Red Hat's Real-Time Linux Move
Hoping to counter rival Novell in the Linux market, Red Hat officially shipped the second release of its Enterprise MRG Real-Time Linux.
According to some Fort Lauderdale computer repair experts, Red Hat's MRG product is essentially a variant of the core Red Hat Enterprise Linux stack.
The new 1.1 version of Enterprise MRG, originally due to ship by December 31, is intended to replace the standard generic Linux kernel with a real-time kernel.
Specifically, one based on the config_preempt_rt patch.
According to sources, that patch that was jointly created by IBM, Novell, Red Hat, Silicon Graphics, and several smaller companies.
How, specifically, does the new version differ?
Instead of supporting simple and reliable functions for both database and applications serving like Enterprise Linux, MRG is sculpted to best carry out a combination of messaging, real-time, and grid computing workloads.
Breaking down the constituent parts of MRG, Red Hat officials note that the R is relevant mostly to financial services companies and defense contractors.
The M is for those products, such as IBM's WebSphere MQ, that tend to pass a lot of messages between applications and servers.
The G in MRG, until release 1.1, has been the missing link.
Red Hat officials said it has been the case largely because its programmers have been working overtime to integrate the Condor grid into the Enterprise MRG product.
They said that when MRG went into beta in late 2007, Condor technology was yet to be stitched in because it had yet to support JBoss Enterprise stack.
Red Hat nemesis Novell is, at the same time, chasing this same market with its SUSE Linux Enterprise Real-Time product, a variant of its SLES 10.
The two have been trying to best each other with lower levels of latency and higher throughput rates over the past year or two.
Red Hat has also added native InfiniBand and Remote Direct Memory Access drivers designed to aid in accommodating lower-latency clustering.
The drivers, which also work with Ethernet, make it possible for nodes on a cluster to transport data among nodes directly into and out of memory in those nodes.
Continue reading this article in Information Week ...
