IBM research aims to automate virtual, cloud computing
- 27 March, 2008 08:28
IBM on Wednesday launched a collaborative research initiative with two US universities Big Blue said could ultimately deliver software applications that enable self-managing virtual datacentres and automation in cloud computing environments.
IBM partnered with Georgia Institute of Technology and Ohio State University to conduct the research aimed at developing self-managing technologies within IBM's autonomic computing division, created in 2001. Big Blue officials say the increased complexity virtualization and cloud computing bring to environments require technologies that can keep up with an increased rate and frequency of change. The company expects the research to produce capabilities that would be integrated across IBM products, in particular its Tivoli management software portfolio.
"IBM is very active in autonomic computing and more recently with cloud computing. This research brings two areas together to address customer challenges," said Matt Ellis, IBM's vice president of Autonomic Computing. "More adaptive and flexible computing environments promise benefits, but they increase the complexity of management. This research will bring automation to support these dynamic environments."
The project, slated to be launched at a ribbon-cutting ceremony today in Atlanta, combines Georgia Tech's expertise in creating technologies for managing diverse distributed service-oriented systems and applications with Ohio State's focus on IT process and management issues. The two universities will work closely with IBM Watson and Austin Research Labs and Raleigh development teams to test applications and processes on infrastructure awarded by IBM. The project is one among many in IBM's Shared University Research and Academia Initiative programs, in which researchers at colleges and universities propose research projects to Big Blue.
Being selected by IBM means the universities can conduct their work using IBM BladeCenter H chassis running HS21 servers, IBM System Storage DS34000, network equipment and software that includes IBM Tivoli, WebSphere and Information Management. Along with VMware virtualization and other infrastructure technologies, the initiative will create a prototype computing cloud that links datacentres from the two educational institutions, called the Critical Enterprise Cloud Computing Services (CECCS) facility.
Cloud computing, often used in software-as-a-service delivery models, can also be adopted in-house among IBM clients, Ellis said, and Big Blue plans to be prepared to manage change and complexity in both environments for its customers.
"Customers are expressing interest in cloud computing because they enable them to set up and modify environments in more aggressive manners to meet business demand," Ellis explains. "When you calculate virtual servers, virtual storage and virtual networks into the equation, this type of environment extends server clusters to a much grander scale and it delivers an on-demand environment."
For researchers, the initiative brings many opportunities to test theories and prove that such environments could support critical business applications with the proper management tools and processes.
"The question is can these kinds of open, flexible environments support critical, time-sensitive applications with the use of autonomic technologies," said Karsten Schwan, director of Center for Experimental Research in Computer Systems (CERCS) at Georgia Tech. "We have benchmarks from developing very dynamic and critical applications, for example working with airlines for back-end systems for reservations, that have extremely hard requirements in terms of response times to see if these flexible environments can support such applications."
The shared facility, Schwan said, will provide a "test bed for modern management tools" and train his group on using real-world products such as those from Tivoli in research and development. The IBM project also provides the opportunity for Georgia Tech researchers to learn more about relating business needs to technology applications.
"Ohio State's expertise in business process alignment will help us better understand the high-level requirements that mandate an application, for instance, have a 3-second response time," Schwan said. "It's easy to see that understanding the business process can help us in developing the critical applications that support the business."
Researchers plan to pursue use cases around load balancing, server consolidation, disaster recovery, automated failover and reliability -- all of which would benefit from enhanced autonomic capabilities. And Schwan said the joint team will take this research opportunity, scheduled to run about three years, to also test their theories about power and cooling in virtual datacentres and cloud computing environments.
"We have some fun use cases to pursue, including a joint project with the mechanical engineering power and cooling experts that isn't as directly related to IBM, but could yield very interesting results around better managing infrastructure," Schwan said. "We are hoping we can extend the facility in the future because working with IBM and other experts gives us more and more interesting problems and challenges to try to solve and grounds our research much better in real world use cases."