The type of storage deployed into the data center directly affects both the overall potential availability of applications, and the organization's ability to acquire sufficient storage to meet capacity needs. Three of the options available when it comes to choosing hard drives for a server or storage array are:
Serial ATA (SATA) disks, the capacity workhorses of the IT industry. They are available in capacities of up to 4 TB, but are at the bottom of the food chain when it comes to performance and overall reliability.
Near Line SAS (NL-SAS) disks, in which the actual disk platters are enterprise grade SATA platters, and the connector is of the SAS variety. This combination allows customers to enjoy SATA's capacity benefits while also enjoying some advanced features afforded by the SAS connector.
Serial Attached SCSI (SAS) disks, a staple in the data center when it comes to enterprise level storage. Compared to SATA and NL-SAS disks, SAS disks are both faster and an order of magnitude more reliable, but they lack the capacity of SATA units.
4. What is solid-state storage and why is it important?
Solid state is the fourth hard drive option for a server or storage array. Expensive solid-state storage has been available for quite some time. Solid-state disks are less expensive today, but remain the most expensive type of storage when it comes to capacity. They are much more affordable when performance is the primary metric. Because they have no moving parts, solid-state disks sport low latency and very good random access performance. They are reliable, require less power than rotating disks, and experience no physical wear and tear or mechanical breakdowns, but they are smaller than hard disks, and are only good for a finite number of erase/rewrite cycles before they are effectively dead.
5. What different RAID levels are available and what is the impact of each?
RAID can stand for Redundant Array of Inexpensive Disks, or Redundant Array of Independent Disks, depending on the source. Either way, RAID, although an aging technology, continues to play a vital role in the data center. Available RAID levels include levels 0, 1, 5, 6, 10, and 50. With the exception of RAID 0, the purpose of each level is to protect your data from hardware failure. RAID 0, which sometimes is used to boost overall storage performance, affords no data protection. For virtualization administrators who must rely on RAID technology for underlying storage, the choice of RAID level carries long-term ramifications, so it's important to thoroughly understand the pros and cons of each.
6. How do enterprise level storage features play important roles in virtualization?
Enterprise-class storage features can have a significant impact on how the environment operates, and require careful consideration. Three of these features include:
Thin provisioning, which allows administrators to overprovision virtual server storage capacity without actually wasting any storage space. Downsides include physical disk space constraints and a potential performance penalty.
Deduplication, which provides significant benefits to organizations that want to maximize the return on investment for their storage environments. The deduplication engine attempts to match patterns in blocks and files with existing data. When a match is found, the deduplication engine simply writes a pointer to the original copy of the data instead of creating a duplicate, so storage capacity need is reduced.