When it comes to storage efficiency and performance, some of the basic key terms to understand are:
- Latency
- IOPS
- Throughput
Having a full grasp around these key terms will help pick the best storage for your application.
In addition to picking up the right the volume type for your application , it is also important to confirm that the volume picked is compatible with the instance type you plan on using.
I know it seems like a lot of figure out but thankfully AWS compute Optimizer can help alleviate that burden by identifying optimal configurations using machine learning.
To learn more about instance and volume compatibility, check the resource that AWS has put together here
Latency
Latency is the true end-to-end client time of an I/O operation; in other words, it is the time elapsed between sending an I/O to EBS and receiving an acknowledgement from EBS that the I/O read or write is complete.
It is measured in unite of time and the lower the latency is the faster is the transaction time.
It is worth mentioning that Several other factors can affect the latency as well. Among them are block size and QueueLenght. The QueueLenght the number of transaction request pending. However the QueueLenght can be also controlled by assigning a value to it. That value will determined how many IP operation can line up.
IOPS
IOPS is the Input Output Operation per second, in other terms, it is the measurement of the number read and write operations that can be carried out one second and that is used to asset the performance of a storage . SSD are the best type of storage when higher IOPS is needed .
Throughput
Throughput measures the ability for a storage to handle read and write operation of large sequential data files and it is measured in Megabytes per Seconds (MB/s).
If your application for example is serving video files online, your need will be redirected toward a HDD base EBS Volumes since they are optimize to work very will with transactions where the dominant performance attribute is Throughput.
Top comments (0)