These days, everything is about cloud. Cloud computing, cloud killing, cloud blocking, cloud storage, cloud email, cloud shearing, clouds ... I do not do the performance test can not lose to them.
The previous introduction of the use of LoadRunner to manipulate MySQL and the use of LoadRunner manipulation memcached, a typical business system is the cache + application + database. Where the application implements the business logic of the close-up. Similarly, doing performance testing is the same:
The test scenarios we designed are implemented through test scripts, which are equivalent to applications in the business system. And the test process is also the data, usually we use the LoadRunner parameter function to save the data. Unfortunately, LoadRunner's support for large file reads is bad, and the default is simply to read the parameter file. Writing or modifying a file does not provide a ready-made method and requires a self implementation.
Writing a file is not as good as writing a database, which is why there is a database.
In order to improve the performance of the system, people often add caching on the front end.
Do the same with performance testing, we need data, and test scripts are not the bottleneck of testing.
Now that you have been able to manipulate the database and cache directly using LoadRunner, how do you build a team that belongs to your own test cloud?
The principle is actually very simple, refer to the principle of memcached cluster. When accessing data, hash and modulo a key value. This determines where the data is stored (database cluster members) and where the data is obtained (caching cluster members). Figure diagram: