For the last several years, Red Hat has been pushing forward the development of real-time enhancements for Linux. Yet the company has made no formal product announcement of how it would attempt to productize its real-time Linux innovations.
That changed today, with the announcement of the Red Hat MRG (Messaging, Real Time and Grid) platform. The product is expected to be available as a public beta this month, with a generally available release set for early 2008.
The product’s real-time enhancements provide deterministic performance for the kernel that enables actions to occur within the same amount of time, every time — a feature critical for a number of industries including telecommunications, the military, healthcare and financial services.
But Red Hat MRG goes farther, aiming to better appeal to enterprises by adding two more features to the mix: messaging and grid computing.
“Red Hat sees a lot of synergies in combining real-time and grid technologies together,” Bryan Che, Red Hat Product Manager for MRG, told InternetNews.com. “It’s an extension of Red Hat’s automation strategy of ‘any application, anywhere.’ What MRG adds is incredible scalability, performance and reliability.”
Red Hat rolled out its Linux Automation strategy early last month, with new on-demand, virtualization and appliance offerings for Red Hat Enterprise Linux (RHEL).
At the core of the new MRG offering is Red Hat’s Real-Time Linux kernel, which supplants the stock RHEL kernel. Bryan Che, Red Hat Product Manager, told InternetNews.com that the new kernel is just a replacement, so if an enterprise is already running a RHEL-certified application, it will continue to work.
In April, Tim Burke, director of emerging technologies at Red Hat first talked about Red Hat’s real-time efforts for RHEL 5 in a product that he dubbed “RHEL Real Time Edition,” which was under development.
Che explained that the new MRG platform is what RHEL Real Time Edition has evolved into for enterprise consumption.
Real-time features alone didn’t seem to be enough of a selling point, however. As a result, Red Hat bundled in messaging and grid capabilities, he said.
“Enterprises aren’t looking for deployment of real-time on its own, but rather in conjunction with other needs like reducing network latency, ” Che said. “Real-time doesn’t just come out of a separate need, it needs to integrate with enterprise workloads.”
The new messaging components of MRG are intended to leverage real-time to provide low-transaction latency communications. Such a feature could prove critical in scenarios like financial services trading environments, where transactions need to be executed at a deterministic rate (that is, the same rate every time).
The grid side of MRG provides Red Hat users with the ability to scale up computing power to handle workloads. The MRG grid isn’t just for servers — it will enable what Che referred to as “cycle stealing” from desktops. So if an enterprise has idle desktops, whether Linux or even Windows, Red Hat MRG can take advantage of the CPU power on those machines as part of a larger grid computing effort.
The technology behind Red Hat’s cycle-stealing grid is from a University of Wisconsin effort called “Condor.” Under a deal with Red Hat, Condor will soon be available under an open source license, and Red Hat will be working jointly with the university to further develop the technology.
“The technology has been around, but you haven’t seen it adopted pervasively in open source because of licensing,” Che said. “From an enterprise point of view, it wasn’t adopted because it had been mostly used by academia and because it lacked enterprise capabilities.”
To use the cycles of idle Windows boxes, Che said MRG will install a virtualization client on each Windows desktop, which can then run a Linux workload.
The real-time kernel at the heart of Red Hat MRG includes patches that are not yet in the mainline kernel.org Linux kernel. Nevertheless, Che said it has always been Red Hat’s goal to get true real-time features into the mainstream kernel. He added that to date, nearly two-thirds of the patches in the company’s real-time implementation have already been merged into the main kernel.org kernel.
Even when all of Red Hat’s real-time patches are part of the mainline kernel, Che still expects that MRG will remain a needed, standalone product, since it offers more than just real-time.
Red Hat’s real-time platform announcement comes barely a week after rival Novell released its Novell SUSE Enterprise Linux Real Time 10 software. Che said the timing was coincidental, and that Red Hat has been working on Real Time for years.
He went on to note that Novell’s latest offering actually uses technology originally developed by Red Hat.
“With their latest release, Novell has moved to the Red Hat-developed real-time patch set,” Che said. “If you look at it from a technology standpoint, Novell has moved to a version aligned with what Red Hat is doing. It validates our approach that you have to work with the upstream Linux kernel.”