Nimbus Goes After HPC Market with Disk-Priced Flash Array

By Michael Feldman

July 20, 2010

Nimbus Data Systems has unveiled its new high-density enterprise flash memory system, delivering 10 terabytes of solid state capacity per 2U shelf. The S1000 can scale up to 250 TB per system and is being priced to challenge spinning disk appliances head on. For HPC and other enterprise users looking to turbo-charge performance of terascale-sized data sets, Nimbus offers a compelling case for making the switch to flash technology.

Four-year old Nimbus is headquartered in San Francisco, Calif., and had been aggressively pursuing the emerging flash-based storage market with its S-class storage arrays. The company has managed to collect about 200 customers, the largest being the US Department of Defense. They’ve also corralled OEM wins with IBM Tivoli and AMCC. As a result, Nimbus says they’re profitable and debt free — not bad for a company that grew up during one of the worst economic downturns in modern times.

The general idea behind employing flash memory for I/O drives is to take advantage of Moore’s Law in order to close the performance gap between external storage and the other computer components. Over the past 10 years, hard drives have not become appreciably faster or more power efficient, while the performance of a computer’s solid state components has increased several-fold. “We believe storage is on an unsustainable trajectory in the datacenter,” says Nimbus CEO Tom Isakovich. “While CPU, memory and network performance have all grown exponentially, storage performance and storage efficiency really have not kept pace.”

External storage demand is escalating, though. Virtualization, data warehousing, and high performance computing are multiplying the need for more I/O, especially random-access I/O. Isakovich says more hard drives, storage tiering, and cache solutions are not the answer. According to him, while they may boost performance a bit, they’re really not addressing the underlying inefficiency of the spinning disk technology. “Drives have really run their course,” says Isakovich.

Nimbus’ mission to drive a stake through the heart of the hard drive was launched in April, with its first all-flash memory S-class storage arrays: the S250 and S500, which provided 2.5 TB and 5.0 TB per shelf, respectively. All of the S-class offerings use Micron’s Enterprise Multi-Level Cell (EMLC) NAND flash, which is five times more durable than vanilla MLC used in consumer devices and much less expensive than Single-Level Cell (SLC) NAND commonly used for most enterprise SSDs. SLC remains the more robust technology, but at about quadruple the cost and a quarter of the density of EMLC silicon.

Nimbus has managed to layer even more reliability on top of the EMLC silicon by incorporating write amplification, wear leveling and dual-parity RAID into its design. They also over-provision the storage by 28 percent to account for the inevitable degradation of the NAND devices over time. The S1000 employs the higher density 34 nm EMLC NAND from Micron, which makes it possible to offer 400 GB of storage per blade. (Because of the over-provisioning, there is actually 512 GB per blade.) The product comes with a one-year warranty, which is upgradeable to three or five years, although Isakovich believes the hardware will actually be just fine for up to 10 years.

Although S1000 performance may be less than the more expensive SLC-based flash memory products out there, the Nimbus offering easily outruns 15K RPM disk array technology typically found in tier 1 storage. Compared to disk, the S-class products deliver up to 24 times more IOPS (1.65 million), up to 16 times faster data transfer (7.2 GB/sec), and 95 percent lower latency (300 microseconds). Space-wise, a single S1000 2U shelf can deliver the same number of IOPS as in four racks of spinning disks.

Since no moving parts are involved, power savings are equally as impressive. Nimbus is claiming 90 percent lower energy usage — as low as 15 watts per terabyte — and a 70 percent reduction in BTU cooling demand. And since there is less heat generated and no motors to wear out, fewer replacements will be needed.

An S1000 shelf is made up of 24 hot-swappable storage blades. Up to 25 shelves can be stacked via 6G SAS ports, making it possible to deploy a 250 GB file system all in flash. A storage shelf is powered by two Intel quad-core Nehalem processors, although Isakovich says expansion shelves don’t require CPUs or the associated memory. According to him, the flash is so much faster than a disk that the CPUs are rarely tied up waiting for I/O to complete, so you just need less of them to manage the storage.

Since all S-class gear speaks iSCSI, NFS, and CIFS, the hardware can act as both a SAN device and a NAS device. Systems come standard with four 10GbE ports (SFP+ or 10GBASE-T) per appliance, which can auto-negotiate down to GbE when needed. Nimbus is also now offering an upgrade to twelve 10GbE ports, using a technology they’re calling “FlexConnect.” It employs triple active-active 10 GbE network controllers, and, in some cases, will eliminate the need for a standalone SAN switch.

The combination of off-the-shelf 10GbE components, Intel CPUs and EMLC NAND chips has enabled Nimbus to achieve cost parity with 15K disks products. All the S-class products, including the new S1000, are priced at $10,000 per terabyte, which is more or less in line with other tier 1 disk-based appliances.

Of course, any vendor could assemble similar hardware, but the S1000 is more than just flash-in-a-box. The real secret sauce is Nimbus’ HALO operating system, a full-featured software stack that comes standard in all S-class platforms. It includes snapshots, replication, mirroring, deduplication, compression, thin provisioning, real-time analytics, proactive notification, and a Web management interface. In late 2010, the company is planning to make a programmable API available as well. Because all this functionality is baked in, there is no need to purchase third-party software or hardware to make the system enterprise-capable. “We think that gives us a sustainable advantage since it has taken us five years to write all this software,” says Isakovich.

Because the company has been able to solve the acquisition cost penalty for flash, while at the same time offering a feature-rich enterprise storage platform, it may be carving a unique home for itself in the IT landscape. Competitors like NetApp, EqualLogic (Dell) and EMC all offer SSD capabilities to one extent or another, but there are no pure flash offerings to match the Nimbus S-class. On the other hand, pure flash array vendors may offer better performance with SLC NAND, but typically bundle little if any software with their systems. And because those systems are based on the more expensive SLC technology, they come at a price premium.

With the S-class platform, Nimbus is looking to go after IOPS-critical storage applications, especially virtualization, traditional database processing, and On-Line Transaction Processing (OLTP). Now, with the higher capacity S1000, they have a credible entry for the HPC market. Data-intensive applications like seismic analysis, image rendering, and many science codes are I/O bound and thus ideally suited for flash-based storage. Isakovich says they have a proof of concept deployment at one of the big supercomputing centers and also have a couple of oil and gas companies looking at systems. He expects to see some customer deployments by the end of the quarter.

The new platform currently tops out at 250 TB per system, but the dedupe and compression technology can boost the effective storage by a factor of 3 to 10, pushing the S1000 into the petascale realm. According to Isakovich, they’re planning to expand system capacity even further later this year. “The demand we’re seeing from the HPC community is rather significant and we think we can continue to push the density envelope even more,” he says.

Subscribe to HPCwire's Weekly Update!

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

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

Nvidia Appoints Andy Grant as EMEA Director of Supercomputing, Higher Education, and AI

March 22, 2024

Nvidia recently appointed Andy Grant as Director, Supercomputing, Higher Education, and AI for Europe, the Middle East, and Africa (EMEA). With over 25 years of high-performance computing (HPC) experience, Grant brings a Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Houston We Have a Solution: Addressing the HPC and Tech Talent Gap

March 15, 2024

Generations of Houstonian teachers, counselors, and parents have either worked in the aerospace industry or know people who do - the prospect of entering the fi Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire