Feature: Grids and clouds - reaching for the next phase |
||
|
||
"This is not a replacement technology," says Ignacio Llorente, "This is the next phase of evolution for grids." Llorente coordinates 'virtual machine' management for RESERVOIR, an EC-supported project standing for Resources and Services Virtualization without Barriers, that works to enable deployment and management of complex IT services across different administrative domains. This project began collaborating with EGEE in June 2009 to marry the advantages and practicalities of cloud computing with grid technology. After starting in February 2008, the project is beginning to show concrete results: their open source cloud toolkit 'OpenNebula' can be integrated with the grid middleware - gLite or any other - without alterations to the code. In mid-January 2010, the project achieved a milestone with the stable release of OpenNebula 1.4. (A description of the release along with download links is available on the OpenNebula blog.) "This is an excellent technical collaboration to see how cloud technology could impact the way the future European Grid Infrastructure is delivered," says Steven Newhouse, technical director of Enabling Grids for E-sciencE and interim director of the European Grid Initiative (for more details see the iSGTW 16 September 2009 story, "EGI from the interim director's view"). |
||
The Goldilocks solution? Grids and clouds can work together in several ways, at least three of which are currently being explored. Cloud-technology can make it possible to create virtual grid sites on commercial cloud resources (See iSGTW story "Grid in a cloud: Processing the astronomically large") - a model which the StratusLab collaboration is pursuing. It seeks to virtualize an entire site, running all basic grid services in the cloud, thereby making their resources more reliable and giving providers more flexibility. The ability to rapidly increase the available resources on demand being the primary advantage, also termed 'scaling up.' In theory this could be infinitely scalable, however providers still need to iron out the technicalities of this approach through testing and trial deployment. In another grid-cloud model, the CERN IT department - which hosts a real grid site - has shown how RESERVOIR's OpenNebula technology can deploy 'virtual machines' specific to certain applications or 'jobs.' One piece of hardware can then run several operating systems simultaneously (within a 'hypervisor'). This gives the user communities much more flexibility; however, the providers have less control over the application environment. This is not necessarily a problem, the codes of policy need to be set before it is used more widely. A third possibility is a 'hybrid' solution: combining private infrastructure with cloud resources to expand the computing capacity of an EGEE site - in this system both 'real' grid resources and 'virtual' grid resources would be used. "I think this is where the future of grid computing lies," says Llorente. "This technology will help us save energy and resources and, in the end, be more reliable and flexible for users and providers." -Danielle Venton, EGEE |