Shulou(Shulou.com)06/02 Report--

Have you ever used Amazon EFS? Its main purpose is to enable you to build cloud-based network file systems for Linux-based application loads and programs. Because the structure and function of EFS are similar to the local infrastructure, you can easily transfer a variety of existing files. However, before you plan to move any application load to EFS, give EFS a "test drive" and perform regular performance optimizations after moving in.

Of course, partners with practical project experience know that in most cases, performance is not an exact science, and it will involve the real environment and usage status of the application. Therefore, when optimizing performance, we are often unable to achieve it overnight, and we need to continue, or even anyway, to set up and try automation or manual work.

In this article, you will learn the basic concepts of AWS EFS, how to configure and try it, and how to keep the application load of EFS running smoothly and efficiently through seven important aspects.

What is AWS EFS?

Elastic File system (Elastic File System,EFS) is a storage service in AWS that can be used to recreate network file systems in the cloud. EFS based on network file system v4 (NFSv4) is mainly designed for Linux-based application loads and programs. You can smoothly integrate it with other AWS services, as well as local resources.

Compared with other storage services, AWS EFS can provide the function closest to the local existing file storage (see-- https://cloud.netapp.com/blog/ebs-efs-amazons3-best-cloud-storage-system for details). With it, you can easily move existing local file structures to the cloud and enjoy the same smooth operation experience as accessing local files. You can do this either by hosting the application in EC2 (Elastic Compute Cloud) and then attaching the instance to EFS, or you can use EFS as a stand-alone file system.

In EFS, you can choose the following two access methods:

Standard access (Standard Access) is designed for your basic application load. It is characterized by higher cost in exchange for lower access latency.

Infrequent access (Infrequent Access) is designed for files that need to be stored for a long time but are rarely used. It is characterized by lower cost in exchange for higher access latency.

Of course, no matter which access method you choose, EFS is charged according to use. In other words, it only charges for the storage and transmission bandwidth you actually use.

How to try EFS?

You often need to try out the service before transferring files to EFS. Based on this, you can not only familiarize yourself with its relevant configuration and evaluate its running performance, but also estimate the level of resource allocation required when using EFS by comparing it with the current local system. Next, let's start with the "free package (free tier)" offered by EFS.

Free EFS tier provides 5 GB storage space that can be used for 12 months. If you already have an instance in EC2 that you can mount the file system to, it is not difficult to try out EFS. Of course, if you don't like the package, you can also choose a free tier for EC2, which provides 750 hours of use, enough for you to take your time to complete the tests.

To set up your own demo, you can create various resources through the command line interface (CLI) or the administrative console (Management Console). We won't go into detail here on how to use the administrative console to set up and connect the corresponding services. If you are interested, you can refer to the relevant documentation of AWS by linking to https://docs.aws.amazon.com/efs/latest/ug/getting-started.html.

It is worth mentioning that during the trial, please simulate the various usage operations of the target file system as realistically as possible. For example, you can add multiple instances, use both single-threaded and multithreaded methods to write files, and use standard and infrequent access methods to access all kinds of data.

Seven important hints about AWS EFS performance

Once you have decided to adopt EFS services, you need to shift your focus to ensuring maximum return on investment. Here are seven important tips I've summarized for you about optimizing and improving AWS EFS performance. And you don't have to pay extra for resources you don't need.

1. Monitor your EFS burst credit

Burst credits is one of the most neglected aspects of EFS. These credits are designed to temporarily improve your application performance when traffic is high. However, they only last a short time and take a while to recover. Therefore, this means that if the credits is used up, the performance of the application may degrade sharply.

Each EFS volume usually starts at a transfer rate of 0.5 MB/s. Burst credits can supplement and boost throughput to 100 MB/s for 7.2min. The only way to increase such metrics is to increase the amount of data stored in the EFS volume. In other words, you can continue to achieve good performance when the amount of data can reach up to 10 GB

two。 Use asynchronous write operations

Asynchronous write operations allow you to reduce the latency caused by write operations by subtracting part of the path of data transfer. When asynchronous writes are enabled, the data is buffered to an instance of EC2 before being written to EFS.

It is worth noting, however, that this operation may affect data consistency. Therefore, it is best not to use this method if you require the file system to be synchronized in real time.

3. Do not run applications from EFS

If you try to run the application directly in EFS, you may encounter poor performance. After all, EFS is not designed to run or host applications. It cannot perform reads of large files, nor can it provide the required speed for managing the code base or deploying applications.

Instead, you should host or run the application from EBS (Elastic Block Store) and EC2 (see-- https://dzone.com/articles/confused-by-aws-storage-options-s3-ebs-amp-efs-explained) and re-store the EFS for the attachment file. For example, you can allow EFS to provide only the content and limit its corresponding services to EC2 to ensure that the application achieves the expected performance.

4. Monitor average IPUP O throughput

Tracking the average IPUP O throughput can help us identify bottlenecks in existing applications and then try to compress the storage area. For example, if an application needs to use multiple files, we can combine these files into one, thereby reducing the need for throughput. Conversely, operations that frequently request multiple small files are likely to quickly reach and exceed predetermined throughput limits in a relatively short period of time.

It can be seen that by reducing the number of possible file requests, we can reduce the number of Icano operations per second. This not only helps to reduce latency overhead, but also maintains a certain performance range.

5. Carefully manage backups

Since there is no snapshot feature available, you may encounter difficulties if you want to create a backup in EFS. Typically, you must use the AWS Backup service or execute an EFS-to-EFS to make the backup. However, either way, you need to keep copies of various backups and closely monitor the storage space required for backups.

In addition, when you create a backup, it is best to schedule it during non-working hours and create it at a speed-limited replication. This can not only avoid affecting the performance of various standard operations, but also prevent the occurrence of unexpected burst credits.

6. Choose the right performance mode

When using EFS, you can choose two performance modes:

General intent (General Purpose)-provides lower throughput in exchange for lower latency operations. It is designed for standardized tasks, such as web hosts or content management. Also, this is the default performance setting for EFS.

Maximum Max O (iPink O)-provides higher throughput in exchange for higher latency. It is designed to handle tedious computing tasks, such as media processing or big data analysis (see-- https://dzone.com/articles/a-beginners-guide-to-big-data).

7. Adopt concurrent connection

EFS is designed to allow large-scale concurrent connections and can attach up to thousands of EC2 instances at the same time. Accordingly, you can significantly improve overall performance by sharing different applications and loads. Of course, when you configure these connections, you are still limited by the total throughput of each volume, as well as the transfer speed.

Summary

Amazon EFS is a powerful tool, but like many other self-configuring (self-configured) services, it still requires you to optimize it on a regular basis. To avoid unnecessary overhead, you should monitor EFS's burst credits and average Icano throughput on a regular basis. At the same time, to ensure optimal performance, you can use asynchronous write operations while avoiding running applications directly from the EFS. In addition, in order to avoid additional billing, you not only need to manage various backups, but also choose the correct performance model for the application load. Finally, you can also take advantage of concurrent connections to improve overall performance.

Original address: https://www.linuxprobe.com/notice-aws-efs-performence.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.