In this video SC12, Mike Stoltz from Xyratex describes the company’s ClusterStor products for HPC. Xyratex recently announced that HP and Dell will resell ClusterStor to power their advanced HPC clusters.
In this video from SC12, Doug Johnson from Aeon Computing describes the company’s innovative Data Oasis technology powered by the Lustre file system.
What advantages does Lustre offer as a foundation for a storage system? Bandwidth. Its performance scales out linearly as the file system scales in build out. The more object servers you have, the more network paths you have, the faster your potential. It is the opposite of a large scale monolithic NFS appliance with one spigot.”
For more information, check out our exclusive interview with Aeon’s co-founder, Jeff Johnson.
JackRabbit is a tightly coupled high performance computing and storage platform that integrates processing and network connectivity at prices starting at less than $1 per GB. siFlash is intended as a metadata server for parallel file systems, fast NFS and CIFS data storage, and similar high performance use cases. Featuring PCIe-flash cards, SSD-flash devices, or RAID-attached flash, siFlash units have achieved random reads and writes exceeding 950k IOPs to and from storage using 8k block sizes and a 30% write mixture in manufacturer benchmarks.”
Read the Full Story.
In this video from SC12, Molly Rector from Spectra Logic describes the latest advancements in the company’s tape-based storage technologies. Along the way, she debunks the myth that Amazon’s AWS Glacier offering will be the death of tape.
Over at Datacenter Knowledge, Rich Miller writes that CERN’s new data center in Budapest is set to be one of the first beneficiaries of a new terabit network created by GÉANT, a European data network for researchers and scientists.
GÉANT’s migration to the latest transmission and switching technology is designed to support up to 2Tbps (terabits per second) capacity across the core network. 500Gbps capacity will be available across the core network from first implementation, delivering circuits across Europe that will allow individual users to transfer data at speeds of up to 100Gbps, or multiples thereof, thereby enabling faster collaboration on critical projects and meeting the rapidly increasing demand for data transfer.
Talk about Big Data–the CERN Large Hadron Collider generates over 100 petabytes of data per year at its home near Geneva, Switzerland. Read the Full Story.
In this video, Panasas Chief Scientist Garth Gibson describes how the company brings Hadoop support to bear in the world of Big Data and HPC.
Hadoop is a great platform for taking a gigantic amount of information and reducing it down to the central core that you then want to do the second level of analysis on. And that’s what’s happening across the enterprise, data warehousing, and HPC. So the fundamental issue is that after you’re done crunching with that commodity Hadoop cluster, you now have valuable assets. You want those valuable assets on a system you trust. You want it on a good, high-quality NAS. But it has to keep up. You need the high speed of a direct-flow environment. And then, it turns out, that once you can process from off-board quickly, you can optimize Hadoop and go faster in many cases because you’re using your off-board NAS.”
We are big fans of Lustre here at insideHPC. And when I learned that Aeon Computing was building their Data Oasis storage platform on the open source file system, I sought out the company’s co-founder Jeff Johnson to learn more
insideHPC: Why build a Lustre appliance? What was the customer problem you set out to solve?
Jeff Johnson: This is all about performance. We originally set out to plan a Lustre deployment and brought in storage solutions from various well known manufacturers for Lustre benchmarking. We were surprised to discover that many storage devices that looked to be high performance on paper performed quite poorly in a Lustre environment. As a result we decided to develop one ourselves. In this particular case the customer needed Lustre performance without all of the extra, and costly, active failover high-availability features found in the very high end Lustre solutions available. That is what we built.
insideHPC: What sets the Oasis Appliance from other available storage solutions available out there?
Jeff Johnson: There are many storage solutions available in the market but not all of them do Lustre well. We set out to design a Lustre platform that was good at Lustre data and I/O profiles. Part of that design, in addition to performance, is that it follows Aeon Computing’s business philosophy in that there is no unnecessary, extraneous bull___t that gets in the way.
insideHPC: Is Oasis just for the HPC market, or does it have appeal to the Big Data and Cloud-based applications?
Jeff Johnson: The Data Oasis filesystem is specifically an HPC application. The EclipseSL appliance is well suited for any application where reliable, dense high-performance storage is needed. We have mainly focused on Lustre, the EclipseSL would perform well in Gluster and other distributed filesystem applications.
insideHPC: How does the Aeon design change the way you architect a Lustre system?
Jeff Johnson: We designed the EclipseSL with no bottlenecks. The data flow from the network interfaces through bus infrastructures and memory to block storage and back out encounter no bottlenecks or oversubscription. I don’t know that what we did necessarily changes the industry in any way. We simply started with a blank sheet of paper.
insideHPC: What advantages does Lustre offer as a foundation for a storage system?
Jeff Johnson: Bandwidth. Its performance scales out linearly as the file system scales in build out. The more object servers you have, the more network paths you have, the faster your potential. It is the opposite of a large scale monolithic NFS appliance with one spigot.
insideHPC: What was your first reaction when you learned you won the Best of Show Award?
Jeff Johnson: We checked the calendar to make sure it wasn’t April Fools Day. There may have been some expletives of disbelief uttered in the office. Once it sunk in, humbled, flattered and still a bit surprised.
Check out Aeon this week at SC12 booth #2119.
Colfax Hadoop appliance will help customers overcome the perception that the big data analytics technology is hard to deploy and achieve business value,” says Gautam Shah, CEO of Colfax International. “With an ideal balance of performance, scalability and price, the Colfax Hadoop appliance enables organizations to overcome the barriers to big data analytics, providing a perfect platform to extract the most value from their data.”
Mellanox’s Unstructured Data Accelerator (UDA) enhances the Hadoop appliance performance by nearly 50 percent, providing data analysts with near real-time processing power. UDA brings RDMA capabilities into the Hadoop framework, enabling efficient data transfer and lower CPU overhead. The UDA package is an open source plug-in tool to the Apache Hadoop Map Reduce frame work.
Colfax and Mellanox will demonstrate the Hadoop appliance at SC12 in the Mellanox booth #1531. Read the Full Story.
This week at SC12, DDN announced a $100 million investment in its research and development efforts, specifically directed at resolving key challenges to achieving Exascale levels of performance in scientific computing. The new investments by DDN represent a substantial percentage of DDN’s engineering resources and will be directed toward technology challenges, which become critical at Exascale proportions, including: I/O Acceleration; Converged Infrastructure; Information Value Extraction; and Energy and Data Center Efficiency.
Data-intensive computing impacts individuals, organizations, industries and governments by enabling the creation of valuable information based on massive volumes of highly complex data,” said Alex Bouzari, CEO and cofounder, DDN. “Significant investment is required to allow researchers to address challenges such as the design of new materials needed for better electric car batteries, the improvement of multi-physics models for more accurate severe weather modeling, and the development of high-resolution cosmological simulations to help understand dark matter and the universe around us. With today’s announcement, DDN is establishing a clear direction for our Exascale computing agenda and reaffirms DDN’s continued central role in the future of supercomputing.”
DDN disclosed that the investments will center around the following critical technologies for Exascale computing:
- I/O Acceleration: New file system, middleware and storage tiering methods will be required to eliminate scalability barriers associated with conventional methods of file, object and database access in order to achieve 1,000x scalability, TB/s performance and million-way application CPU parallelism.
- Converged Infrastructure: The convergence of computing, storage and networking technologies will give rise to intelligent and accelerated data storage infrastructures which can co-locate pre-processing and post-processing routines natively within the storage infrastructure to enable applications to access data with increased acuity.
- Information Value Extraction: Leveraging converged infrastructures, DDN R&D efforts will support the development of scalable data analytics environments to extract actionable insights from vast volumes of unstructured data.
- Energy and Data Center Efficiency: With the emergence of storage-class memory and software tools, infrastructures can be built with fewer components compared to today’s disk-based technologies. These initiatives will serve to significantly reduce hardware acquisition costs but will also make data centers much more space and power efficient by reducing storage footprint by more than 75%.
This is exciting news for the HPC community looking to see more vendors step up to the plate for the incredibly daunting goal of Exascale computing within this decade. I can tell you that the DDN booth was packed with people all week at SC12. Read the Full Story.
This week Cray announced the Sonexion 1600 storage system. As the storage technology that powers the Blue Waters supercomputer, the Cray Sonexion 1600 system scales from terabytes to petabytes with a modular architecture that enables incremental performance and capacity scalability for HPC workloads.
Storage has become essential to scientific computing, and what Cray has done is design and build a storage system that is immense. I don’t know of any other system with the requirements for scalability that is being put together for such huge applications,” said Bill Kramer, deputy project director for the Blue Waters Project at NCSA. “With the Sonexion 1600 storage system and its Cray Lustre environment, Cray has built a file system that is highly reliable, productive and very performant. NCSA also has recently demonstrated over 1.1TB/sec of I/O bandwidth with this file system for multiple real-world use cases, and we believe this to be the fastest Lustre file system ever built.”