A surge of web requests is common phenomena for a web server that handles popular contents. If the server has limited resources (I/O bandwidth, memory, CPU utilization, network bandwidth, etc.), the server may suffer a scalability problem. Optimization on the application side is the only option. This paper compares two techniques dealing on access scalability of a web service that handles popular contents, such as student admission results. This paper explains method how the PHP-FPM worker handle a lookup query between data array stored in a single file and data array split into multiple files. The empirical test results on a web server with 2 GB of RAM and 2 CPU cores show that the lookup process on multiple files can handle more than 63.3% requests, compares to 8.41% requests on a single file. The lookup query of data array split into multiple files outperforms the lookup query of data array stored in a single file.
展开▼
