RAID Turns Rad

Enterprise Storage Forum content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

Last fall, Panasas, a developer of high-performance clustered storage solutions, unveiled a new “Tiered Parity” architecture that the company and some analysts said dramatically increases data reliability.

In addition to providing a more reliable solution to enterprises with large-scale data storage needs, Tiered Parity may also be heralding the beginning of the end to solutions based on “traditional,” or single parity, RAID, and even those based on RAID 6, known as double parity RAID.

Classic RAID vs. New RAID

Tiered Parity renders old, non-Panasas RAID implementations obsolete, said Garth Gibson, the computer scientist and co-founder and CTO of Panasas, who pioneered RAID technology back in the late 1980s. That’s because the original RAID technology mainly addressed the loss of complete disks, not media failures (also known as unrecoverable read errors or media faults) on one or more disks (see Gibson Discusses Learning from Storage Failures).

As Gibson explained, “RAID was a strategy for striping data over multiple fault domains and using inexpensive error-correcting codes across those fault domains so that we could recover from common storage failures, in particular the catastrophic loss of complete disks.

“In that era, the fault domain was a physical disk, and we striped the data logically one bit at a time across a fixed number of disks. What’s happened since then is that we’ve realized that the simple striping one bit at a time across a fixed number of disks for all files in exactly the same pattern across storage is correct, but it is overspecialized.”

It also doesn’t address the problem of media failures, which grow proportionately with a disk’s density. And with disks now 250 times denser than they were 20 years ago, when RAID first came on the scene, that’s a big problem for enterprises and institutions that depend on high-performance computing and the secure storage and retrieval of terabytes of data. According to Panasas research, media failures occur approximately 30 times more often than disk failures, and a single media failure can cause a RAID reconstruction to fail.

Even powerful solutions such as RAID 6, said Gibson, are not up to the task.

“It [RAID 6] does a pretty good job of solving some of these problems, but it doubles the amount of parity and doubles the amount of work you do for small updates for particular classes of workloads,” he said. “And it isn’t as powerful as you’d like it to be because with RAID 6 if there are, in fact, two failed disks, you still have to deal with those media failures, because all RAID 6 really did was fix one failed disk and a media failure.”

Put another way, RAID 6 treats the symptom of the failure, not the root cause — and can be expensive in terms of cost and lost performance.

Tiered Parity, on the hand, “is a set of novel solutions that … improve the resistance to media failures, the silent hardware failures, and decrease the long reconstruction times that commercial commodity arrays are taking today,” said Gibson.

So does this mean that the father of RAID has abandoned his first born? Yes and no. Tiered Parity, which makes use of RAID (specifically Panasas’s ObjectRAID), is really “new and improved RAID,” a fully grown, mature RAID, said Gibson, designed to address today’s enterprise and institutional large-scale storage reliability needs.

So What Exactly Is ‘Tiered Parity’?

Tiered Parity is a combination of three independent tiers, or solutions — Horizontal Parity, Vertical Parity and Network Parity — aimed at detecting and correcting problems before they cause a disk failure.

  • Horizontal Parity uses Panasas’s ObjectRAID to provide protection from disk failure across the entire storage array. It also provides scalable RAID recovery, parallel reconstruction and per-file fault isolation.
  • Vertical Parity, a new technology from Panasas that can also be thought of as RAID within an individual drive, detects and fixes media errors on disks — the leading cause of failed RAID recovery operations — before the failure can interfere with RAID recovery.
  • Network Parity, which provides the industry’s first end-to-end data integrity capability, detects errors introduced outside of an enterprise or institution’s storage system.

According to Panasas, when used together (although sold bundled, tiers can be turned off or enabled, depending on the customer’s needs), Tiered Parity, which is sold under the Panasas ActiveScale label, provides 10,000 times the protection of traditional RAID 5 — with no major performance hit.

Put another way, Tiered Parity is RAID, only better.

Back to Enterprise Storage Forum

Jennifer Schiff
Jennifer Schiff
Jennifer Schiff is a business and technology writer and a contributor to Enterprise Storage Forum. She also runs Schiff & Schiff Communications, a marketing firm focused on helping organizations better interact with their customers, employees, and partners.

Get the Free Newsletter!

Subscribe to Cloud Insider for top news, trends, and analysis.

Latest Articles

15 Software Defined Storage Best Practices

Software Defined Storage (SDS) enables the use of commodity storage hardware. Learn 15 best practices for SDS implementation.

What is Fibre Channel over Ethernet (FCoE)?

Fibre Channel Over Ethernet (FCoE) is the encapsulation and transmission of Fibre Channel (FC) frames over enhanced Ethernet networks, combining the advantages of Ethernet...

9 Types of Computer Memory Defined (With Use Cases)

Computer memory is a term for all of the types of data storage technology that a computer may use. Learn more about the X types of computer memory.