With the OpenPOWER Foundation, IBM is bucking the trend of other chip and server manufacturers’ proprietary business models and has released detailed technical specifications for its POWER8 processor, inviting collaborators and competitors alike to innovate on the processor and server platform, providing a catalyst for new innovation.
IBM unveiled the new POWER8 Systems at an Open Innovation Summit Big Blue held with the OpenPOWER Foundation in San Francisco on April 23. Built on IBM’s POWER8 technology and designed for an era of big data, the new scale-out IBM Power Systems servers culminate a $2.4 billion investment, three-plus years of development.
The systems are built from the ground up to harness big data with the new IBM POWER8 processor, a sliver of silicon that measures just one square inch, which is embedded with more than 4 billion microscopic transistors and more than 11 miles of high-speed copper wiring.
“This is the first truly disruptive advancement in high-end server technology in decades, with radical technology changes and the full support of an open server ecosystem that will seamlessly lead our clients into this world of massive data volumes and complexity,” said Tom Rosamilia, senior vice president of IBM’s Systems and Technology Group, in a statement. “There no longer is a one-size-fits-all approach to scale out a data center. With our membership in the OpenPOWER Foundation, IBM’s POWER8 processor will become a catalyst for emerging applications and an open innovation platform.”
IBM’s POWER architecture is the cornerstone of innovation for the OpenPOWER Foundation, creating a computing platform available to all. The Foundation – representing 25 global technology providers and growing – was founded by IBM, Google, NVIDIA, Mellanox and Tyan. The group announced an innovation roadmap detailing planned contributions from several of its members, with IBM’s Power Systems as the first servers to exploit OpenPOWER technology.
The first POWER8-based systems to debut are five Power Systems S-Class servers designed for large, scale-out computing environments. The new line-up helps to reduce floor space, power and cooling costs. IBM has designed these systems to perform at a sustained 65 percent utilization — a rate higher than common x86 utilization levels, IBM said. With twice the data throughput compared to an x86-based server, the new Power Systems can help cut data center footprints in half, the company claims
With availability beginning June 10, the new scale-out S Class servers include two systems that run Linux exclusively – the Power Systems S812L and S822L servers. The three additional offerings, the Power Systems S814, S822 and S824 servers, provide users the choice of running multiple operating systems including Linux, AIX and IBM i. Available in 1 and 2 socket and 2U and 4U configurations, the starting price of the new servers is $7,973.
FIS, a provider of payment processing and banking solutions that manages financial data for thousands of companies around the world, expressed its anticipation over the combination of IBM POWER8 and IBM FlashSystem technology.
“We expect higher utilization and performance capabilities along with the flexible computing resources needed to meet our client’s application processing and business delivery requirements,” the company said in a statement. “POWER8’s secure-key cryptographic accelerator and cryptographic coprocessor functions combined with FlashSystems’ enterprise ready extreme performance and application latency reductions capability provides an infrastructure that is critical to the success of today’s core banking application environments.”
IBM also announced three new Power Systems solutions optimized for big data and analytics solutions. The new technologies, IBM Solution for BLU Acceleration, IBM Solution for Analytics and IBM Solution for Hadoop, are optimized for IBM’s new Power Systems to deliver quick insights on both structured and unstructured data. For example, the new IBM Solution for Analytics provides speed of insight for today’s data driven analytical, computational and cognitive workloads through integration with Cognos, SPSS and DB2 with BLU Acceleration, IBM said
According to IBM test results, the IBM Power Systems are capable of analyzing data 50 times faster than the latest x86-based systems. Certain companies have reported analytics queries running more than 1,000 times faster, reducing run times from several hours to just seconds, IBM said.
IBM also announced two Linux developments that fortify rapid cloud innovation on POWER8 systems. One is the availability of Ubuntu Server 14.04 LTS, Ubuntu OpenStack and Juju service orchestration tools, on all POWER8 systems; and the introduction of PowerKVM, a Power Systems-compatible version of the popular Linux-based virtualization platform KVM, on all POWER8 systems that run Linux exclusively.
IBM Launches POWER8 Systems, OpenPOWER Roadmap
IBM’s collaboration with Canonical, the commercial sponsor of Ubuntu with more than 20 million users worldwide, provides easy migration for applications to Linux for cloud deployments to deliver big data and mobile software applications and to boost the performance of existing applications across cloud platforms. IBM is offering the latest release of Ubuntu Server, Ubuntu OpenStack and Canonical’s Juju cloud orchestration tools on the new Power Systems announced today and all future POWER8-based systems.
This complements the existing support by IBM for Red Hat and SUSE Linux operating system distributions on its complete lineup of Power Systems.
Meanwhile, at the Open Innovation Summit, the OpenPOWER Foundation, showed the first reference board and OEM systems, and innovations including many forms of acceleration, advanced memory and networking. OpenPOWER has grown to more than two dozen members including global hardware and software thought leaders. The foundation makes POWER hardware and software available for open development, as well as POWER intellectual property licensable to other manufacturers. OpenPOWER is greatly expanding the ecosystem of innovators providing value back to the industry and end users.
“We are very pleased with the growth of the OpenPOWER community and the progress made by the Working Group members even at this early stage,” said Gordon MacKean, Chairman, OpenPOWER Foundation, in a statement. “The projects feeding the innovation pipeline to date will greatly enhance the performance of the next generation of servers by eliminating system-level bottlenecks.”
At the summit, the OpenPOWER Foundation presented its first white box server details including a development and reference design from Tyan, and firmware and operating system developed by IBM, Google, and Canonical. The OpenPOWER Software stack in this white box design is targeted for ease of implementation in hybrid deployments. IBM noted it will be deploying systems leveraging this OpenPOWER hardware and software stack in Softlayer later this year.
Speaking on why the time is now right for OpenPOWER, Brad McCredie, president of the OpenPOWER Foundations, said, “The first thing is the decline of Moore’s Law. It’s very hard to just take the silicon and keep going. The second is the decline of PC volumes. The PC isn’t going to be dominating all of the specifications. The third is we’ve entered the third generation of scale-out computing, which we think will look a lot like OpenPOWER.”
McCredie noted that IBM’s POWER technology has been known for building big systems that have doors on them – that you can open a door and walk into them because they are so big. However, “We’re taking that technology that builds computers that have doors on them and we’re decomposing it into its elements and we’re making it available to the ecosystem to innovate. We’re taking the processor and we’re allowing people to license that processor technology – the very technology that’s in the world-famous computer that’s run banks, and let them innovate on that and make scale-out computers. We’re building firmware stacks. They’re going to be open source to enable people to take that and innovate to build computers with scale-out performance.”
MacKean, who also is the engineering director for the platform group at Google, said Google has always considered itself an innovation leader in platform technology. He said Google looked at OpenPOWER as the opportunity to launch that third generation of scale-out computing McCredie described – to really break down the barriers that exist between the components.
“We wanted to create something where the whole is greater than the sum of its parts, to really fix those bottlenecks that exist between those different components that make up a system – networking, memory subsystems, I/O subsystems, storage subsystems and acceleration,” MacKean said.
He added that Google is doing a thorough evaluation of POWER for the kinds of workloads the company sees and is helping to develop and harden a lot of the codebase and development tools, the operating system and the compilers. “We’re working very closely with the rest of the member companies in helping to harden that software stack,” he said.
Founding OpenPOWER Foundation member NVIDIA is adding CUDA software support for NVIDIA GPUs with IBM POWER CPUs. IBM and NVIDIA are demonstrating the first GPU accelerator framework for Java, showing an order of magnitude performance improvement on Hadoop Analytics applications compared to a CPU-only implementation. NVIDIA will offer its NVLink high-speed GPU interconnect as a licensed technology to OpenPOWER Foundation members.
Twenty-five members have joined OpenPOWER including Canonical, Samsung Electronics, Micron, Hitachi, Emulex, Fusion-IO, SK Hynix, Xilinx, Jülich Supercomputer Center, Oregon State University, and several others since OpenPOWER formed as a legal entity in December 2013.