DEV Community: OverFlow

Challenges in Implementing Business Monitoring for Banking IT Systems

OverFlow — Thu, 04 Apr 2024 03:49:59 +0000

Other Languages

In our previous discussion, we identified common business indicators and monitoring rules. However, several difficulties arise during implementation, primarily in the following areas:

1. Acquiring basic transaction information

To effectively monitor transactions, the system must accurately collect data for each transaction, including the request system, transaction code, important transaction flags, service system, processing node, start and end times, serial number, institution, amount, service system return code, and this system's return code, among other details. Unfortunately, many systems cannot provide complete basic information, such as the service system return code. If the production system wasn't initially designed to record this field, it's challenging to provide it. Without this data, calculating the service system transaction success rate is impossible. Additionally, all systems provide transaction detail data to the monitoring system, resulting in large data volumes that challenge the monitoring system's processing performance.

2. Development specifications for production systems

Monitoring rules apply to all production systems, so standardization and normalization of production system development are essential. For example, there must be a strong dependence on the transaction code parameter. To achieve this, the transaction code parameter should include the request system, transaction code, transaction name, service system, financial flag, busy flag, and idle flag, among other details. Ideally, a single transaction code should perform only one function. For instance, transaction code A009 has non-standard functional decomposition, causing an atomic transaction code A009 program to use a sub_type-like field for functional branching internally. As a result, a single transaction code implements multiple business functions, making it difficult to clearly describe the business functionality of A009.

Standardizing development code is another significant challenge. The accuracy of alarm triggers depends on the response code and log recording provided by the production system. While working on Bank X's central clearing system, the author encountered a classic response code: 201999-System error. This error typically indicates a serious system fault that the program cannot process from a technical perspective or determine the cause. However, this response code is returned approximately 800,000 times daily, resulting in an overall daily transaction success rate of less than 90%. Moreover, return codes may have multiple uses. For example, return code R00001 should signify "No record found," but developers may customize it to mean "Insufficient balance," "Card number xxx error," "Address error," or even the opposite, like "Query returned xxx records." In such cases, the return code cannot determine whether a transaction is successful. This problem becomes more pronounced in business monitoring that drills down to the transaction code level, where transaction volume is smaller, and sensitivity to return values is higher. Standardizing return codes is urgently needed.

Log specifications should also include error log standardization. Accurate log monitoring alarms depend on standardized log recording; otherwise, false alarms and missed alarms may occur frequently.

3. Setting numerous running parameters

If the previous challenges are addressed, operations personnel must continuously fine-tune various running parameters. Business-class monitoring requires parameter settings for each transaction code, including business success rate, system success rate, surge ratio, drop ratio, indicator deviation baseline ratio, and the number of abnormal transactions. These parameters describe the transaction code's running situation. Larger systems have many transaction codes, requiring long-term adjustment of tens of thousands of running parameters for the transaction code parameter alone. However, if parameter setting rules are established, an algorithm can be designed to automatically modify running parameters based on the baseline running situation.

Designing a monitoring system is not difficult, but the challenge lies in making monitoring more relevant to the business, fully exploiting the value of business monitoring, and driving the transformation from IT maintenance to IT operations.

Original article

Netis specializes in Business Performance Analysis (BPC) and Network Performance Monitoring (NPM). Each day, our transaction monitoring solutions ensure the smooth completion of over 30 billion transactions. We prioritize the security and autonomy of systems and data. Utilizing real-time analysis through network traffic mirroring, we ensure zero intrusion into business operations, without the need for any agent installations. Headquartered in Shanghai, China, Netis is a global software company renowned for its high-performance, cost-effective products. If you're interested in our offerings, please don't hesitate to reach out. We eagerly await your trial and partnership.

For further details, visit: www.netis.com/en/

Contact me via email at: dennis.li@netis.com

Crafting Business Monitoring Metrics for Bank IT Systems

OverFlow — Sun, 31 Mar 2024 01:00:18 +0000

Other Languages

As the financial sector evolves, banking services have grown in complexity alongside a diversification in information system architectures. The challenge of identifying issues through purely technical monitoring rules is escalating. There's a clear demand for business-centric monitoring strategies that are agnostic to technology architectures, focusing solely on depicting business operations. Drawing from extensive operational experience and a deep understanding of business monitoring, the author delves into the metrics and rules essential for effective business monitoring.

Evolution of Monitoring Practices

First Stage: IT operations staff manually establish monitoring metrics, rules, and parameters. A common approach involves configuring scripts within a monitoring system like Zabbix, setting up specific alarm conditions based on thresholds, where crossing these thresholds (either above or below) triggers an alert.

Second Stage: Monitoring metrics and rules are standardized, with IT operations staff responsible for the upkeep of operational parameters. This stage builds upon the first by organizing and standardizing monitoring criteria to ensure broad compatibility and uniformity across various systems.

Third Stage: The monitoring system autonomously determines the metrics, rules, and optimal operational parameters by analyzing both its own performance and that of the production systems, minimizing the need for manual intervention by IT staff.

The focus of this discussion is the pivotal second stage, proposing a set of universal monitoring metrics and alert rules derived from operational insights into business activities.

Key Metrics and Alert Rules for Monitoring Banking Services

Banking information systems are defined by their transaction codes or interfaces, characterized by return codes, transaction volumes, and response times. This guide outlines essential monitoring metrics and alert rules based on these characteristics, offering a reference for effective business monitoring.

1. Transaction Success Rate

A cornerstone metric in system monitoring is the transaction success rate, essential for assessing the basic health of a system. Traditional success rate monitoring deems a transaction successful in the absence of systemic errors (e.g., network issues or file access errors). Yet, a lack of systemic errors does not guarantee the absence of business process issues. Take transaction A001, "query posting", which normally indicates "no record found". An unusual spike in postings indicates a potential issue, suggesting a need for differentiating between system success rates and business success rates. The system success rate evaluates technical failures, while the business success rate measures transaction outcomes against expected business return codes. This nuanced approach allows for more precise monitoring, moving beyond mere success rates to consider specific expected outcomes for each transaction.

Figure 1: Transaction Success Rate Page

2. Service System Transaction Success Rate

A unique aspect of banking systems is the clear distinction between calling and service systems for each transaction. Understanding the performance of service systems is crucial for pinpointing issues. Simply determine whether a transaction invokes a service system and which one it is. Analyzing the success rates of transactions linked to a specific service system provides insight into its reliability. Further granularity, such as assessing individual transaction codes within the service system, can offer deeper diagnostic capabilities.

3. When Nodes Go Quiet: The No-Transaction Alarm

In the whirlwind world of high-frequency trading, there's a clever trick: setting up an alarm for when, surprisingly, nothing happens. Imagine a bustling marketplace that suddenly goes silent. This alarm rings when a specific transaction doesn't occur within a set period at a particular node - think of it as a "missing action" alert. It's a bit like expecting a daily call from a friend and starting to worry when it doesn't come. However, figuring out when to expect the call (or transaction) isn't straightforward, as busy and quiet times vary, just like weekends and weekdays. It's about tracking every transaction's ebb and flow, picking out the ones that should be happening now. And so, we bring in the idea of "rush hours" and "down times" - for instance, marking weekdays from 8 a.m. to 7 p.m. as peak times. If transaction code A002 is buzzing all day, it's watched non-stop. But if A003 only wakes up during the rush, it gets a break during the lulls. And A004? If it barely whispers throughout the day, it's left alone, unmonitored.

4. The Sudden Swings: Tracking Transaction Tsunamis

Every day, transactions flow smoothly like a river through its channel. But when the system hits a snag, this flow can turn into a flood or a drought. Spotting these drastic changes is crucial. Imagine tracking a transaction, A005, during the afternoon rush. Comparing its activity to a "normal" day, we calculate the change. If it falls within the range of typical ups and downs (let's say, 0.8 to 1.2 times the norm), all's well. But if it plunges to 0.4 or soars to 1.6, it's time to sound the alarm. This method, however, stumbles a bit with the less frequent transactions, where even minor changes can seem like big waves.

5. Monitoring the Pulse: Tracking Abnormal Transaction Volumes

Diving into transaction success rates with a microscope, particularly at the granular level of individual codes, reveals a challenge: this approach doesn't play well with the extremes of low or ultra-high-frequency transactions. Picture this: low-frequency transactions end up setting off alarms like fireworks, while ultra-high-frequency ones might fly under the radar, missing the alert that trouble's brewing.

Take transaction A006, bustling with 10,000 transactions in just 10 minutes, 50 of which hit a snag. That nudges the success rate down by a mere 0.5%—seemingly minor yet potentially disastrous. Meanwhile, A007 logs just two transactions in the same timeframe, but given its heavyweight financial status, even a single hiccup is hard to ignore. The solution? Tailoring alert thresholds—over 10 misfires for A006, just one for A007—based on the beat of each transaction's rhythm, efficiently sidestepping the issue. And imagine a dashboard that lays out the nitty-gritty of all transactions gone awry, from the processing node to the service system, making emergency troubleshooting a breeze.

6. The Time Crunch: Response Time Overshoots

Bank systems, like any well-oiled machine, can sputter, slowing down or even stalling. Traditional monitoring might blink at response rates, but what about when delays linger just shy of the alarm threshold? Especially in high-traffic transactions, a slight delay doesn't always mean crossing the red line.

Here’s a fix: keep an eye on transactions straying from the norm in response times, cycle by cycle. For instance, if A008 typically wraps up in 300ms, stretching beyond 600ms or 900ms could signal trouble. A trio of these outliers within a cycle? Time to sound the alarm.

7. Closing the Day, Keeping the Books

The curtain call of a banking system's day, end-of-day processes, holds the key to a smooth opening act the next morning. Monitoring is straightforward: transactions tagged for day's end need to kick off and wrap up within set times, with a bow tied by a success signal. And when it comes to balancing the books, consistency is king—minor fluctuations in discrepancies are the green light for business as usual.

That wraps up our exploration into crafting smart, sensitive business monitoring tools. Stay tuned for a peek behind the curtain at the tech challenges that come with bringing these concepts to life.

Original article link: https://mp.weixin.qq.com/s/qlvqPsz2fX0AyxMdXdVzxw

High-Frequency Trading System Monitoring

OverFlow — Sat, 23 Mar 2024 04:57:50 +0000

Other Languages

In the evolution of trading systems, traders have consistently sought speed. With the growing popularity of quantitative trading, the demand for lightning-fast systems has only increased.

Challenges in Monitoring High-Frequency Trading Systems

The advent of high-frequency trading systems has significantly increased the complexity of application operations and maintenance management, demanding unprecedented accuracy, performance, and stability from monitoring tools.

Nanosecond-level monitoring precision required: High-frequency trading necessitates the rapid processing of orders and low-latency market feedback, with transaction processing speeds typically in the microsecond range. Traditional monitoring tools fall short of achieving microsecond or even nanosecond precision, unable to keep up with the high-speed demands of high-frequency trading systems.
Need for specialized protocol decoding: Traditional monitoring tools lack the comprehensive decoding capabilities needed to fully support the automatic decoding of mainstream high-frequency trading systems.
Minimizing impact on system performance: Traditional tools that log or insert code consume significant system resources, compromising stability and failing to offer precise monitoring of high-frequency trading systems.

Thus, application operation and maintenance teams require monitoring tools that are high in precision and performance, with robust decoding capabilities. Such tools can outpace high-frequency trading itself, tackling issues like difficult troubleshooting, pinpointing problems, and complex analysis to ensure the seamless operation of high-frequency trading systems.

BPC High-Frequency Trading Edition

For years, Netis BPC has been pivotal in ensuring operational continuity for clients in the securities sector. By capturing network traffic via bypass deployment, Netis BPC delivers essential features like four key indicators, intelligent alerts, individual transaction tracking, and multi-dimensional statistical analysis. The upgraded "BPC High-Frequency Trading Edition" introduces nanosecond-level monitoring, offering comprehensive maintenance support for high-frequency trading systems.

Nanosecond Timestamp Technology

The core challenge for high-frequency trading systems lies in managing penetration, uplink, and full-path latencies, mostly measured in microseconds. This necessitates monitoring solutions capable of nanosecond precision. The "BPC High-Frequency Trading Edition" elevates precision from microseconds to nanoseconds, allowing for the swift calculation of time spent on each trading segment and across the entire chain. With accurate latency measurements, it immediately alerts users to any significant order delays, pinpointing issues with features like latency-thruput correlation and one-click tracking.

A Business-Centric Monitoring Approach

"BPC High-Frequency Trading Edition" intelligently maps out the logical relationships between systems following actual business processes, offering a service-oriented perspective. It uses service path diagrams for real-time monitoring of core middleware and exchange components, providing essential metrics such as volume, response times, rates, and success rates, enabling quick issue resolution.

Figure 1: Service Path View

Multi-Segment Transaction Tracking

This feature smartly associates data from individual orders across the trading chain and key nodes. With just a click, users can analyze the processing times of each order at various nodes.

"BPC High-Frequency Trading Edition" comprehensively analyzes order requests, internal responses, and exchange response messages for high-frequency trades. Utilizing high-precision timestamps from high-performance network cards, it calculates comprehensive metrics like full-chain and middleware processing times, aiding in detailed analysis and allowing for the flexible creation of securities monitoring views adjusted to the second.

Figure 2: Multi-Segment Transaction Tracking Feature

Full Support for Mainstream High-Frequency Trading Systems for Immediate Monitoring

Leveraging Netis's leading decoding capabilities, "BPC High-Frequency Trading Edition" now fully supports major high-frequency trading platforms such as Hengsheng, Huarui, and Dingdian, along with exchange multicast systems like SSE and SZSE, and private market data systems. BPC's adaptable decoding supports quick integration with third-party protocols, facilitating instant deployment and monitoring.

How to Gracefully Implement Business Performance Monitoring

OverFlow — Sat, 16 Mar 2024 12:51:24 +0000

Other Languages

In an era overflowing with monitoring tools, how can we streamline the complexity and assist users in gracefully implementing business performance monitoring?

1. A Business-Centric Monitoring Perspective

Considering the disparities between monitoring metrics and intentions, application performance monitoring (APM) solutions offer varied alerting logics, broadly categorized into two types: "Bottom-up" alerting involves establishing alerts based on packet headers and other foundational data, then ascending to app and service utilization levels, ultimately identifying anomalies such as DNS errors or slow response times. On the flip side, "Top-down" alerting monitors business application nodes and select network nodes, setting alerts for high-level aggregated data. This approach starts with overarching business challenges and delves into underlying data. Due to the high false alarm rate and troubleshooting challenges of bottom-up data, it's not only time-consuming but also less precise. Conversely, top-down alerting can swiftly trigger alerts at the first sign of system or service failures, a crucial feature within a business-focused operational framework.

Figure 1: Knowledge Hierarchy Diagram

Netis BPC employs a "top-down" monitoring strategy, offering end-to-end coverage that enables a comprehensive view from the macro to the micro. By centering on business needs and bridging the divide between operations and management, it ensures unified monitoring and business perspectives, greatly enhancing the precision and clarity of alert analysis.

2. Designing with the User in Mind

Imagine using piano keys to track time and a simple four-quadrant grid to display crucial business metrics. This innovative approach sits at the heart of Netis BPC's design ethos.

Figure 2: How piano keys and a four-quadrant grid play a role in Netis BPC's alert system and troubleshooting.

Why is each minute represented by a piano key?

Netis BPC uses piano keys as a metaphor for time to give users a real-time sense of their business operations in a way that turns repetitive and stressful tasks into a more enjoyable experience. Each minute is a key on this timeline piano, with different colors conveying various states of business health. Interacting with these keys, users can "play" through their data, zooming in on specific moments with a mouse hover, much like enlarging a key under a pianist's finger. This attention to detail ensures users can easily and precisely access the information they need.

Furthermore, the choice of minute-level alerts strikes a balance between promptness and practicality. From extensive experience with financial clients, Netis learned that too detailed an alert system overwhelms with repetition, while too broad a system risks delays and oversights. Thus, setting alerts to minute intervals—each piano key—was a deliberate choice for efficiency and effectiveness.

However, minute granularity doesn't fit every scenario. For instance, in stock trading, where the pace is faster and stakes higher, even seconds matter. A transaction delay of more than 10ms or 10 specific events within a second can trigger alerts in this high-speed context. Thus, Netis BPC also offers second-level alerting, a faster, more sensitive detection system tailored for such critical needs.

Displaying the Golden Signals

Netis has distilled the essence of business monitoring into four golden signals: transaction volume mirrors application load, response rate echoes system health, success rate reflects business well-being, and transaction response time inversely indicates customer satisfaction. This revelation came from analyzing and comparing industry metrics like the Four Golden Signals, RED, and USE, tailoring them specifically for the financial sector's unique needs.

But how do you make these critical indicators clear and engaging for users? Netis's breakthrough came from marrying thorough analysis with creative design, resulting in the pioneering four-quadrant grid.

This design leap was inspired during a visit to New York's Museum of Modern Art by Netis co-founder & Chief Product Officer, Wizard. A piece showcasing America's geography and population through a clear, grid-based diagram sparked the idea. It showed that the power of information lies not in its volume but in the precision of its presentation. From highway signs to information guideboards, clarity reigns supreme. This insight, coupled with the need for quick, clear alert information in IT, birthed the iconic four-quadrant grid.

Figure 3: New York's Museum of Modern Art's classic four-quadrant grid:

Figure 4: Netis BPC's signature four-quadrant grid:

At its core, BPC's product philosophy is about deeply understanding user needs and designing solutions that meet those needs precisely. By embedding the users' perspective into the design, notably with the piano keys and the four-quadrant grid, Netis champions a user and service-centered approach. This philosophy not only addresses user pain points but also bridges the gap between product functionality and user value.

Cloud Network Traffic Collection and Analysis White Paper (2022)

OverFlow — Sat, 02 Mar 2024 07:32:55 +0000

Other Languages

IDC's recent study forecasts that in the upcoming two years, a majority of companies will lean towards hybrid and multi-cloud strategies to develop their cloud-native platforms. As businesses expand their cloud-based activities, the internal (east-west) traffic has unexpectedly skyrocketed, posing a significant challenge: how to precisely collect and manage this data flow. In light of this, Gartner advises: "IT leaders should revamp their traffic management strategies, enhancing visibility within hybrid environments to bolster future performance tracking and control."

Netis posits that analyzing cloud network traffic observability is a key strategy for improving performance management in both cloud-native and hybrid settings. Utilizing advanced, comprehensive, and non-invasive techniques for data collection and processing, organizations can achieve a detailed view of their cloud network traffic across various setups and frameworks, leading to total network transparency. To aid companies in navigating these hurdles, Netis presents the "Cloud Network Traffic Collection and Analysis White Paper (2022)." This guide is organized into six main sections: the challenges presented by the cloud era, the benefits of cloud network traffic observability, frameworks and technologies for data collection and analysis, practical deployment strategies, examples from typical scenarios, and best practices across industries. This comprehensive approach aims to furnish every sector with the most effective cloud migration strategies.

The white paper highlights the following key insights:

1. Key Capabilities

Versatility Across Environments:

Comprehensive support for various settings: compatible with leading operating systems, chipsets, and both virtualized and containerized frameworks.

Microservice Compatibility: Adaptable to a range of deployment models within a microservice architecture, alongside support for diverse container runtimes.

Broad Skill Set:

Packet Handling: Facilitates key operations on data packets, including collection, filtering, distribution, trimming, and eliminating duplicates.

Flow and Telemetry: Essential for cloud network traffic visualization, this involves preprocessing collected raw data packets, generating Flow logs, and enabling Telemetry functionalities.

Data Packet Tagging: Allows for the tagging of collected raw data packets, integrating details from business, network, and security contexts directly into the data.

Reliable TCP Transfers: Utilizes TCP protocol for the secure and dependable transmission of collected data packets.

Non-intrusive Operations: Designed to be minimally invasive, it requires no alterations to existing business processes.
Autonomy in Management: Encompasses integration with management platforms, real-time status updates on collectors, self-monitoring, automatic problem resolution, and comprehensive visualization capabilities.
Scalable Architecture: Supports seamless scaling to adapt to changing demands.

2. Key Technologies

Collection Technologies:

Application Scenarios: Classified into major categories such as public and private clouds from the application standpoint; from a technical angle, broken down into VMs and container technologies.

Capabilities: Initially, traffic collection must span diverse scenarios with specific collection strategies tailored for each. Furthermore, in mixed environments, layered traffic collection requires integration into a singular management platform.

Layered Approach/vTAP: Netis advocates for a comprehensive approach in cloud traffic management, encapsulating layered collection, processing, refinement, and management. This strategy is vital to maintain granularity and overall observability amidst vast arrays of managed objects and substantial traffic.
Fundamental Collection Techniques: Given the cloud's intricate and varied base layer, selecting the ideal collection technology tailored to distinct environments is crucial for maximizing performance and efficiency.
Support for Autonomous Environments: Cloud network traffic and observability analysis should cater to self-sufficient servers, including those running on ARM architecture or bespoke operating systems.
Enhanced Flow Support: Implementing detailed Flow analytics tailored to specific consumer scenarios across networks, applications, and security can significantly address the balance between value and cloud network expenditures.
Advanced Telemetry: Integrating Open Telemetry features enhances observability capabilities while concurrently minimizing the resource footprint on cloud network traffic.

3. Deployment Strategies

This section outlines five key strategies: comprehensive collector setup, host-based traffic gathering, virtual machine traffic monitoring, containerized environment data collection, and vTAP segmentation. The white paper provides a thorough breakdown of each strategy.

4. Key Use Cases

Security Monitoring: Effective cloud network traffic systems are crucial for precise, real-time monitoring of internal cloud service traffic.

Situational Awareness: In the era of cloud integration, understanding the security landscape within business processes is critical. This demands accurate cloud traffic analyses supported by foundational security mechanisms.

Enhanced Security Compliance (Level Protection 2.0): Ensuring high-level security standards such as Level Protection 2.0 necessitates comprehensive cloud network traffic oversight.

Business and Application Monitoring (BPM/APM): Netis highlights how BPM, aided by network traffic insights, swiftly pinpoints user experience issues. Coupled with APM’s in-depth application analysis, this offers a holistic approach to cloud observability.
Network Performance Monitoring and Diagnostics (NPMD): Cloud monitoring diverges from traditional methods, incorporating elements like Availability Zones (AZ), Virtual Private Clouds (VPC), and services. This entails advanced capabilities in cloud structure analysis and in-depth, multi-tiered session tracking.

Cloud Structure Insights: Understanding cloud network architecture requires robust monitoring, supported by specialized cloud traffic collection and tagging techniques, to navigate the cloud's dynamic nature and prevent data bottlenecks.

In-depth Network Tracking: Detailed session tracking in cloud networks demands sophisticated collection and tagging techniques to follow traffic through multiple layers.

Database Management and Security: The advent of distributed architectures necessitates a new approach to database monitoring. Frequent interactions between data and access nodes heighten the risk of issues. Opting for cloud traffic observability for distributed database oversight offers a superior solution.

Download Here

Netis NPM Empowers Customers' IPv6 Upgrade

OverFlow — Sun, 25 Feb 2024 06:20:59 +0000

Netis NPM Empowers Customers' IPv6 Upgrade

Other Languages

Background

In July 2021, the Cyberspace Administration of China, the National Development and Reform Commission, and the Ministry of Industry and Information Technology collectively released a directive to accelerate the widespread adoption and application of Internet Protocol Version 6 (IPv6). The directive specifically targets "advancing the IPv6 upgrade within financial institutions, continuously enhancing the IPv6 support capabilities of internet-facing financial service applications, and strengthening the IPv6 monitoring and maintenance framework." This initiative is identified as a critical step towards the broad deployment and use of IPv6.

Overview

One bank took this initiative seriously, dedicating efforts to fulfill the IPv6 upgrade mandate. By early 2022, it had developed a network infrastructure that supports both IPv4 and IPv6 (a dual-stack network), enhancing IPv6 compatibility across various customer-facing applications like its website, online and mobile banking, and WeChat banking services. To boost the operational effectiveness of this dual-stack network and fortify its maintenance structure, the bank adopted Netis NPM, a cutting-edge, comprehensive traffic analysis system. This system is designed to perform real-time monitoring of internet traffic across critical network segments and applications.

Value of the Case

Network Traffic Visualization: Offers a fast and intuitive view of the network's topology and details, setting the stage for uncovering the network's internal business potential.
Service Path Mapping from a Business Perspective: Moves beyond the isolated worlds of traditional network and application maintenance;
Comparative View of Internet Applications in IPv4 and IPv6: Stay informed about the usage, performance metrics of IPv4 versus IPv6 applications, enabling dynamic adjustment of network resources to swiftly meet application needs;
Accurate Network Anomaly Detection and Quick Alerts: Reduces time to resolve issues, enhancing network service quality.

1. Netis NPM: A Comprehensive Network Traffic Management Platform

Built on proven network mirroring technology, the full-traffic network analysis system duplicates network traffic in real-time to a Traffic Analysis Probe (TAP). This TAP then aggregates, tags, and distributes traffic to various analysis probes based on predefined rules. These probes dissect and process the traffic, after which an all-in-one management platform consolidates and presents the findings.

Figure 1: Netis NPM's Logical Deployment Blueprint

Through comprehensive traffic analysis on internet exit pathways and the networking hardware surrounding security mechanisms, it crafts distinct, visual perspectives for both IPv4 and IPv6. This allows for real-time monitoring of network health, enabling IT staff to swiftly grasp daily network operations.

Figure 2: View of an IPv4 Dedicated Exit Link - Application Performance Metrics

Figure 3: View of an IPv4 Dedicated Exit Link - Network Traffic Load

Figure 4: View of an IPv6 Dedicated Exit Link - Network Traffic Load

2. The Innovative SPV: Network Insights Through the Application Lens

Normally, network and application management are siloed into separate teams, leading to a fragmented response when issues arise, which slows down troubleshooting and problem-solving. Netis NPM breaks new ground with its Service Path View (SPV) technology. By creating a comprehensive, end-to-end monitoring framework centered around the business's perspective, it shifts the focus onto how applications interact with the network. The SPV illuminates an application's network condition across four key areas: traffic volume, network performance, application performance, and availability. This empowers IT teams to quickly pinpoint and resolve any issues as soon as they arise, ensuring smooth application operation.

Figure 5: The Netis NPM's Service Path Diagram from a Business Viewpoint

3. Everything at a Glance: IPv4 vs. IPv6 Network Metrics Comparison

At the gateway to the internet, one bank has leveraged the professional capabilities of Netis NPM to create a large display that contrasts the network performance of IPv4 and IPv6. This setup allows for real-time monitoring of the exit paths to the internet, significantly boosting user satisfaction by ensuring smooth connectivity.

Figure 6: The Grand Display of Netis NPM

For critical services accessible to the public, the bank designs detailed reports on a daily, weekly, and monthly basis, tailoring these insights to meet the demands of oversight bodies.

Figure 7: Weekly Insights for Key Services

Netis's suite of products offers robust support for IPv6 analysis, playing a key role in maintaining the reliability and stability of IPv6-based services for their clients.

At Netis, we're at the forefront of analyzing business(APM) and network performance(NPM), ensuring that over thirty billion transactions run smoothly every day. Our commitment to system and data security is paramount, and we pride ourselves on our ability to conduct real-time analysis through network traffic mirroring. This means we can get the job done without the need for any agents or disruptions to your existing operations. Based in Shanghai, China, Netis is a global software company known for its high-performance products and cost-effective solutions. Interested in what we offer? Don't hesitate to get in touch. We're excited about the possibility of working together and welcoming new partners.

Discover more at: www.netis.com/en/

Or drop me a line directly: dennis.li@netis.com

User Case: Smart Business Performance Monitoring in Financial Private Cloud Hybrid Architectures

OverFlow — Fri, 16 Feb 2024 01:43:39 +0000

Other Languages

As the foundational platform supporting the operation of information systems, financial institutions' IT infrastructures face the twin challenges of digital transformation and technological autonomy. Consequently, the financial private cloud has emerged as a preferred solution for many financial institutions to address these challenges.
In recent years, a city commercial bank has dedicated itself to the vigorous development of private clouds as a new embodiment of fintech-enabled business. It has partnered with leading cloud service providers for joint research and breakthroughs in financial cloud infrastructure and services, financial scenario innovation, and the development of cloud-native business systems, collaboratively advancing the deployment of the bank's financial private cloud technology architecture. This strategy of consistently deploying or migrating critical business systems to the financial private cloud has yielded numerous benefits, such as enhanced business agility, cost optimization, and improved security. However, it has also introduced significant challenges to operational management, including:

Achieving unified traffic collection and management across hybrid cloud architectures like cloud servers and container clouds;
Navigating complex business access relationships and monitoring blind spots in hybrid architectures;
Dealing with more concealed failures in cloud environments, complicating troubleshooting efforts;
Streamlining operations to enhance efficiency under the hybrid cloud architecture. In response to these challenges, Netis and the city commercial bank have collaboratively developed the "Financial Private Cloud Hybrid Architecture Intelligent Business Performance Monitoring Solution," bravely confronting and advancing through these difficulties.

Challenge One: Overcoming Traffic Collection Hurdles in Hybrid Cloud Architectures

As the city commercial bank embraces its financial private cloud, an increasing number of business systems find their new homes on cloud servers and container clouds. The first hurdle is how to efficiently and securely gather traffic across this mixed cloud setup, merge it with traffic from traditional setups, and manage it all under one roof. Netis has ingeniously addressed this by introducing a suite of probe deployment strategies (involving both bypass and host probes) tailored to user scenarios. This approach not only facilitates comprehensive traffic collection within the hybrid cloud framework but also, through the Cloud Probe Manager, ensures centralized control over these collections. The culmination of this effort is the integration of traffic data, both from traditional and hybrid cloud sources, into Netis's Business Performance Management (BPC) system.

Figure 1: Deployment architecture of the monitoring solution.

Challenge Two: Navigating Business Path Analysis in Hybrid Cloud Architectures

As cloud-based systems proliferate, delineating business paths becomes increasingly crucial within architectures that blend traditional and cloud environments. Leveraging wire data decoded by BPC, the system autonomously navigates business processes, rendering the entire business system's service paths in a visually accessible format. This not only facilitates real-time visibility but also enables the intelligent amalgamation of applications at identical nodes. This comprehensive visibility aids city commercial bank clients in demystifying the service nodes and applications across their operations, allowing for meticulous path analysis and the elimination of monitoring blind spots. This ensures that the tech department can masterfully manage and understand the business system's intricacies.

Figure 2: SPVD's automated business path analysis.

Challenge Three: Navigating Troubleshooting in Hybrid Cloud Architectures

The nature of cloud architectures—characterized by their flexible connections and ever-changing application nodes—means that failures within the cloud are sneakier and their causes more complex than ever. This complexity turns troubleshooting in cloud-based business systems into a formidable task. Without the right tools, delays in failure notifications, imprecise fault localization, and inaccurate analyses can seriously disrupt business operations, potentially impacting customer-facing services.

To safeguard the smooth operation of cloud-based systems, Netis has implemented a strategy using five sets of alarm presets, specifically designed to monitor for common yet critical financial transaction failures. This system is bolstered by automated fault localization and root cause analysis, enabling a thorough evaluation of the failure's impact. The end goal? To quickly identify the root cause and any abnormal performance metrics, facilitating swift resolution. This approach ensures that the maintenance response times of cloud-based systems are on par with those of traditional setups.

Figure 3: Illustration of intelligent alarm scenarios and triggers.

Challenge Four: Overcoming Operational Management Hurdles in Hybrid Cloud Architectures

The city commercial bank's financial private cloud has embarked on a groundbreaking journey, thanks to its evolving partnership with cloud service providers. This evolution has propelled operational management, especially within a hybrid cloud framework, to the forefront of the technology operations department's agenda. Consequently, the deployment of a unified intelligent monitoring platform has become a necessity. Netis Business Performance Management (BPC) stands out by seamlessly integrating hybrid and traditional environments into one monitoring interface, ensuring comprehensive, end-to-end coverage without any blind spots. This achievement not only aligns with the operational management goals of the city commercial bank but also significantly boosts efficiency in this domain.

Figure 4: Comprehensive Service Path - Unified Monitoring Across Cloud and On-Premises Environments

Through its sophisticated multi-layer tracking, Netis BPC empowers the bank to seamlessly track transactions across cloud and on-premises systems using business transaction numbers. This capability simplifies the task of achieving thorough, end-to-end monitoring of services, irrespective of their location.

Figure 5: Multi-Layer Tracking - End-to-End Monitoring for Individual Transactions

Netis BPC's advanced application performance metrics component vividly displays the real-time status of the entire business system. Its intuitive graphical interface efficiently showcases various key metrics, facilitating visualized performance monitoring. The platform adeptly presents essential application performance metrics, including the "Five Golden Metrics," along with a wide range of business dimensions through dynamic curves, pie charts, and bar graphs. This includes detailed views of application layers, snapshots, multi-dimensional statistics, and transaction tracking features.

Figure 6: Netis BPC - Bridging Business and Performance Analysis