routes and BGP can easily handle that, whereas an interior gateway routing protocol could not. That’s why organizations use BGP to connect to external entities like AWS.

BGP Routing preferences sequence:

Largest weight
Largest local preference
Shortest number of autonomous system hops
Lowest Med (metric)

In order to select the next hop BGP checks who has the maximum weight that route it will take, next it will prefer maximum local preference, next shortest number of system hops, next lowest metric. This way it optimizes the route selection and data transfer over the internet.

You are designing a video streaming software. What should you use TCP or UDP for? Which and why?

It must be UDP. TCP is used for reliable transport. UDP is used for Realtime transport. Why? If I send this “Rupesh like design” via UDP and I receive “Rupesh Design” I lost like If I send this “Rupesh like design” via TCP and I receive “Rupesh Design Like” I lost, then TCP will retransmit the Like and you get the data in different order. For video streaming you do not want this behavior. All streaming things must be UDP! Because it is better, faster for these applications. There are no sliding windows. Performance is going to be what it’s going to be and there is no re-transmission on lost data.

What is CXO and what does it mean to present differently to a CXO vs an Engineer?

CXO means a C-Level executive:

COO
CFO
CEO
CIO (chief information officer)

CXO means a C-level executive. They are extremely busy people. Because they are so busy they have an attention span of a few seconds at max. So you got to get your concept to the executive out to the point.

And you must talk about things that they care about.

The CEO is tasked with the organization’s strategy and increasing shareholder value meaning revenue growth or profitability growth.

The CFO cares about the organization’s finances. They are the gatekeeper for the organization’s money. When you are presenting to the CFO, you better be really good at basically doing some ROI (return on investment) modeling and showing that the value provided by your solution, provides greater value and savings or profitability to the company than it’s cost. You have to do that.

The CIO wants to know your technology solution is going to meet the CEO’s goals and needs. So you have to present this and it’s going to work. Now when you are dealing with the Engineers, they might need a lot of technical depth.

Presenting to the executive should be “SHORT to the POINT and the Solution”

Presenting to the engineer should be to dive deep and present how it works technically.

In most companies, the board of directors and the founders are at the top of the corporate hierarchy followed by the C-level executives namely the CEO, COO, CFO, etc. … Usually Vice Presidents (VPs) and Senior Vice Presidents (SVPs) report to C-level executives. https://resources.workable.com/hr-terms/c-level-executive

What does organization strategy mean?

An organizational strategy is a plan that specifies how your business will allocate resources (e.g., money, labor, and inventory) to support infrastructure, production, marketing, inventory, and other business activities.

https://getsling.com/blog/organizational-strategy

What is a Shareholder?

A shareholder is a party that legally owns shares of a company’s stock. Shareholders who own less than 50% of a company’s stock are known as ‘minority shareholders’, whereas shareholders who own 50% or more of a company’s stock are called ‘majority shareholders’. Shareholders are owners of the company, but they are not liable for the company’s debts.

From <https://www.investopedia.com/ask/answers/08/difference-between-a-shareholder-and-a-stakeholder.asp>

From <https://sumup.co.uk/invoices/dictionary/shareholders/>

What is a stakeholder?

Stakeholders can be:

Owners and shareholders
Employees of the company
Bondholders who own company-issued debt
Customers who may rely on the company to provide a particular good or service
Suppliers and vendors who may rely on the company to provide a consistent revenue stream

From <https://www.investopedia.com/ask/answers/08/difference-between-a-shareholder-and-a-stakeholder.asp> A shareholder can sell their stock and buy different stock; they do not have a long-term need for the company. Stakeholders, however, are bound to the company for a longer term and for reasons of greater need. For example, if a company is performing poorly financially, the vendors in that company’s supply chain might suffer if the company no longer uses their services. Similarly, employees of the company, who are stakeholders and rely on it for income, might lose their jobs.

What is AWS?

Amazon Web Services (AWS) is the world’s most comprehensive and broadly adopted cloud platform, offering over 200 fully featured services from data centers globally. Millions of customers—including the fastest-growing startups, largest enterprises, and leading government agencies—are using AWS to lower costs, become more agile, and innovate faster.

From <https://aws.amazon.com/what-is-aws/?nc1=f_cc>

What is cloud computing?

Cloud computing is the on-demand delivery of IT resources over the Internet with pay-as-you-go pricing. Instead of buying, owning, and maintaining physical data centers and servers, you can access technology services, such as computing power, storage, and databases, on an as-needed basis from a cloud provider like Amazon Web Services (AWS).

From <https://aws.amazon.com/what-is-cloud-computing/>

Cloud Computing Models?

From <https://docs.aws.amazon.com/whitepapers/latest/aws-overview/types-of-cloud-computing.html>

Infrastructure as a Service (IaaS)

IaaS contains the basic building blocks for cloud IT. It typically provides access to networking features, computers (virtual or on dedicated hardware), and data storage space.

From <https://aws.amazon.com/what-is-cloud-computing/>

Platform as a Service (PaaS)

PaaS removes the need for you to manage underlying infrastructure (usually hardware and operating systems), and allows you to focus on the deployment and management of your applications

From <https://aws.amazon.com/what-is-cloud-computing/>

Software as a Service (SaaS)

SaaS provides you with a complete product that is run and managed by the service provider.

From <https://aws.amazon.com/what-is-cloud-computing/>

Six Advantages of Cloud Computing

From <https://docs.aws.amazon.com/whitepapers/latest/aws-overview/six-advantages-of-cloud-computing.html>

Go global in minutes
Trade capital expense for variable expense
Benefit from massive economies of scale
Increase speed and agility
Stop spending money running and maintaining data centers
Stop guessing capacity

Cloud Computing Deployment Models?

Cloud
Hybrid
On Premise

What is Region?

An AWS Region is a physical location in the world where we have multiple Availability Zones. Each Amazon Region is designed to be completely isolated from the other AWS Regions. This achieves the greatest possible fault tolerance and stability. AWS has 25 AWS regions with 240 countries.

What is the Availability Zone?

Availability Zones consist of one or more discrete data centers, each with redundant power, networking, and connectivity, housed in separate facilities. These Availability Zones offer you the ability to operate production applications and databases that are more highly available, fault tolerant, and scalable than would be possible from a single data center. Each Availability Zone is designed as an independent failure zone. This means that Availability Zones are physically separated within a typical metropolitan region and are located in lower risk flood plains (specific flood zone categorization varies by AWS Region).

What is NAS?

Network attached storage. Centralize storage device for storing data on a network. I will have multiple hard drives in a RAID configuration. Used in home and small side business. It has a NIC so that you can connect this with a switch or router to be in the network and then you can access it from laptop, desktop mobile etc. It can be accessed as a shared drive. It has a single point of failure, if power supply fail in the NAS then other devices will not able to access it.

What is SAN?

Storage area network is special. SANs are a highly scalable, highly redundant and high speed network that stores and provides access to large amounts of data. SAN is fault tolerant since data is shared among several disk arrays. If a switch, disk array or server goes down data can still be accessed. Server accesses this data as if it was a local attached hard drive. In NAS it is recognized as a network drive. All other devices in SAN are connected to each other using fiber channel (fiber optics ) 2 -128 gbps speed. Fiber channel is very costly. The other option is iSCSI (Internet Small Computer System Interface) which is a cheaper alternative to using Fiber channel but they are not very fast. SANs are not affected by network traffic happening in LAN because SANs are not part of LAN. SANs are partitioned off, it’s basically a network all by itself. SANs are very expensive therefore, only large company, enterprises can afford this.

Some important slides from NAS vs SAN - Network Attached Storage vs Storage Area Network

What is Scalability?

Scalability is the ability of a system to increase workload by adding resources. Scale up, down, out or in. Scalability vs Elasticity in 99 seconds

What is Elasticity?

Elasticity is the ability to automatically match the resources sufficient for a given workload size.Scalability vs Elasticity in 99 seconds

Network load balancer vs Application load balancer

Application load balancer:

URL routing can be done in
In level -7 (application)
Low performance
Load balance frontend servers

Network load balancer:

In level -4 (transport)
High performance
Load balance backend servers

I will publish one more article where I will explain what are the soft skills you need to be successful in your next cloud architect interview so stay tuned.

References

Go Cloud Videos

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Server Deployment Strategies and Storage Spaces Direct Fault Tolerance in Azure

Rupesh Tiwari — Sat, 08 Jan 2022 00:00:00 +0000

Learn how Storage Spaces Direct provides you a Data fault tolerance at verity of levels. You don’t loose data at any stage. You get Data resiliency out of the box in Azure.to

You can deploy servers in below types:

Node
Chassis for Blade Servers
Rack Server

What is Planned Maintenance?

A planned maintenance event is when the underlying Azure fabric that hosts VMs is updated by Microsoft.

Security Patch, Performance patch, add and update features are called as planed maintenance.
After planned maintenance you must reboot your VMs.

What is Unplanned Maintenance?

Unplanned maintenance events involve a hardware failure in the data center, such as:

power outage
disk failure

What is Node?

Node is nothing but a server. It could be Server Virtual Machine. Or Database Virtual Machine.

What is Rack?

One Rack has servers, data disks and it’s own Network Switch and Power Supply.

Fault Tolerance in Azure

You can arrange your servers in all different types of deployment to manage fault tolerance at Drive, Node or Rack level.

Learn More about Failover Clustering Fault Domains

Node Fault Tolerance

In Windows Server 2016 clustering allows multiple servers called nodes to work together for High availability is called as Node Tolerance.

Chassis Fault Tolerance

Chassis is a hardware where you put servers. This is a Chassis, it has 4 servers they share common components like power supply and switch network. They are less expensive.

If 3rd number Chassis common components fails then all nodes within the Chassis will go down.

Windows Server 2016 Clustering has Chassis Fault Tolerance.

Rack Fault Tolerance

If network switch for one Rack is out of order then in this photo 2nd Rack is out of order then all of the servers (Nodes) are down within the Rack only. Rest Rack will be still working. This is called Rack Fault Tolerance.

Storage Spaces Direct Fault Tolerance Support

Spaces are the set of disks to store data. Spaces provides Data resiliency by copying the data in several redundant drives.

Data resiliency

Spaces provides Data resiliency by keeping multiple copy of every thing in on different drives.

So if one fails then all data is safe and accessible.

The remaining copy will be used automatically as the failed drive to restore it. This feature is already in Windows 2010.

Node Fault Tolerance

Spaces also stores every copy of data not only to different drive but also in different nodes.

So that if one node is out of order then other node will automatically start and working.

Rack Fault Tolerance Storage

Storage Space Direct can guaranty that it will place every copy not on every Drive, different Node but in different Rack. To provide data even at Rack level outage.

Become full stack developer 💻

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Introduction to Azure Infrastructure Architecture

Rupesh Tiwari — Sat, 01 Jan 2022 00:00:00 +0000

Understanding Azure architecture gives you power to learn lots of higher level concepts on Azure like Virtual Machine scale sets etc. Therefore, in this article you will learn what is Azure overall architecture and its individual components like region, zones, data-centers etc.

Azure Infrastructure Architecture

Geography consists of many regions
Each Region must have minimum 3 Availability Zones.
Each Availability Zone can have one or more Data Centers equipped with independent power, cooling, and networking.
No two zones can share a data center.
Each Data-Center has Availability Sets
Each Availability Set can have 3 Fault Domains & up to 20 Update Domain
Fault Domain is grouping of virtual machines, independent network switch and power supply provides hardware fault tolerance.
Update Domain is logical grouping of virtual machine and included hardware that can be updated and rebooted at one time.

Azure Geography

Azure Geography contains most of the countries in the world.

An Azure geography is a discrete market, typically containing at least one or more regions, that preserves data residency and compliance boundaries. Geographies allow customers with specific data-residency and compliance needs to keep their data and applications close. Geographies are fault-tolerant to withstand complete region failure through their connection to the dedicated high-capacity networking infrastructure of Azure.

What is Azure Regions?

Each Geography has multiple Regions. Example The United States has regions like Center US, East US, East US2, East US3, North Central US, South Central US, West Central US, West US, West US 2, West US 3.

An Azure region is a set of datacenters, deployed within a latency-defined perimeter and connected through a dedicated regional low-latency network.

What is Availability Zone in Azure?

Azure Availability Zones are physically and logically separated datacenters with their own independent power source, network, and cooling. Connected with an extremely low-latency network, they become a building block to delivering high availability applications. Availability Zones ensure that if there’s an event impacting a datacenter site—for example, if someone cuts the power, or there are issues with cooling—your data will be protected.

Running a VM with one or more replicated copies on different Availability Zones , providing resiliency against data center failure.

🍹 Tip

Azure Availability Zones are unique physical locations within an Azure region and offer high availability to protect your applications and data from datacenter failures. Each zone is made up of one or more datacenters equipped with independent power, cooling, and networking. The physical separation of availability zones within a region protects apps and data from facility-level issues. Zone-redundant services replicate your apps and data across Azure Availability Zones to protect from single points of failure.

What is the Azure global network?

The Azure global network refers to all of the components in networking and is comprised of the Microsoft global wide-area network (WAN), points of presence (PoPs), fiber, and others.

Azure Data-Center

In the Data Center the servers are divided into multiple physical and logical groups.

What is Availability Set?

An availability set is a logical grouping of VMs that allows Azure to understand how your application is built to provide for redundancy and availability.

What is azure fault domain?

The physical grouping of servers (virtual machines) is called Fault Domain. It is Each server share a common power source and network switch within a Fault Domain. A fault domain is a set of hardware components that share a single point of failure.

What is update domain in Azure?

Azure SLA

Azure Availability Zone give maximum industry best 99.99% of availability.

scenario	Manage VMs manually	Use VM Scale Set
Adding extra VM	Manually create, configure and ensure compliance	Automatically create from central configuration
Traffic Balancing & distribution	Manually	Automatically create and integrate Azure load balancer or Application Gateway
High availability and redundancy	Manually create Availability set or distribute and track VMs across Availability Zones	Automatic distribution of VM instances across Availability Zones or Availability Sets
Scaling of VMs	Manual monitoring and Azure Automation	Auto scale based on host metrics, in-guest metrics, Application Insights, or schedule

Become full stack developer 💻

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

What is Azure Virtual Machine Scale Sets

Rupesh Tiwari — Sat, 25 Dec 2021 00:00:00 +0000

Now a days distributed architecture is common. We deploy our services into many different servers to scale them up and meet our demand. However, managing all servers for load-balancing, scaling, make application highly available is super challenging on cloud. Azure Virtual Machine Scale sets is the great tool which does all of these automatically with no extra cost for you. Lets learn more about Scale sets in this article.

For Black Friday sell you want your services to be running in 50 instances and on regular day you want 5 instances. So in cloud Virtual machine scale set you get automatic scale up and down of your VM instances.

Scalability vs Elasticity

Scalability means to increase from 5 to 50 instances. Elasticity is to reduce 50 instances to 5. With VMSS scalability and elasticity is possible automatically.

Vertical vs Horizontal Scaling

Two types of scaling vertical and horizontal.

Vertical Scaling or Scale Up/Down

In Vertical scaling if you have 1 core CPU and 4GB memory and if you can upgrade this configuration to 2 core CPU and 8 GB memory. This is called as Scale UP. You can reduce the configuration back to baseline which is called as Scale Down.

Horizontal Scaling or Scale Out/In

Horizontal Scaling is you start with one VM and you keep adding more VM with same configuration is also called as Scale Out. You can also reduce back to 1 VM once your Sale or Business demand is over this process is called as Scale In.

Distributed Architecture with Scale Sets

You must need Distributed Architecture to provide redundancy and improved performance , applications are typically distributed across multiple instances. In order to give your customer faster speed and low latency you may need load balancer that distributes requests to one of the application instances.

What is Azure Virtual Machine Scale Sets?

Virtual Machine Scale Set (VMSS) let you Create and Manage a group of load balanced VMs. The number of VM instances can automatically increase or decrease in response to demand or a defined schedule.

With virtual machine scale sets, you can build large-scale services for areas such as compute, big data, and container workloads.
Scale sets provide high availability to your applications , and allow you to centrally manage , configure , and update a large number of VMs.
Consistent Configuration : Virtual Machine scale sets are the objects that are used to run multiple instances of your application and maintain a consistent configuration across your environment.

VMs in a scale set are identical, so you can create them from the same base operating system image.

How Virtual Machine Scale Set (VMSS) works

VMSS uses minimum instance to start with and you can set the maximum instance of your virtual machine. You can setup rules based on Time, Metrics based to increase or decrease VM instances. VMSS has in build load balancers. Public load balancer works with internet traffic to your VMs. Which looks upon the CPU metrics and if CPU utilization is more than 75% then wait for some time and add another VM instance without any manual steps required.

Minimum Instance

This defines how much instance of VM minimum you need. You can set it to 1.

Maximum Instance

You can go up to 1000 VM instances. For example for your business you can start with 1 VM and scale out up to 3 instances. Therefore, you set Maximum instance value to 3. Therefore, even though the demand is high Azure will not spin up 4th instance. It will stay up to 3 instance only and you save budget.

Metrics Based Scaling

If you don’t know where you are going to get the maximum business, it may be today, next day or any 5 continuous days then you must go for Metrics based scaling. You can set rule like If my VM CPU utilization is > 75% then add one more VM if it is less than ( < ) 25% then remove the VM.

Manually increasing instances : If you know tomorrow only you want to increase your VM instance to 4 instances. So you do that manually and revert it on the day after tomorrow.

Time Based Scaling

Custom : You can do Time Based increment or decrement Here you can schedule your VMs to scale out and scale in. For example every Saturday increase the VM instance to 4 and on Sunday reduce it back to 1. You can schedule these rules.

Load Balancer in Scale Set

An Azure load balancer is a Layer-4 (TCP, UDP) load balancer that provides high availability by distributing incoming traffic among healthy VMs. A public load balancer can provide outbound connections for virtual machines (VMs) inside your virtual network. Public Load Balancers are used to load balance internet traffic to your VMs. An internal (or private) load balancer is used where private IPs are needed at the frontend only. Internal load balancers are used to load balance traffic inside a virtual network.

Virtual Machine Scale Set Architecture

Virtual Machine Scale Set (VMSS) deploys Virtual Machines in a single subnet of a Virtual Network. Figure below shows Architecture of Virtual Machine Scale Set (VMSS) deployed in single subnet with Single Placement group.

Virtual Network with Single Subnet is automatically created during VMSS deployment. All virtual machines specified in create VMSS blade are deployed in the Single Subnet.

Azure Load Balancer is automatically created during VMSS deployment.

Placement group is availability set with five fault domains and five update domains and support up to 100 VMs. Placement group is automatically created by VMSS. Additional Placement groups will be automatically created by VMSS if you are deploying more than 100 instances.

Storage : VMSS can use managed disks or unmanaged disks for Virtual Machine storage. Managed disks are required to create more than 100 Virtual Machines. Unmanaged disks are limited to 100 VMs and single Placement Group.

VM Diagnostic Logs are guest OS performance counters and are streamed to Azure storage Account. VMSS uses this data for making Auto-scaling decisions.

Why use Virtual Machine Scale Sets?

Maintenance Mode Support : If you need to perform maintenance or update an application instance, your customers must be distributed to another available application instance.
Automatically increase VM instances : To keep up with additional customer demand, you may need to increase the number of application instances that run your application.
- Auto scale based on metrics.
- Auto scale based on a defined schedule. Suppose starting next week you are going to have a heavy peak, for next 3 days. You can define a set schedule. For example, at 9 am on Jan 2nd 2021, increase the VM instance count to 50. And at 9PM on Jan 5th when your peak ends, bring the instance count back to your baseline configuration.
Easy to create and manage multiple VMs : It maintains a consistent configuration (VM size, disk configuration) across your environment. All VM instances are created from the same base OS image and configuration. This approach lets you easily manage hundreds of VMs without additional configuration tasks or network management. For basic layer-4 traffic distribution it uses Azure Load Balancer. And for advanced layer-7 it uses Azure Application Gateway.
Provides high availability and application resiliency : If one of these VM instances has a problem, customers continue to access your application through one of the other VM instances with minimal interruption.
Allows your application to automatically scale as resource demand changes : Like it auto increase VM instances. It can also minimizes the number of unnecessary VM instances that run your application when demand is low, while customers continue to receive an acceptable level of performance
Works at large-scale : Up to 1000 Azure VM, and custom VM images up to 600 VM.

Scale set saves money 💰

The management and automation features, such as auto-scale and redundancy, incur no additional charges over the use of VMs. You only pay for the underlying compute resources such as the VM instances, load balancer, or Managed Disk storage.

Benefits of Scale Set

scenario	Manage VMs manually	Use VM Scale Set
Adding extra VM	Manually create, configure and ensure compliance	Automatically create from central configuration
Traffic Balancing & distribution	Manually	Automatically create and integrate Azure load balancer or Application Gateway
High availability and redundancy	Manually create Availability set or distribute and track VMs across Availability Zones	Automatic distribution of VM instances across Availability Zones or Availability Sets
Scaling of VMs	Manual monitoring and Azure Automation	Auto scale based on host metrics, in-guest metrics, Application Insights, or schedule

Using scale set is a wiser decision with zero additional cost!

Building Azure VM using Azure VM Image Builder

Using AZURE VM IMAGE BUILDER SERVICE, you can quickly start building standardized images without needing to set up your own imaging pipeline. Just provide a simple configuration describing your image, submit it to the Image Builder service, and the image is built and distributed. You will incur some compute, networking and storage costs when creating, building and storing images with Azure Image Builder. These costs are similar to the costs incurred in manually creating custom images.

Distribute your VM images

SHARED IMAGE GALLERY enables image distribution to different users, service principals, or AD groups, across multiple subscriptions within your organization and regions through a centralized image management platform.

Run below command for creating shared image gallery:

`New-AzGallery -GalleryName “azdemoGallery” -ResourceGroupName “Azure-demo” -Location “West US”`

Next, we need to create Gallery definition for storing windows images:

New-AzGalleryImageDefinition -GalleryName “azdemogallery” -ResourceGroupName “Azure-demo” -location “West US” -Name “winserverimages” -OsState generalized -Ostype windows -Publisher demo -offer windows -sku ‘win2016’

How to monitor scale sets

Use Azure Monitor for VMs, to see metrics about VM CPU, memory, disk and network performance
Enable monitoring virtual machine scale set application with Application Insights to collect detailed information about your application including page views, application requests, and exceptions.
Configure availability test to check application availability to simulate user traffic.

Learn more about Azure Virtual Machine Scale Set

Become full stack developer 💻

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

Create Update and Delete Tags on Azure Resource

Rupesh Tiwari — Sat, 18 Dec 2021 00:00:00 +0000

Learn more about how to use tags to organize your Azure resources and management hierarchy. Also learn the cmdlet used to fetch tags from Azure resources.

What is Tag in Azure Resources?

You apply tags to your Azure resources, resource groups, and subscriptions to logically organize them into a taxonomy.
Each tag consists of a name and a value pair. For example, you can create tag and apply the name “Environment” and the value “Stage” to all the resources in stage. Learn more about Tagging ideas
Tag names are case-insensitive : Example: “Environment” or “environment” they are same tag names.
Tag values are case-sensitive : Example: “Stage” or “stage” they are different tag values.

Learn more about how to use tags to organize your Azure resources and management hierarchy



Get-AzResource # To fetch all resources

Apply tags to Azure Resource

Use New-AzTag to apply tag on resource in Azure.



# Create Tag variable
$tags = @{"Dept"="Finance"; "Status"="Normal"}
# Get the resource that you want to appply tag on.
$resource = Get-AzResource -Name "Application Insights Smart Detection" -ResourceGroup AdmissionApplication
# Apply Tag
New-AzTag -ResourceId $resource.id -Tag $tags

View Tag on Azure Resource

Use Get-AzTag cmdlet to view tag on Azure resource. You have to pass the resource Id though. You can also use Get-AzResource cmdlet to view the tags directly under a particular resource.



# Get the resource that you want to view tag on.
$resource = Get-AzResource -Name "Application Insights Smart Detection" -ResourceGroup AdmissionApplicationGet-AzTag -ResourceId $resource.id
# View tags
Get-AzTag -ResourceId $resource.id

Delete Tag on a Azure Resource

To remove specific tags, use Update-AzTag and set -Operation to Delete. Pass in the tags you want to delete.



# Get the resource that you want to view tag on.
$resource = Get-AzResource -Name "Application Insights Smart Detection" -ResourceGroup AdmissionApplication
# View tags
Get-AzTag -ResourceId $resource.id
# Decide which tags to remove
$removeTags = @{"Dept"="Finance"; "Status"="Normal"}
# Delete the tags
Update-AzTag -ResourceId $resource.id -Tag $removeTags -Operation Delete
# Confirm Tags are deleted by fetching them again
Get-AzTag -ResourceId $resource.id

Delete Tag on a Azure Subscription

To remove all tags from your Azure subscription, use the Remove-AzTag command.



# Get subscription by it's name
$subscription = (Get-AzSubscription -SubscriptionName "Example Subscription").Id
# Remove the tags
Remove-AzTag -ResourceId "/subscriptions/$subscription"

Merge Tag



$tags = @{"Location"="Canada"}
# Merge Tag
Update-AzTag -ResourceId $resource.id -Tag $tags -Operation Merge

Replace Tag



$tags = @{"Dept"="Sales"; "Status"="High"}
# Replace tags
Update-AzTag -ResourceId $resource.id -Tag $tags -Operation Replace

Watch the entire videos on Full Stack Master YouTube Channel related to Managing Tags in Azure Resources

Become full stack developer 💻

Please subscribe to All-Access Membership PRO plan to access current and future angular, node.js and related courses.
Please subscribe to All-Access Membership ELITE plan to get everything from PRO plan. Additionally, you will get access to a monthly live Q&A video call with Rupesh and you can ask doubts/questions and get more help, tips and tricks.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com

How to improve Reliability in the cloud?

Rupesh Tiwari — Sun, 12 Dec 2021 00:00:00 +0000

Did you get a phone call from your customer saying they want to improve their application reliability while you move their workload to the cloud? If yes then continue reading this article. I will use Microsoft Azure to explain the Reliability concept and its improvement steps. You can certainly do the same in Google Cloud or Amazon AWS.

What is Reliability?

Well reliability is a derived concept. Basically you want your application to be available for your customer. Also you should ask what latency is acceptable for your application. If your application is slower or normal, your application users should not keep retrying to finish any transaction. How should your application fidelity be while your application is slow and not rendering properly. So in order to define reliability of your application you have to answer how much your application should be Available and what is your Latency that you want for your application.

Reliability = Availability + Latency

So basically reliability is something your Business has to define. You must talk to your customer and understand what kind of business they are running, what problem they want to solve, who are the target users for the application.

How to increase Reliability?

Reliability of your workload is a shared responsibility between you and your cloud provider. Platform Reliability such as datacenter, network and other hardware infrastructure is your cloud provider responsibility. However, application reliability is customer responsibility.

In order to increase application reliability you must do below:

Choosing and Configuring correct cloud building blocks.
Your application must be cloud aware. Like your code should have Retry logic to make sure that application is handling transient failures which is common in the cloud.

Right Building Blocks to increase Reliability

Make sure you choose the right infrastructure building blocks in the cloud to protect their reliability impacts. You have to isolate the concerns and identify them like below:

Isolated VM failure such as OS, disk HDD(hard disk drive) issue.
Hardware failure such as server rack issues.
Entire datacenter failure such as power/network issues.
Entire Region failure such as natural disasters.

In order to protect VM disk and OS issues you need to use Premium Storage from Azure Cloud. To protect from hardware failure you must create a VM inside the Availability sets. To protect from entire data center failure you must consider deploying your workload in multiple Availability Zones. Also to protect from natural disasters make sure you deploy your workload in another region as well. So in summary you have to choose correct building blocks as per your requirement to improve reliability.

How to Improve Reliability of an Existing Workload?

Start backup for your virtual machine, databases that you can enable for existing resources in Azure cloud. However, if your VMs are not in the Availability set then you can not put them in the availability set for existing resources. You have to do workaround and re-build some of them and put them in Availability sets again.

In order to improve reliability of existing workload you must follow below steps on each use case:

Find Key components for the critical use case (which business is more concern)
Failure Mode Analysis for the critical use case
Calculate Composite SLA for the critical use case
Improve Composite SLA to meet business requirement

Failure Mode Analysis Technique for Reliability

First you should talk to customers to find out which business capability they want to be up for what time. Example the purchase application written as asp.net needs to be up for 99.99 and 4 min per month down time. If business defines this then next you must learn their existing workload design consideration to achieve 99.99 SLA.

Designing for failure is the philosophy to take while designing applications in the cloud. Exploring what happens if something fails? Take the architecture, look at the portions of the application piece by piece and apply failure mode analysis. Break this down into your solution as a series of key processes. Like in your application, the customer goes and searches for an item to buy.

Step 1: Find the key components of the business use case

So now let’s search what component has to be working to make search possible?

Fo below architecture example in order to make sure Search is working below components must be up and running

AKS with .net application
Products service
Cosmos DB

Step 2: Failure Mode Analysis

Try to write down each component failure mode and their effect, impact and chances and work with your customer to learn more. This is a collaborative effort. Solutions Architect and Customer together must do this to come up with correct decisions. For search use cases you want to check what happens if AKS is unavailable, Website is slow, CosmosDB unavailable for read and update use case.

AKS is unavailable so we have to change the SLA to 99.95% on the API server of Kubernetes which is a controller plane. This is the SKU version we have to upgrade. In the worker plane we have a VM whose SLA is just like regular. So businesses have to think of their own requirements to opt in if they want to upgrade the SLA or kubernetes.
Web site is down, because of fluctuation of load due to holiday session. Probability is unknown. So you must use a monitoring tool setup alert to learn more about this kind of problem. LOAD OVER TIME can be only learnt by monitoring and it’s trends.
Cosmos DB can not read or update. Cosmos DB has 99.99% SLA for a single region. Cosmos DB can not be updated so customers can not see the latest catalog, it is also covered by 99.99% SLA out of the box by Azure cloud.

Step 3: Calculate Composite SLA

AKS SLA is 99.95% and Cosmos DB SLA is 99.99%

So overall composite SLA of AKS and Cosmos DB is 99.99 x 99.95 = 99.94%

Step 4: How to increase Composite SLA

To increase the composed availability add redundancy to the solution. Let’s make 2 AKS, if one breaks the other will work. You can put them in multiple Azure Region East/South USA. They have to be identical. Deployment of the cluster and the containers would have to be deployed in an automated way to keep them identical. IAC (Infrastructure as Code) should be in a repository and used to deploy your resources. Since AKS is in multiple regions you need DB in both regions. Cosmos DB has a multimaster database you can use in 2 regions so any change you do in one region is synchronized in both regions out of the box. You have to use Microsoft Front Door to route between multiple regions.

AKS (Azure Kubernetes Service) with redundant will give you SLA 99.99% per region with multiple region SLA is 99.999975%
Front Door SLA is 99.99%
Cosmos DB with multiple writable region SLA 99.999%

Composite SLA = 0.99999975 x 0.9999 x 0.99999 = 99.99%

How to know if something is wrong in your web app?

Create alert rules specify conditions and for actions you can alert humans with sms, or create action groups and execute some azure functions.

Enable VM Replication for improved Reliability

In addition to backup lots of customers are focused on disaster recovery and making sure the workloads are redundant and not just backed up. Meaning that if we had a planned or unplanned type of outage. Planned means upgrades to different applications, or we are testing our disaster recovery or may be unplanned like some natural disaster or things of nature. So introduce resilience in the form of being able to access backups that are in another region and to make those come online as quickly as possible to make sure that we reduce downtime to the extent possible.

Upgrade VM to Premium Disks to increase Reliability

For production workloads, especially Big Data, SQL and other transactional DB upgrades to premium disk instead of standar disk offering. Premium disks also offer bursting capabilities.

Burst means performance on heavy load.

On-demand bursting: you pay whenever you have burst. When you have an inconsistent workload.
Credit-based bursting: Harvest or credit yourself with additional bursting capability when you are operating below the maximum thresholds that we have. In this case, you’re able to burn those credits by bursting when you need it without any other cost implications. So you accrue your bursting credits when you are not using them and use it when you need.

Upgrade VM to Managed Disks to increase Reliability

Un-Managed Disks where you have to set the size, throughput limit, and add more resources to scale are called unmanaged disks.

However in Managed Disk you get below processes automated for you by your cloud provider:

Set the size, throughput limit, add more resource to scale handled by azure
Abstract the entire management layer of storage
Compatible with availability set and availability zones, you get 99.99% availability for production workloads.
Resiliency : 3 replicas across multiple hardware devices
Granular role based access control

How do I know what I don’t Know?

You have to have a long term learning process. Just wait for another disaster, unexpected event and learn. The other option is to simulate/force failures and see what happens. Say stop AKS and check if your redundancy is working. Take a more experimental model, introduce real world scenarios and see how our workload works. Game day, setup scenario, one team generating un-predictable load by un-usual way, another team observing, are we seeing alerts are coming, are we seeing redundant machine status.

Summary

In order to get Reliability increased for your customers workload you must speak with them and understand their critical use case which they care most about. Follow the steps that I suggested in this article to come up with the desired reliability state defined by business by doing failure mode analysis in depth one by one for each important use case. You may have to teach your customer about premium disks for VM and their SLA improvement because of upgrading to premium disk. What is managed and unmanaged disks after migrating to managed disk resilience can be increased since it automatically replicates the disk into 3 multiple hardware devices. Consider replicating virtual machines to protect them from regional outages. Setup Azure Virtual Machine backup for improved reliability and to protect from human error or DB corruptions. Also make sure you set up alerts for your applications to send you email or create zendesk tickets whenever it is down. I hope now you have a good understanding about reliability improvement.

💖 Say 👋 to me!

Rupesh Tiwari

Founder of Fullstack Master

Email: rupesh.tiwari.info@gmail.com

Website: RupeshTiwari.com