How do I test the performance of an ECS?

First of all, you need to install HD Tune on your server, which is a free hard disk detection software abroad, the default is the English interface, but you can also find the Chinese version of HD Tune software by searching on the Internet.

Use HDTun to perform a read performance test on the drive. In the "Benchmarks" tab of the software window, click the drop-down list in the upper left corner to select the hard drive you want to perform the performance test, the software defaults to the "Read" test, and then click "Start" on the right, as shown in the following figure:

After the ECS is activated, the default disk is not partitioned, and the write rate must be tested by the software before starting the test.

2. Network uplink and downlink speed test.

Opening instructions: Please use IE browser to test the speed, it is best not to use the window of the world and Maxthon, there may be an error) You can see the interface shown in the figure below, click the "begin test" button to test.

End 3. Server stability test.

Here are a few things to keep in mind during server testing:

1.Whether the server is down frequently.

2.The number of times the server went down during the duration of the test.

3.How long does it take for the server to recover after an outage.

4.Whether there is severe data loss after downtime.

5.Whether it has been attacked by an intrusion attack.

6.Whether the application is abnormal.

Fourth, network stability test.

Delay. 2.Packet loss rate.

Whether the upload is frequently interrupted.

Whether the server's ** is turned on properly.

5.Check whether the server is remote.

End 5. Software compatibility test.

1.Whether the software installation is normal.

2.Whether the software is compatible with the operating system.

3.Whether the software is running abnormally and often reports errors.

4.Whether the software data is compatible. Hope.

Generally, you don't need to test, just look at your server configuration.

After selecting an ECS vendor, you can compare the technical specifications of specific ECS products to see whether they meet the existing and future business needs. Because cloud computing has requirements for the structural flexibility, computing density, and storage capacity of the cloud server itself. So choose carefully.

However, Tianhu data is a good choice.

If you use LoadRunner to perform stress testing on Alibaba Cloud server deployment projects, then you must first ensure that the internal network test, that is, the LR is on your own machine, and the load machine must be added to another Alibaba Cloud server Win or Linux, or you can purchase another Alibaba Cloud server for Windows to install the LR load machine and add it to the local computer.

2 In this way, you don't have to consider the impact of network bandwidth (public network) when you test, and then you can use LR to record the script, and the recorded script is debugged (useless ** delete, add things, assembly points, parameterization, etc.).

3. After the script is successfully debugged, it is time to set up the scene, set your test scenario and test content according to the actual needs of the project, and pay attention to the clear purpose of the test, such as: find out the system bottleneck. For example, look at the maximum amount of concurrency that the system can withstand.

The 4 scenarios are all finished, the data collection is completed, and the most difficult part is the analysis of the data collection results. I won't say more, I won't be able to say it deeply! It's all made clear!

This is generally tested with a test tool software.

You go to "Haidilao" for dinner at noon.

We can think of the restaurant "Haidilao" as a system under test.

When you go to dinner, you make a request to the system under test, which causes a certain load on the system. The more people you bring with you, then the busier this restaurant will be, and arguably the greater the load on the restaurant.

You start ordering. At this time, the person at the table next to you also starts ordering. Then the two of you have concurrent requests to this system.

At the same time, some of the other tables are eating and some are waiting for food, and these are all concurrent transactions. A complete meal transaction can be defined as four steps: ordering, ordering, serving, and paying.

For a CS system, it can correspond to: establish a connection, send a request, accept a reply, disconnect.

An important factor that affects the quality of a restaurant's business is the speed of serving. The speed of serving is reflected in two aspects:

Many factors will affect the speed of serving, such as the number of waiters and the number of chefs. For a CS system, the waiter is equivalent to the access layer, and the chef is equivalent to the back-end service. If there are too few waiters, orders are very slow, and the chefs in the back are idle, then the serving speed will not be fast; If there are enough waiters and the order is fast enough, but there are too few chefs to place the order in time, the service speed is also very slow; If there are a lot of waiters and a lot of cooks, but there are few customers who come, then most of the waiters and cooks are idle and all the resources are wasted.

Therefore, the number of access layer and backend service processes, as well as the resource ratio, need to be optimized according to the actual situation.

The number of customers who come is beyond the control of the restaurant itself, but the speed of serving food and the number of seats in the restaurant will restrict the flow of customers. There must be a peak flow of guests, and when the number of guests exceeds this peak, then those guests will wait for a table, or the serving speed will be too slow for the guests to tolerate. The capacity test is to simulate the transaction of enough customers to eat through the tool, hoping to find such a customer flow to generate a certain load on the restaurant, at this time, the restaurant can not only receive the most customers but also ensure the shortest waiting time.

The number of customers who follow up is also related to the reception capacity of the restaurant. The more customers you come to dine with, the more likely you are to be complained about and the more likely you are to be served the wrong dish.

1.The processing time of a customer request, the waiting time between the time the order is placed and the time it is served, is called the response time.

2.The frequency with which this restaurant serves food to multiple customers at the same time is what we call throughput.

What needs to be tested.

Internet speed: You can check the speed with the speed test tool.

Hardware: Master Lu can be installed to stress test the hardware.

Do you need a cloud host, I can test it for free on my side, mainly to see the stability of the network.

It's generally a fool's errand, how is your internet speed.