- Is the Cisco MARS mission going to abort?
- First iPhone worm spreads Rick Astley wallpaper
- 10 stunning 3D buildings made with Google SketchUp
- Open source software ready for big business
- Four reasons to buy (and one reason to avoid) the Droid
A new breed of server company is targeting customers unsatisfied with the processing power of x86 system clusters and unwilling to pay steep prices for current proprietary symmetric multiprocessing boxes.
Two of these start-ups, Fabric7 Systems and Liquid Computing, have rolled out multiprocessor servers built on AMD Opteron chips that promise high-end compute power at prices lower than those HP, IBM and Sun usually charge for SMP boxes. In addition, the systems include network components that enable end users to allocate not just CPU but also I/O and bandwidth resources on the fly to satisfy changing application demand.
It's an idea similar to that used by developers of standards-based blade servers and rack-based server clusters, which use high-speed interconnects to enable workloads to be shared across systems. But Fabric7 and Liquid Computing, whose founders worked at such network companies as Nortel and Procket Networks, take the clustering idea a step further by including virtualization and network capabilities within a single SMP box.
"The real trick here is marrying the interconnect strength with commodity or off-the-shelf processors, operating systems and middleware," says Jonathan Eunice, president and principal analyst at Illuminata. "We're talking Windows; we're talking Linux. It's standard middleware. It's not a specialized design that needs huge amounts of customization of the whole software stack."
Armed with about $45 million in venture funding between them, Fabric7 and Liquid Computing are targeting different markets: The former positions its systems as the answer for enterprise IT administrators battling with underutilized, proprietary midrange servers, while the latter is eyeing the high-performance computing segment looking for unbounded compute power.
But their approach - combining Opteron processors and advanced network features - is similar, and one that analysts expect to see more of as enterprises more widely embrace the idea of data center virtualization.
Roger Carpenter, vice president of design at Magma Design Automation in Santa Clara, Calif., which makes software for chip design, has been testing Fabric7's Q160 since earlier this year and brought in a smaller Q80 about a month ago. Carpenter says he likes Fabric7's approach because it will enable him to get the processing power he needs without spending a lot of money on high-end systems. A 16-way Fabric7 Q80 with dual-core 1.8-GHz Opteron processors, 32G bytes of memory and two 73G-byte disk drives starts around $83,000, whereas IBM's Web site lists a 16-way p570 with dual-core 1.65-GHz Power5 processors, 32G bytes of memory and two 73G-byte disk drives as starting at around $219,000.
"Much of [electronic design automation] software used very expensive multiprocessing systems in the past," he says. "The trend lately has been toward Linux and two-way Linux boxes."
But now multithreaded EDA applications are demanding more memory and bigger multiprocessor systems at different points in the design process, Carpenter says.
Partner Content
www.bmc.com
Gartner 2009 Magic Quadrant for Job Scheduling
Gartner has positioned BMC CONTROL-M in the Leaders Quadrant of their "2009 Magic Quadrant for Job Scheduling." The report assesses the ability to execute and completeness of vision of key vendors in the marketplace. Read a full copy today, courtesy of BMC Software.
Download whitepaper
Dell's SMART Approach to Workload Automation
Read a compelling case study by EMA, Inc. to learn how Dell uses BMC CONTROL-M to cut cost and increase productivity with workload automation.
Download whitepaper
Workload Automation Cost Savings 2 Minute Video
A major computer manufacturer uses BMC CONTROL-M and just four people to schedule and run over 85,000 jobs every month. By switching to BMC CONTROL-M, they more than quadrupled the workload without adding a single staff member. See how in this 2-minute video overview.
Go to video
Comment