All performance charts
Partners
The Games selection
violent :
Friday the 24th
Exterminate Santa Claus's elves. Use the arrows to move, S to grab the elves, and A or W to attack them with your sword.
|
violent :
Interactive Buddy
Unwind on your interactive buddy: Do anything you want to him, it will earn you money, and you can buy other stuff to torture him with.
|
Sponsored links
Table of contents
- 1 – The RAID 6 Storage Bodyguards
- 2 – How Redundant Arrays Work
- 3 – Performance Of Redundant Arrays
- 4 – Adaptec 2820SA W/ Advanced Data Protection Suite
- 5 – Array Creation
- 6 – Areca ARC-1220
- 7 – Promise SuperTrak EX8350
- 8 – Array Creation
- 9 – Test Setup
- 10 – Benchmark Results
- 11 – Benchmark Results
- 12 – Conclusion: Areca

Redundant Array of Inexpensive/Independent Drives (RAID) is the general solution for requirements in which performance, storage capacity or data safety of a single drive isn't sufficient. While this usually is the case in server systems, workstations and high performance desktops sometimes take advantage of RAID setups as well. Since two or more drives must be used, a RAID array can be somewhat expensive, and this cost can be inflated further depending on your controller choice.
Our recent roundup of eight-channel SATA II RAID controllers revealed that most manufacturers offer great value. However, only one of these cards actually offered a RAID mode that can satisfy data protection requirements that exceed the capabilities of a RAID 5 setup. Although RAID 5 allows one array drive to fail without leading to data loss, this means that as soon as a drive does fail, you are without protection. Your stored data will be at risk until the array is rebuilt, and a second drive could potentially fail any moment.
Almost any enthusiast class motherboard offers basic Serial ATA RAID capabilities today, with RAID 0 and 1 the most popular options. RAID 0, however, does not offer any data redundancy - its only purpose is increasing throughput, but this also means a high risk of losing data: one dead drive will cause the whole array to be lost. RAID 1 does the opposite, by mirroring one drive's information onto another one in real time, so either drive can fail without data loss.
In business environments, RAID 1 or 0+1 (a combination of both RAID 0 and 1 using four drives) is considered the basic setup. Since the data safety of mirrored array configurations comes hand in hand with a halved total storage capacity, RAID 3 or RAID 5 is favored for larger drive arrays. Both generate parity data from all of the drives in the array but one and store it on the remaining drive. RAID 3 has almost disappeared because this parity data is stored onto one dedicated drive, which creates an unwanted bottleneck. In contrast, RAID 5 distributes the parity information across all drives; this means that the more drives you add, the better a RAID array will perform.
RAID 6 is based on the same concept as RAID 5, but it adds another layer of security by generating parity data across multiple data rows. Recalculating this requires complex matrix and reverse matrix algebra and can keep even the latest hardware pretty busy.
- Build a Cheap and Fast RAID 5 NAS [TomsNetworking]
- Reccomendations for Xtreme PC [CPU & Components]
- Vista Problem: Control Panel [Windows Vista]
- Opinions for new rig (yet another one) [CPU & Components]
- Server motherboard: Z8NR-D12 or Z8PE-D12? [Motherboards & Memory]
Questions? Ask Tom's community!
Sponsored links
Related forums topics
- I converted my Antec 900 fans from Molex to 3-pin
- Installing a Floppy Drive Still Makes Sense?
- AI NOS
- rediculously bad drive speeds with onboard raid
- Installing drivers via floppy vs internet
- RAID0 and SSD. Match made in heaven?
- RAID + Partitions
- RAID 0. Risky?
- Raid 1 + A8N SLI Premium + New Build.. Help..
- 12 drive dual RAID-5 config need advice
