Picking the Right Memory Configurations
There are many kinds of computer memory. For our purposes, let’s focus on the most important basic attributes required to select the right memory configuration for a specific workload. These attributes include size of memory, speed of memory, and type and amount of memory cache.
Memory cache provides temporary storage for frequently used instructions and data for quicker processing by the processor. The cache is an extension of a compute platform’s main memory. Its use is a given in most cases. You might look for more cache to speed up processor requests that are more routine, which can be held in cache and thus are not required to be reloaded at the cost of performance. However, you can allocate too much cache that won’t be used, or not enough, and that will also impact processor performance.
The speed of the memory is just that. It’s the amount of time it takes memory to receive a request from the processor and then read or write data. RAM speed or frequency is measured in megahertz (MHz), and that’s often how cloud providers will tell you about the speed of the memory for your compute configuration.
Finally, there’s the type and size of the memory, or how much memory will be available to the application. Typically, memory size is measured in gigabytes (GBs). If, for some reason, you don’t allocate enough memory, your application won’t just stop processing. Instead, the operating system will begin to write to disk storage instead of writing to memory. Your application performance will suffer greatly when this occurs. Memory size is also a tunable parameter in the operating system, as to when and how the operating system leverages storage instead of actual memory.