Panasas Invents ‘Tiered Parity’

By Michael Feldman

October 12, 2007

In 1988 Garth Gibson at the University of California, Berkeley, co-authored a paper titled “A Case for Redundant Arrays of Inexpensive Disks (RAID) [PDF],” which outlined the basic principles of using big, cheap disks to increase data reliability and I/O performance. RAID went on to become a widely adopted storage technology throughout the industry, while Gibson co-founded Panasas Inc., a storage cluster vendor for high performance computing applications.

This week, Gibson and company claim that they have implemented the most significant extension to disk array data reliability since the original RAID paradigm was developed. Their new architecture is called “tiered parity.” In this model, Panasas has built “vertical parity” and “network parity” on top of their existing RAID 5 “horizontal parity” implementation.

The RAID 5 approach, as it was outlined in the original paper, consists of striping data and parity across multiple disks. It enables error recovery for single disk failures and increases performance via parallel reads and writes. This technology is widely used in storage systems today. Panasas’ own implementation of RAID 5, called “ObjectRAID,” is based on storage objects rather than blocks. The added intelligence is designed to reduce reconstruction times when a disk failure occurs.

But no RAID 5 technology can handle a media error, also know as an unrecoverable read error (URE), if it occurs during reconstruction of a failed disk. When this occurs, the RAID data cannot be rebuilt from disk; a backup (usually on tape) has to be used to recover the entire array. Ten years ago, this wasn’t a serious problem. With 50 GB SATA disk drives, a media error was very unlikely to occur while reading a single disk, since the rate of failure is about one error every 10^14 bits (12.5 terabytes), a rate that has remained constant for over a decade. And when a media error did happen to occur during reconstruction, a 50 GB disk took only a few hours to recover from tape.

Times have changed. Disks have become much bigger and denser. Capacities of 500 to 750 GB are common today, and one terabyte disks will soon be the norm. That means when a disk goes south, the odds of hitting a media error during recovery are much greater, and recovery from tape can take days or weeks.

Imagine a RAID array of seven 1 TB disks. When one disk fails, the chances of hitting a URE while recovering the data from the six remaining disks is now about 50/50. When two terabyte disks hit the market in 2009, the disk failure plus media error scenario becomes almost a sure bet. Recovering the storage array from backup tape could take a month. For high end computing applications that use tens or hundreds of terabytes of data, this would be a disaster.

“I think what people are becoming aware of is that the data integrity provided by RAID 5 is basically no longer sufficient,” says Robin Harris, senior analyst at Data Mobility Group. “RAID 5 will only protect across a single disk failure, so it’s going away as a [standalone] data protection strategy.”

To address this problem, Panasas invented vertical parity. Essentially, they’ve added RAID within each disk, by generating a parity sector from the other sectors. The local parity sector can be used to recompute the missing data in case of a media error. According to Panasas, vertical parity gets the error rate down to between one in 10^18 and one in 10^19, which is 1000 to 10,000 times better than the URE rate. The extra parity information uses 10 percent of the disk capacity, but Panasas claims there is no performance hit. So scalability is built in.

A word here should be said about RAID 6 technology (also known as double parity), which some vendors use for an additional level of data protection. This scheme was designed to guard against a double disk failure, which it does. Sort of. The problem is that RAID 6 doesn’t protect against subsequent media errors after the second disk goes down, which, as discussed above, is becoming increasingly more likely. Here, it has the same problem as RAID 5. However, RAID 6 can be used to recover from the single disk failure plus media error scenario. But the performance hit for dual parity compared to single parity is significant. So it’s a mixed bag and doesn’t directly address the media error problem.

On top of its horizonal and vertical parity schemes, Panasas has added an additional layer of network parity protection. At this level, parity checking is done on the client side, to make sure the data delivered by the storage system wasn’t corrupted on its way to the user. Because of increasing I/O bandwidth and the number of hardware and software components between the external data and the application, there are increasing opportunities for good data to go bad. Firmware, server hardware, server software, network components and transmission media can all potentially mangle valid data unbeknownst to the application. With network parity, the client receives an error notification when bad data is detected.

The tiered parity technology will be included in the next version of Panasas’ ActiveScale operating environment, version 3.2. The beta will be out next month and will be generally available by the end of the year. The additional parity levels can be turned off if the user believes they’re not needed for a particular environment. According to Panasas, the tiered parity technology doesn’t exact a performance hit on top of the existing RAID 5 implementation, but, as stated above, the vertical scheme does eat an additional 10 percent of the storage — that’s in addition to the 10 percent used by the RAID 5 implementation.

Although the overall concepts of the three-tiered architecture are fairly general, Panasas is attempting to protect its new invention. “We actually have a patent pending on this tiered parity concept, particularly the vertical parity,” says Larry Jones, VP of Marketing at Panasas. “Could someone copy it? Who knows? But we are trying to protect this specific idea.”

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

Data Vortex Users Contemplate the Future of Supercomputing

October 19, 2017

Last month (Sept. 11-12), HPC networking company Data Vortex held its inaugural users group at Pacific Northwest National Laboratory (PNNL) bringing together about 30 participants from industry, government and academia t Read more…

By Tiffany Trader

AI Self-Training Goes Forward at Google DeepMind

October 19, 2017

DeepMind, Google’s AI research organization, announced today in a blog that AlphaGo Zero, the latest evolution of AlphaGo (the first computer program to defeat a Go world champion) trained itself within three days to play Go at a superhuman level (i.e., better than any human) – and to beat the old version of AlphaGo – without leveraging human expertise, data or training. Read more…

By Doug Black

Researchers Scale COSMO Climate Code to 4888 GPUs on Piz Daint

October 17, 2017

Effective global climate simulation, sorely needed to anticipate and cope with global warming, has long been computationally challenging. Two of the major obstacles are the needed resolution and prolonged time to compute Read more…

By John Russell

HPE Extreme Performance Solutions

Transforming Genomic Analytics with HPC-Accelerated Insights

Advancements in the field of genomics are revolutionizing our understanding of human biology, rapidly accelerating the discovery and treatment of genetic diseases, and dramatically improving human health. Read more…

Student Cluster Competition Coverage New Home

October 16, 2017

Hello computer sports fans! This is the first of many (many!) articles covering the world-wide phenomenon of Student Cluster Competitions. Finally, the Student Cluster Competition coverage has come to its natural home: H Read more…

By Dan Olds

Data Vortex Users Contemplate the Future of Supercomputing

October 19, 2017

Last month (Sept. 11-12), HPC networking company Data Vortex held its inaugural users group at Pacific Northwest National Laboratory (PNNL) bringing together ab Read more…

By Tiffany Trader

AI Self-Training Goes Forward at Google DeepMind

October 19, 2017

DeepMind, Google’s AI research organization, announced today in a blog that AlphaGo Zero, the latest evolution of AlphaGo (the first computer program to defeat a Go world champion) trained itself within three days to play Go at a superhuman level (i.e., better than any human) – and to beat the old version of AlphaGo – without leveraging human expertise, data or training. Read more…

By Doug Black

Student Cluster Competition Coverage New Home

October 16, 2017

Hello computer sports fans! This is the first of many (many!) articles covering the world-wide phenomenon of Student Cluster Competitions. Finally, the Student Read more…

By Dan Olds

Intel Delivers 17-Qubit Quantum Chip to European Research Partner

October 10, 2017

On Tuesday, Intel delivered a 17-qubit superconducting test chip to research partner QuTech, the quantum research institute of Delft University of Technology (TU Delft) in the Netherlands. The announcement marks a major milestone in the 10-year, $50-million collaborative relationship with TU Delft and TNO, the Dutch Organization for Applied Research, to accelerate advancements in quantum computing. Read more…

By Tiffany Trader

Fujitsu Tapped to Build 37-Petaflops ABCI System for AIST

October 10, 2017

Fujitsu announced today it will build the long-planned AI Bridging Cloud Infrastructure (ABCI) which is set to become the fastest supercomputer system in Japan Read more…

By John Russell

HPC Chips – A Veritable Smorgasbord?

October 10, 2017

For the first time since AMD's ill-fated launch of Bulldozer the answer to the question, 'Which CPU will be in my next HPC system?' doesn't have to be 'Whichever variety of Intel Xeon E5 they are selling when we procure'. Read more…

By Dairsie Latimer

Delays, Smoke, Records & Markets – A Candid Conversation with Cray CEO Peter Ungaro

October 5, 2017

Earlier this month, Tom Tabor, publisher of HPCwire and I had a very personal conversation with Cray CEO Peter Ungaro. Cray has been on something of a Cinderell Read more…

By Tiffany Trader & Tom Tabor

Intel Debuts Programmable Acceleration Card

October 5, 2017

With a view toward supporting complex, data-intensive applications, such as AI inference, video streaming analytics, database acceleration and genomics, Intel i Read more…

By Doug Black

Reinders: “AVX-512 May Be a Hidden Gem” in Intel Xeon Scalable Processors

June 29, 2017

Imagine if we could use vector processing on something other than just floating point problems.  Today, GPUs and CPUs work tirelessly to accelerate algorithms Read more…

By James Reinders

NERSC Scales Scientific Deep Learning to 15 Petaflops

August 28, 2017

A collaborative effort between Intel, NERSC and Stanford has delivered the first 15-petaflops deep learning software running on HPC platforms and is, according Read more…

By Rob Farber

Oracle Layoffs Reportedly Hit SPARC and Solaris Hard

September 7, 2017

Oracle’s latest layoffs have many wondering if this is the end of the line for the SPARC processor and Solaris OS development. As reported by multiple sources Read more…

By John Russell

US Coalesces Plans for First Exascale Supercomputer: Aurora in 2021

September 27, 2017

At the Advanced Scientific Computing Advisory Committee (ASCAC) meeting, in Arlington, Va., yesterday (Sept. 26), it was revealed that the "Aurora" supercompute Read more…

By Tiffany Trader

How ‘Knights Mill’ Gets Its Deep Learning Flops

June 22, 2017

Intel, the subject of much speculation regarding the delayed, rewritten or potentially canceled “Aurora” contract (the Argonne Lab part of the CORAL “ Read more…

By Tiffany Trader

Google Releases Deeplearn.js to Further Democratize Machine Learning

August 17, 2017

Spreading the use of machine learning tools is one of the goals of Google’s PAIR (People + AI Research) initiative, which was introduced in early July. Last w Read more…

By John Russell

Nvidia Responds to Google TPU Benchmarking

April 10, 2017

Nvidia highlights strengths of its newest GPU silicon in response to Google's report on the performance and energy advantages of its custom tensor processor. Read more…

By Tiffany Trader

GlobalFoundries Puts Wind in AMD’s Sails with 12nm FinFET

September 24, 2017

From its annual tech conference last week (Sept. 20), where GlobalFoundries welcomed more than 600 semiconductor professionals (reaching the Santa Clara venue Read more…

By Tiffany Trader

Leading Solution Providers

Graphcore Readies Launch of 16nm Colossus-IPU Chip

July 20, 2017

A second $30 million funding round for U.K. AI chip developer Graphcore sets up the company to go to market with its “intelligent processing unit” (IPU) in Read more…

By Tiffany Trader

Amazon Debuts New AMD-based GPU Instances for Graphics Acceleration

September 12, 2017

Last week Amazon Web Services (AWS) streaming service, AppStream 2.0, introduced a new GPU instance called Graphics Design intended to accelerate graphics. The Read more…

By John Russell

EU Funds 20 Million Euro ARM+FPGA Exascale Project

September 7, 2017

At the Barcelona Supercomputer Centre on Wednesday (Sept. 6), 16 partners gathered to launch the EuroEXA project, which invests €20 million over three-and-a-half years into exascale-focused research and development. Led by the Horizon 2020 program, EuroEXA picks up the banner of a triad of partner projects — ExaNeSt, EcoScale and ExaNoDe — building on their work... Read more…

By Tiffany Trader

Delays, Smoke, Records & Markets – A Candid Conversation with Cray CEO Peter Ungaro

October 5, 2017

Earlier this month, Tom Tabor, publisher of HPCwire and I had a very personal conversation with Cray CEO Peter Ungaro. Cray has been on something of a Cinderell Read more…

By Tiffany Trader & Tom Tabor

Cray Moves to Acquire the Seagate ClusterStor Line

July 28, 2017

This week Cray announced that it is picking up Seagate's ClusterStor HPC storage array business for an undisclosed sum. "In short we're effectively transitioning the bulk of the ClusterStor product line to Cray," said CEO Peter Ungaro. Read more…

By Tiffany Trader

Intel Launches Software Tools to Ease FPGA Programming

September 5, 2017

Field Programmable Gate Arrays (FPGAs) have a reputation for being difficult to program, requiring expertise in specialty languages, like Verilog or VHDL. Easin Read more…

By Tiffany Trader

IBM Advances Web-based Quantum Programming

September 5, 2017

IBM Research is pairing its Jupyter-based Data Science Experience notebook environment with its cloud-based quantum computer, IBM Q, in hopes of encouraging a new class of entrepreneurial user to solve intractable problems that even exceed the capabilities of the best AI systems. Read more…

By Alex Woodie

HPC Chips – A Veritable Smorgasbord?

October 10, 2017

For the first time since AMD's ill-fated launch of Bulldozer the answer to the question, 'Which CPU will be in my next HPC system?' doesn't have to be 'Whichever variety of Intel Xeon E5 they are selling when we procure'. Read more…

By Dairsie Latimer

  • arrow
  • Click Here for More Headlines
  • arrow
Share This