June 09, 2011
Today IBM formally launched a range of cloud computing services specifically for the needs of technical and scientific computing users. This is the first in a series of announcements that will emerge over the coming year that put HPC and clouds front and center.
These upcoming announcements will all focus on specific verticals in the realm of high performance computing, beginning with today’s launch of the IBM Engineering Solutions for Cloud. According to IBM’s Deep Computing Marketing and Strategy Manager, Teri Dewalt, this purpose-built HPC cloud solution has been refined for modeling and simulation needs common in automotive, electronics and manufacturing sectors.
According to IBM, this cloud, once this becomes available in the third quarter of this year, it can “enable electronics companies and automotive and aerospace manufacturers to ramp up HPC function as needed to support 2D and 3D product design cycles, helping improve their speed and dexterity in delivering new offerings and services to market.”
This engineering-specific cloud will come equipped with IBM’s Rational app development suite and will provide ISV offerings from Ansys, Cadence and Magma with others in the works, said Dewalt.
Instead of reaching into the ether for case studies to highlight this cloud solution, IBM pointed to their own use of the engineering cloud as satisfactory evidence of its viability. The technology behind the engineering-specific cloud spun out of work from the IBM Systems Development organization that has used the same solution set for its 3,000 distributed engineers that were behind the company’s POWER7 processor line.
According to Dewalt, in IBMs own experiences using these same solutions that are being wrapped into their engineering cloud, they were able to harness modeling and simulation to realize a 50% reduction in developer costs and a product lifecycle that ended up six months under what they usually budgeted for such projects.
Beyond the engineering cloud, the company’s more general announcement of HPC cloud services is targeted at a broad range of technical and scientific computing. Aside from the distinct focus on modeling and simulation, other application areas that emphasize large-scale analytics are at the center of the cloud. Climate modeling, genomics, and applications for financial services or climate modeling are all targets—and some that could find unique service offerings similar to the engineering cloud. For instance, it is reasonable to expect that IBM will make a power play to become the IBM vendor of choice for financial markets in the coming months.
IBM said that the cloud can scale up to many thousands of nodes and users will be able to spin up a cluster in a matter of moments based on their specs. While the details about a GPU angle were not given during our briefing, one can assume this might play a role. IBM states that customers will be able to use any x-64-based servers as well as, of course, IBM’s System X, BladeCenter and Power Systems.
At the center of all of their upcoming industry-specific HPC cloud announcement is IBM’s Intelligent Cluster Solutions portfolio, which is an integrated HPC cluster package that comes complete with servers, storage and switches that they claim comes ready to plug into a data center.
IBM’s workload management suite will handle resource allocation, scheduling, metering and energy use in much the same way other HPC workload managers operate—and with many of the same features in a private cloud context.
IBM’s Dewalt noted that one of the features of this release is the implementation service, which help users get off the ground quickly with installation and configuration of either a private cloud or IBM-hosted private cloud.
Dewalt said that customers were demanding a way to integrate all of their HPC resources into private cloud models with the ability to extend out to other resources when needed. IBM is claiming that they are the only private cloud vendor to offer a solution that is specifically tailored for HPC users, which is hard to verify, but the list of other contenders with such a specific focus is not long.
Today’s announcement is set to complement the enterprise side of IBM’s cloud equation. Its SmartCloud platform, which is aimed at business users, leverages many of the same technologies and core functionality but with integrated business solutions for managing big data—a topic that IBM’s Dewalt sees as relevant to both their SmartCloud and now HPC-tuned offerings.
This cloud service for HPC users could have more appeal for users that Amazon’s HPC-flavored instance type (Cluster Compute Instances) due to the fact that there is no layer of virtualization to muck up the performance required. After all, these are private clouds that are the focus here, therefore, as Dewalt claims, the performance overhead that has prevented many organizations from adopting public resources is mitigated.
This offering stands apart from their Deep Computing On Demand services, which was one of the first HPC on-demand offerings that emerged before the “cloud” buzzword applied to managed services for specialized uses.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
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?