Blade Servers: A Wonder Drug For Some Business Ailments, But Not All

Blade servers will heal only some of an IT department's most common complaints--they definitely don't cure every annoying symptom.

December 5, 2003

3 Min Read
NetworkComputing logo in a gray background | NetworkComputing

Every new technology promises to be a panacea for every IT ailment imaginable. However, like the latest medicines hitting pharmacy shelves, Blade servers are just what the doctor ordered for these common server ills:

1) Too many servers, too little time. Blade servers' plug-and-go format makes deploying hundreds or even thousands of servers quick and easy. Customers can reap timesavings of up to 80 percent and savings of up to 90 percent of blade servers will heal only some of an IT department's most common complaints--they definitely don't cure every annoying symptom. total costs (labor and hardware), RLX Technology estimates.

2) Over-inflated IT costs. Gartner points to labor and servers as the two largest cost areas in the data center (accounting for 60 percent to 85 percent of total budget). Blade servers substantially lower IT employee-to-server ratio: One IT person can manage two to three times more blade servers compared with standard servers), according to Hewlett Packard.

3) Underutilized servers. Companies with large usage spikes only may achieve 30 percent to 50 percent server utilization. Blade servers allow for 70 percent to 80 percent utilization rates by allowing the speedy reallocation of server capabilities as applications and workloads shift.

4) Mind-bending complexity. Blade servers, which slip into an enclosure, offer simplified power and cooling, and cable consolidation of up to 80 percent, according to HP's Industry Standard Server division.5) Need for quick (and fewer) fixes. Since blade servers have fewer working parts, they are more reliable. When repairs are needed, a new card can be slipped into the slot without powering down or redoing cabling. Replacing a traditional rack-mounted server takes about 10 minutes compared with just a minute or two for a blade server, according to Sun Microsystems.

6) Overpowering power costs. A SPARC blade uses 18 watts of power for average operation, while traditional servers eat up to 60 or 70 watts, Sun Microsystems finds. By lowering power usage enterprises can save substantially in terms of electricity usage.

But find a different prescription if:

1) Processing power is paramount. Today, blade servers are not ideally suited for 64-bit environments, or when you need 8, 16 or more processors.

2) Huge amounts of storage are needed. Blade servers allow for only up to two disks (168 GB) of local storage.3) Expandability is vital. Traditional servers offer up to a dozen or more PCI slots. If you need to extend the capabilities of your servers, then blade servers may not be the answer.

4) A smaller operation. Small businesses won't be able to justify the up-front enclosure infrastructure costs necessary for blade servers, particularly if they need only a few servers.

5) Support is needed for non-standard or legacy technologies. For companies that need to support older standards (such as Token Ring), rather than Fibre Channel and Ethernet, traditional servers will be the answer.

Before you decide whether blades or rack-mounted servers will ease your pain, ask your local IT doctor to diagnose your particular headache, and then decide which server cure is the right fit for you.

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like


More Insights