Several ways of optimization, which allows to act on certain steps of the request processing, can be considered : adjustment of certain parameters of the configuration file : processes management, desactivation of useless modules, and other parameters allowing to desactivate the execution of certain steps. acceleration of the processing time of the requests during the response phase when the required documents are PHP files, by using a solution of caching PHP. Concerning the processus management of the processes, the parameters to adjust are: MaxRequestsPerChild : maximum number of requests served by the same process. Beyond the process is killed, and a new one is spawned. If this parameter equals to 0, then each process can serve an unlimited number of requests. MaxClients : maximum number of requests that the server can serve simultaneously. In other words, it is the maximum number of process which can be launched by Apache. MinSpareServers : minimum number of available processes. MaxSpareServers: maximum number of available processes. StartServers : number of processes to launch at Apache startup. But the tests carried out do not allow to measure significant time saving in requests processing (details... ). Actually, these parameters must be adjusted according to the resources of the machine, in particular MaxClients. You must ensure that this number is not too small, which results in long latencies with simultaneous requests because there is not enough processes to serve all the requests, nor too large, long processing times, also with simultaneous requests, because the processes launched on the machine are too numerous. Concerning the desactivation of the modules, we can say that a profit can be brought not as regards processing times, but as regards use of the RAM. Desactivating modules makes it possible to decrease the memory size of each Apache process. Contact - version française
Concerning the processus management of the processes, the parameters to adjust are: MaxRequestsPerChild : maximum number of requests served by the same process. Beyond the process is killed, and a new one is spawned. If this parameter equals to 0, then each process can serve an unlimited number of requests. MaxClients : maximum number of requests that the server can serve simultaneously. In other words, it is the maximum number of process which can be launched by Apache. MinSpareServers : minimum number of available processes. MaxSpareServers: maximum number of available processes. StartServers : number of processes to launch at Apache startup. But the tests carried out do not allow to measure significant time saving in requests processing (details... ). Actually, these parameters must be adjusted according to the resources of the machine, in particular MaxClients. You must ensure that this number is not too small, which results in long latencies with simultaneous requests because there is not enough processes to serve all the requests, nor too large, long processing times, also with simultaneous requests, because the processes launched on the machine are too numerous. Concerning the desactivation of the modules, we can say that a profit can be brought not as regards processing times, but as regards use of the RAM. Desactivating modules makes it possible to decrease the memory size of each Apache process. Contact - version française
