Integrator Toolbox: Think small, save big with blade servers
- By J.B. Miles
- Mar 19, 2003
What's the big deal about blade servers? Everything's smaller, from the size of the servers, to the number of cables and wires cluttering up the back of a rack, to the time required for setup and management and, at least in some cases, in the amount of power consumed.
At the simplest level, blade servers are on a single card that plug into a rackmount chassis. They're hot-swappable. Just slide them into the chassis, and they're up and running in seconds with a minimum of fuss. Hundreds of blades can fit into a very small space.
Blade servers are cost-effective because they eliminate many of the complications of rackmount designs and pack a lot of computing power into small spaces. You need less technical expertise than with other servers, and they help eliminate annoying cable clutter.
The term blade server typically refers to a proprietary chassis that can hold a number of blades that act as independent servers. For example, IBM Corp.'s eServer BladeCenter is a 7U chassis that can hold up to 14 server blades for up to 84 high-end servers in the rack.
For large, enterprise-class workloads, choose blade systems with high-end, dual chipsets; quads aren't yet available, though Hewlett-Packard Co. will release the ProLiant BL40p later this spring. J.B. Miles of Pahoa, Hawaii, writes about communications and computers. E-mail him at firstname.lastname@example.org.