Why a server can be bad for your SSD
Unlike traditional hard disk drives (HDDs), a Solid State Drive (SSD) is composed of flash memory that stores data. When you erase data from a flash device it doesn’t immediately go away. It’s flagged for deletion, and processes like TRIM and Garbage Collection, will remove this old data to make room for new. In an HDD when you erase something, it is flagged for deletion like in an SSD, but it is simply overwritten when new data when it becomes available; because of how data stored on an SSD this process will not work, and TRIM and Garbage Collection fill this role. Now that we have a basic understanding of how HDDs and SSDs manage data differently, we can begin to understand why installing consumer grade SSDs into anything like a server, workstation, or Network-attached storage (NAS), is not always such a straightforward option, and may not be recommend.
A server or a NAS, will typically be running 24/7 in most situations, and if you understand Garbage Collection, you will know that it can only run when a drive is in an idle, powered state. This means that SSDs typically will not get a break if they’re running in these types of environments. Without Garbage Collection being able to run and do its job correctly, we are left with TRIM to perform this cleanup. The problem with this is that TRIM is not compatible with most server setups that have any sort of RAID array, or NAS devices ( there are exceptions to this rule). Without TRIM or Garbage collection running, an SSD will start to accumulate erased data. Erased data that accumulates on an SSD will not be visible, and you cannot clean the drive up by simply formatting; formatting an already clogged up SSD would actually make any problem worse, since you’re adding an additional layer of erased data on top of what already exists. This accumulation of old data can have drastic effects on the performance of the drive, as well as potentially causing stability issues if the problem is severe enough.
Understanding the differences between Consumer and Enterprise Grade SSDs
We now understand how a server can be bad for an SSD, but not all SSDs are created equal. Let’s see what some of the differences between consumer grade SSDs and their enterprise counterparts are. Crucial sells consumer grade SSDs, this means our drives are designed for standard day to day use, like you would find in a desktop or laptop, other smaller mobile computing devices, and high-performance gaming/work systems. In an enterprise environment like a server, a drive will be subjected to 24/7 operation, often with a combination of high writes and erases. Enterprise drives like the Micron M500DC or S600DC, for example, are designed to handle continuous operation, with high amounts of erases and writes. Unlike a consumer drive which will lose performance drastically when subjected to continuous writes, an enterprise drive will have a steady level of performance over the given period. When we talk about erasing data from an SSD, this also brings up another important difference between the two classes of drives, and that is endurance.
All flash devices, including SSDs, have a limit to the amount of data that can be written to the memory before reliability of that data will be lost, and the memory starts to go into read only mode. Different drives will have different endurance ratings: our BX200 drives are rated at 72 Terabytes (TB) of erase life, an MX200 1000GB drive will be rated 360 TB, and a Micron 800GB M500DC is 2500 TB, so it’s clear that enterprise drives are rated for significantly more endurance life. For most users, endurance is something they shouldn’t worry about; it would take a typical person many years to use up all of the 72 TB write life on a standard drive. A server that is filling a drive several times over every day, could theoretically use up the entire life of a consumer drive in less than a year; and this sort of wear is not covered under warranty. Along with extended wear life, an enterprise drive will typically have more advanced protection for data on the drive.
At a basic level a consumer grade drive will protect only data at rest, while an enterprise drive is going to cover data in motion as well. Data in motion is data that is in the process of being transferred, or on the drive buffer. With a traditional drive if the system were to lose power during a write sequence, any of this data on this buffer would be lost, and this could actually cause some serious data corruption. This link goes to a Micron document which goes into great detail about these data protection differences. For a more general in-depth comparison of Consumer vs Enterprise drives, this Micron document has some great information as well.
A note on SSDs in Workstations
A workstation in most cases is a server. It has a motherboard that for all intents and purposes is a server board, because it uses ECC (Error Checking and Correction) memory. Since we do not list our drives as compatible with servers, they are excluded from workstations for the same reasons. With that being said, most people will not use their Workstation as a server, they will typically use it for heavy office work for applications like: Illustrator, Photoshop, 3D rendering, or other types of visual editing. So long as you aren’t going to use a workstation for around the clock data management like network caching, virtual machine hosting, and things a server would typically manage, then any of our 2.5” SATA SSD offerings should theoretically be fine in the system.
By installing a consumer grade drive in a server, you are not only getting less than ideal performance, you are also risking the reliability of your data, using a device with a shorter wear life, and possibly voiding the manufacture warranty on the drive. For the above reasons it is crucial to install a storage device that is designed for the type of environment you are going to use it in. If you have any questions or concerns about using a Crucial SSD in your system, please contact our support.