Visit additional Tabor Communication Publications
September 17, 2012
Panasas has launched ActiveStor 14, the company's fifth-generation storage applicance aimed at high performance computing. The new offering adds solid state drives (SSDs) to what has been almost exclusively a hard disk-based (HDD) NAS storage line-up. The inclusion of SSDs into the company's flagship offering is further proof that flash memory has become a mainstream storage technology for accelerating HPC workloads.
It's also a recognition that HPC storage is more that just about streaming lots of data from cheap SATA drives. This has been the case for some time, even if the customers themselves were unaware of it. When Panasas surveyed 10 typical HPC sites (across government, finance, academia and manufacturing), it was found that 50 to 70 percent of their files fell into the "small file" category -- defined as less than 64KB.
This was true even for those users whose storage capacity was dominated by very large files, and who believed high-throughput I/O was the crux of their storage needs. "The reality is that almost all customers are dealing with mixed workloads." says Geoffrey Noer, Sr. Director of Product Marketing at Panasas.
The presence of so many small files suggests that directory information retrieval and random I/0 performance is a critical requirement across many HPC sites. At the same time, these users had a concrete need for high levels of streaming performance to feed at least a portion of their applications. The need for both bandwidth and IOPS from Panasas' customer base was central to the inclusion of SSD hardware alongside high capacity SATA drives in ActiveStor 14.
Using flash technology isn't exactly virgin territory for Panasas. In 2009, they came out with ActiveStor 9, an SSD-accelerated storage appliance that was aimed specifically at small file I/O. Since it depopulated disk slots in favor of SSDs though, ActiveStor 9 suffered on the throughput side.
With ActiveStor 14, the company delivers both IOPS and throughput by marrying big SATA disks with the latest SSD technology. The idea is to store all the file metadata on the SSDs as well as all the small file data. The idea is to put the vast majority of the "hot" data into flash, which should greatly increase I/O performance.
There are three ActiveStor 14 models available, each with its own mix of HDD and SSD capacity on the storage blade to serve different application profiles.
1. For large file throughput-oriented applications, each blade houses two 4TB SATA disks, one 120GB SSD and 8GB of cache. This is aimed at energy exploration, government, manufacturing, and academia. List price for 81.2TB of storage is $125K.
2. For more mixed workloads, they've come up with a blade identical to the one above, but with a 300 GB SSD. This configuration is targeted at analytics for biosciences, especially genomics. List price for 83TB is $145K.
3. For truly file heavy, random IOPS applications, they have a blade with two 2TB SATA disks, one 480GB SSD, and 16GB of cache. Panasas calls this one the ActiveStor 14T (for turbo) and it's aimed at financial analytics, like Monte Carlo simulations for arbitrage modeling. Because of the greater ratio of SSD storage to HDD, this is the most expensive model, with a list price of $160K for 44.8TB.
But you get the performance you pay for. A shelf of ActiveStor 14T, with 27 data drives, delivered 20,745 operations per second (SPECsfs2008_nfs.v3) and an overall response time of 1.99 milliseconds. Using two shelves, operations per second doubled to 41,116 and overall response time was cut to 1.39 milliseconds.
It's no surprise Panasas developed ActiveStor configurations designed specifically for financial services and biosciences, since those two verticals have seen the biggest uptick in sales at Panasas over the last year. According to company chief marketing officer Barbara Murphy, they have seen revenue grow by 5X in the financial sector and nearly 2.5X in biosciences since 2011.
Panasas revenue, in general, has been growing at a nice clip over the past 12 months, and actually has been on the rise for four straight years. But commercial sales are accounting for a much greater share now: from about 55 percent in 2011 to over 70 percent in 2012. With public spending on the wane a bit, and businesses investing more heavily in HPC, Panasas intends to put a lot of energy into serving these markets. "We believe that the commercial space is going to adopt high performance compute very aggressively over the next couple of years," Murphy told HPCwire.
Over that same period, Panasas has grown their customer base from 300 to 400. They've has done so by relying more on their OEM partners, like Dell, HP, SGI, and Bull, to expand the revenue base. Today Panasas claims a nice selection of elite customers, including seven of the top 10 public oil & gas firms, NIH, the Beijing Genomics Institute, UCLA, Leicester University, and BNP Paribas.
In ActiveStor 14, they believe they have a platform that can compete very well in the analytics domain, especially against more expensive offerings from EMC Isilon and NetApp. At a price point of $4/GB price point for the 14T model and under $2/GB for the less SSD-rich models, they compare favorably with the $12/GB price of Isilon's S200 and $9/GB for NetApp's FAS6240. The latter delivers slightly better performance (using the SPECsfs2008_nfs.v3 benchmark), but is more than twice as expensive on a capacity basis.
Despite the ActiveStor 14 design being suitable for a range of analytics-type applications, Panasas will continue to focus on its technical computing/HPC roots, according to Noer. "Our goal here is not to go after the enterprise market," he says. With the HPC on a very healthy growth trajectory, that should be enough to keep Panasas on its upward revenue path.
The company is demonstrating ActiveStor 14 gear this week at the HPC for Wall Street event in New York City and at the Trading Architecture Europe conference in London. Systems can be ordered today, with shipments slated for November 2012.
Jun 18, 2013 |
The world's largest supercomputers, like Tianhe-2, are great at traditional, compute-intensive HPC workloads, such as simulating atomic decay or modeling tornados. But data-intensive applications--such as mining big data sets for connections--is a different sort of workload, and runs best on a different sort of computer.
Jun 18, 2013 |
Researchers are finding innovative uses for Gordon, the 285 teraflop supercomputer housed at the San Diego Supercomputer Center (SDSC) that has a unique Flash-based storage system. Since going online, researchers have put the incredibly fast I/O to use on a wide variety of workloads, ranging from chemistry to political science.
Jun 17, 2013 |
The advent of low-power mobile processors and cloud delivery models is changing the economics of computing. But just as an economy car is good at different things than a full size truck, an HPC workload still has certain computing demands that neither the fastest smartphone nor the most elastic cloud cluster can fulfill.
Jun 14, 2013 |
For all the progress we've made in IT over the last 50 years, there's one area of life that has steadfastly eluded the grasp of computers: understanding human language. Now, researchers at the Texas Advanced Computing Center (TACC) are utilizing a Hadoop cluster on its Longhorn supercomputer to move the state of the art of language processing a little bit further.
Jun 13, 2013 |
Titan, the Cray XK7 at the Oak Ridge National Lab that debuted last fall as the fastest supercomputer in the world with 17.59 petaflops of sustained computing power, will rely on its previous LINPACK test for the upcoming edition of the Top 500 list.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/15/2013 | Bull | “50% of HPC users say their largest jobs scale to 120 cores or less.” How about yours? Are your codes ready to take advantage of today’s and tomorrow’s ultra-parallel HPC systems? Download this White Paper by Analysts Intersect360 Research to see what Bull and Intel’s Center for Excellence in Parallel Programming can do for your codes.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?
Join our webinar to learn how IT managers can migrate to a more resilient, flexible and scalable solution that grows with the data center. Mellanox VMS is future-proof, efficient and brings significant CAPEX and OPEX savings. The VMS is available today.