Yahoo, HP, Intel Give Ivory Towers a Stairway to the Cloud
Three of the biggest names in technology -- HP, Intel and Yahoo -- are joining forces to develop a massive test bed for academics to use in cloud computing research. The project will utilize a hardware infrastructure supplied by HP and Intel, together with open source software systems, including Hadoop, from Apache; Pig, from Yahoo Reasearch; and Tashi, supported by Intel.
Jul 30, 2008 11:03 AM PT
Yahoo, HP and Intel are collaborating on an ambitious research endeavor called "Cloud Computing Test Bed" -- designed to support cloud computing research and education at universities. Users will be able to develop and test software, data center management, and hardware associated with cloud computing on this large-scale grid.
The three companies are partnering with the Infocomm Development Authority of Singapore, the University of Illinois at Urbana-Champaign, and the Karlsruhe Institute of Technology in Germany on the project. The partnership with Illinois also includes the National Science Foundation.
The Issue of Scale
Others have mounted similar educational/corporate ventures, of course. What is different about this one is its scale, said Russ Daniels, vice president and CTO of HP Cloud Services Strategy. The test bed will be a globally distributed, Internet-scale testing environment instead of the limited computing resources typically available to academia.
"Researchers will be able to test their theories in a large-scale environment rather than to do small-scale simulations," he told LinuxInsider. Oftentimes what happens is after researchers have done their simulations, they find the implementation in a larger-scaled environment does not always work out, he added. "The research community has done a lot in this space, but we have only started to scratch the surface of what can be done with scaled testing."
Another benefit of the project is that it is expanding cloud computing resources to new institutions, said Yahoo Research Fellow Raghu Ramakrishnan, chief scientist for audience and head of the community systems group. "It is a significant departure, because we are not creating a cluster where the [same researchers] can grab more time," he told LinuxInsider. "We are opening the vault to people outside of these one or two institutions that have invested in cloud computing."
The test bed will initially consist of six "centers of excellence": IDA facilities; the University of Illinois at Urbana-Champaign; the Steinbuch Centre for Computing of the Karlsruhe Institute of Technology; HP Labs; Intel Research; and Yahoo.
Each center will host a cloud computing infrastructure using HP hardware and Intel processors; there will be 1,000 to 4,000 processor cores in all. The test bed will run Hadoop -- an open source, distributed computing project of the Apache Software Foundation -- and other open source, distributed computing software such as Pig, the parallel programming language developed by Yahoo Research, and Tashi, an open source cluster management system for cloud computing supported by Intel. Some of the test bed is already in operation, Ramakrishnan said. The rest will be deployed throughout the remainder of the year.
For example, Carnegie Mellon University has been using M45 -- Yahoo's supercomputing-class data center -- for cloud computing research since last November, he said. That technology is being deployed in the Cloud Computing Test Bed.
When it is fully deployed, the Cloud Computing Test Bed will support a wide range of research areas. Researchers at HP Labs, for example, will use the test bed to conduct advanced research in the areas of intelligent infrastructure and dynamic cloud services. IDA plans to leverage the test bed and its industry partnerships to train local students and professionals on cloud computing technologies.
"What you are seeing here is watershed in the computing landscape," Ramakrishnan said.
The end results won't be apparent for a while, he added. "This kind of paradigm shift doesn't happen overnight. Different partners will focus on different parts of the grid -- but, hopefully, together we can reinvent the next generation of computing infrastructure."