IBM is taking aim at the venerable x86 architecture with the new eX5 — an enhancement it sees as overcoming some of the shortcomings of x86, like limitations on memory.
The idea is that servers will benefit from being able to access more memory more quickly, rapidly improving efficiency and in time reducing server underutilization and datacenter sprawl. HardwareCentral takes a look.
The company that introduced the IBM Personal Computer almost thirty years ago is launching a solution it believes will “take the PC out of the PC server” and greatly increase the memory capacity of a standard x86 server. The company’s new eX5 Architecture will appear first in blades and racks due out later this year.
“x86 servers today are fundamentally derived from the desktop PC architecture. It’s going on 30 years old, and the limitations of that architecture that still manifest today are the main contributing factor to that sprawl and underutilization today,” Tom Bradicich, IBM fellow and vice president of systems technology for System x told InternetNews.com.
Memory is make or break for performance. The less reading and writing you have to do to the hard disk, the faster everything becomes. There have been a variety of attempts to address the problem, some successful, such as Cisco’s special memory controller on its Unified Computing System blades, and some failures, like the now-defunct MetaRAM.