Looking to upgrade a server? then head this way to Extreme Overclocking for their new guide called Server Power & Performance: A Buyers Guide To Upgrading Your System! Here's a byte.
Servers these days perform such a variety of tasks that it is impossible to give a "one size fits all" solution on how to make it run better. However, one thing that most servers have in common is the programs (or services, or daemons, whatever you prefer to call them) like to cache large amounts of data to maximize performance. For instance, a web page that gets requested often seems common sense to cache it in memory where it can be accessed quickly instead of reading the file from the hard drive each and every time it is requested. Obviously data stored in memory can be accessed many times faster than even the fastest hard drives today.
When a server runs of out allocated memory to cache data, it is forced to start flushing (or emptying) the cache to make room for the newer data that is being read from the hard drive. If the cache settings are too small, then a performance hit is taken from the hard drive having to access commonly used data over and over. So to try and increase performance, the cache settings can be generally be increased via the program's configuration settings. However, if set too large, the system could run out of physical memory and be forced to start swapping to "virtual memory" (sometimes known as a page file or swap file). Virtual memory is really space set aside on the hard drive for RAM to write data to when it becomes full. This second scenario is an even worse situation than the first.
Server Power & Performance: A Buyers Guide To Upgrading Your System
Servers these days perform such a variety of tasks that it is impossible to give a "one size fits all" solution on how to make it run better. However, one thing that most servers have in common is the programs (or services, or daemons, whatever you prefer to call them) like to cache large amounts of data to maximize performance. For instance, a web page that gets requested often seems common sense to cache it in memory where it can be accessed quickly instead of reading the file from the hard drive each and every time it is requested. Obviously data stored in memory can be accessed many times faster than even the fastest hard drives today.
When a server runs of out allocated memory to cache data, it is forced to start flushing (or emptying) the cache to make room for the newer data that is being read from the hard drive. If the cache settings are too small, then a performance hit is taken from the hard drive having to access commonly used data over and over. So to try and increase performance, the cache settings can be generally be increased via the program's configuration settings. However, if set too large, the system could run out of physical memory and be forced to start swapping to "virtual memory" (sometimes known as a page file or swap file). Virtual memory is really space set aside on the hard drive for RAM to write data to when it becomes full. This second scenario is an even worse situation than the first.
Server Power & Performance: A Buyers Guide To Upgrading Your System
