Server Memory

Choosing the right server memory can boost the performance and reliability of your data center by checking for data errors and reducing electrical loads. There are different types of server RAM to prevent crashes, correct inaccuracies in your data or allow for faster operation.

ECC Memory Uses Algorithms to Constantly Check Data for Errors

ECC server memory can prevent system failure and data corruption, which is particularly important in data centers and business environments, such as healthcare and finance. ECC memory adds a ""parity bit"" to each binary code to detect errors. When writing data, this type of system-specific memory creates and stores an ECC code. When it reads data, ECC memory creates a new ECC code and compares it to the one it had previously stored. If there are errors, the memory decrypts the corrupted data and corrects it immediately. ECC computer memory uses algorithms to continually check for single-bit errors as the server processes data. This type of server memory can be slower than others, but it reduces system crashes. It also tends to be more expensive. The other components of the server, such as the processor and motherboard, must support ECC memory.

Registered Memory Puts Less Strain on CPU Memory Controller

There are two types of ECC server memory: unbuffered and registered. When using unbuffered memory, the CPU communicates directly with the data banks for faster performance. Registered memory has registers on each memory bank module. The CPU needs to communicate with the registers first, then the registers communicate with the DRAM. This process means that every instruction from the server requires at least one extra CPU cycle, resulting in slower performance. However, registered server memory puts less strain on the CPU's memory controller, and increases reliability, a useful choice in data centers and business settings. Registered memory works with the DDR, DDR2 and DDR3 modules of many server models.

Load-reduced Dual Inline Memory Modules Reduce Electrical Loading on the Memory Bus

A load-reduced dual inline memory module (LRDIMM) is a type of server memory that reduces electrical loading on the memory bus. LRDIMMs have a memory buffer chip for each module that goes into the parallel memory bus. This way, they do not connect directly to the bus like other types of memory. When a server only has LRDIMMs as dynamic RAM, the memory controllers shift from parallel to serial mode, so the memory buffer handles all the read and write operations. Thanks to the reduced electrical load, LRDIMMs typically work faster than registered memory.

Dynamic RAM Is Common as Main Memory

When considering desktop memory, there are two types of server RAM: static and dynamic. Static RAM (SRAM) has a more complex construction, with latches and transistors. Each block of memory requires six transistors. SRAM consumes less power than dynamic RAM (DRAM) and it is faster, but it is only available in smaller storage sizes. Since SRAM has a short access time, it is most common as cache memory. DRAM has a larger access time, so it is more common as the main memory. It is generally more expensive than SRAM. These factors also apply to laptop memory.

SHOP BY Speed

ECC Memory Uses Algorithms to Constantly Check Data for Errors

Registered Memory Puts Less Strain on CPU Memory Controller

Load-reduced Dual Inline Memory Modules Reduce Electrical Loading on the Memory Bus

Dynamic RAM Is Common as Main Memory