Most IT professionals are familiar with the basic RAID functions. These functions come free with some of today's...
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
operating systems, and include RAID-0 (striping) for enhanced disk I/O performance and RAID1 (mirroring) for enhanced data protection. Even RAID5, which uses parity to provide redundancy, is becoming increasingly familiar to users. On the other hand, "enterprise-class" RAID features that such as "online capacity expansion" and "online RAID level migration", which have been available only with high-end SCSI-based RAID, are less well-known. But the restricted availability of these high-end RAID functions is changing. As often happens with technology over time, high-end features become more widely available as the price of the underlying technology drops.
Enterprise-class RAID features like "N-way mirroring" and "controller spanning" are entering the small-to-medium sized (SMB) market, riding on the coat-tails of low-cost serial ATA storage technology. This month's column explains these less familiar enterprise-class RAID features, starting with a set of features that ensure system uptime and flexibility.
Hot swapping, online capacity expansion and online raid level migration
Several advanced RAID features are commonly used in conjunction with one another. For example, hot swapping, online capacity expansion (OCE) and online RAID level migration (ORLM), when used together, allow major alterations to the storage system without taking the system offline. Hot swapping allows disks to be physically added or removed; OCE allows a data array to be expanded across new disks; and ORLM allows the type of RAID array to be changed to a new type on the fly.
This ability to modify the storage system without interrupting operations is particularly useful to dynamic organizations. As an example, let's look at a small but growing organization with a transaction processing (TP) application. When the organization is new, business-critical data could be protected with a simple RAID-1 array that mirrors the data across two disks. As the business grows and the level of transactions increases, it will need more disk capacity, despite the fact that budgets will still be tight. In this scenario, RAID-5 is appealing since overall data capacity could be doubled by only adding a single, third disk. However, the TP server cannot be offline for any significant amount of time, so hot swapping, OCE and ORLM will be required. Hot swapping allows the third disk to be added; capacity expansion allows the existing RAID-1 array to be expanded to include the third disk; and RAID level migration enables the transformation of the RAID1 array to a RAID5 array. All of this is performed in the background, invisible to each user, without losing a single business transaction.
If our small company continues to expand, transaction processing I/O performance will eventually need a boost and the IT group will have to add and maintain another disk or two. In this case, RAID10 with its fast random reads and writes will be desirable. Again, the admin will have to add at least one disk and transform the array from RAID-5 to RAID 10 without taking the system down. Once again, hot swapping, OCE and ORLM are required to migrate to RAID 10.
Coming soon: Part two of this tip discusses N-way mirroring, drive roaming and more.
About the author: Randy Arnott is a RAID architect at Broadcom specializing in software for embedded, card-based, and host-based RAID solutions. He has over 25 years of experience as both a hardware and software engineer with established storage companies like DEC, EMC, and Adaptec, as well as with two successful startups, Data Kinesis and RAIDCore, recently acquired by Broadcom