Open-source specialist Red Hat has launched a "real-time" addition to its Linux operating system, which it claims will make some features run 100 times faster than rival technologies.
Red Hat's Messaging Real-time Grid (MRG) was launched as a beta on Tuesday with a full release in the first half of 2008. MRG is an addition to the open-source specialist's Red Hat Enterprise Linux platform, and is designed for businesses such as banks that need to carry out transactions on their IT systems as instantaneously as possible, or "real-time".
MRG gives Red Hat's core operating system real-time functionality, said Scott Crenshaw, vice president of Red Hat's infrastructure business unit at a pre-launch event on Monday. With a real-time operating system, IT managers would have guaranteed response times, and no longer have to overprovision to meet required service levels, he explained. "Our prediction is that real-time will become a normal part of the infrastructure."
MRG includes message-queuing middleware, that allows applications to communicate with each other. It is the first commercial implementation of Advanced Message Queuing Protocol (AMQP), an open-source system that Crenshaw claims is 100 times faster than competing proprietary products from Tibco and IBM.
"I don't think this is the last time you'll see a hundredfold improvement," said Crenshaw, who credits the speed-up of messaging services to the community-development model.
MRG is the first fruit of the vendor's Linux automation strategy announced last month which Red Hat applies to its efforts to simplify the building of enterprise infrastructure, for both users and software vendors. "Any application can run, anywhere," said Werner Knoblich, vice president of the EMEA region at Red Hat, speaking at the time. "It can be on a physical server, or on a virtual server, or on a compute cloud, or an appliance. Software vendors only need to certify it once and it automatically runs anywhere."
The other notable feature of the MRG platform is its grid capability, which will allow enterprise applications to "steal" spare processor cycles from Linux desktops. "Previously grid has been for high-performance computing (HPC)," said Crenshaw. "We are extending that to any application — we call it high throughput computing [HTC]." If there is high demand on web servers, for instance, a company's IT system would be able to use more capacity on desktops and on commercial grid services such as Amazon's Elastic Compute Cloud."
The grid features are developed from a joint project at the University of Wisconsin, where Red Hat engineers worked alongside academics, said Crenshaw. The ability to "steal" cycles from desktops is based on integration with Intel's vPro desktop management technology, which also improves RHEL's ability to manage desktop systems.
