Replies

How long have you been running Dell servers? We don't rely on Dell servers because our experience shows that they tend to lose their reliability after 2-3 years...and since we run our datacenter with five year replacement cycles, that's just not acceptable.

FYI: In our datacenter, we log every reboot, lockup, hardware failure, and equipment replacement for every server, and we have computerized service logs dating back to 1986. We've run a limited number of Dell servers since 1991, and relied primarily on HP since 1988 (the old Burroughs mainframe was junked in 1983 in favor of an IBM client/server solution, which in turn was scrapped in 1988 because of IBM's astronomical prices in the late 1980's). Anyway, having these computerized maintenance logs allows me to punch up some interesting, and informative, numbers.

By the 24th month of use (24/7 uptimes under constant load), 11% of our Dell servers have experienced hardware related failures. Our HP servers, in the same period, experienced a 9% failure rate. Not much difference? Read on. By the 30th month of use, the failure rate for our Dell servers climbs to 17%...a 6% climb in as many months. Our HP's, in contrast, only increased 1% to a 10% failure rate. By the 36 month mark, our hardware related failure rate for Dell servers climbs to 24%...nearly 1 in 4 of our Dell servers has suffered a hardware failure by the three year mark. Our HP's have incremented up to 11% by this point. By the four year mark, our Dell servers have reached a whopping 39% hardware related failure rate, while our HP's are still sitting at a fairly comfortable 18%. By the time we reach the five year replacement point, 52% of our Dell servers have suffered a hardware related failure, whereas only 25% of our HP's have experienced the same types of problems.

No marketing claims here, no fancy spin, just real world server performance measured over a 14 year period. I've excluded software issues, viruses, power problems, planned maintenance, and "unknowns", leaving only actual failures of computer hardware. My numbers CLEARLY show that HP servers are historically more reliable than Dell servers. The Dell's seem to do pretty good for the first couple years but they don't seem to be built for the long haul. In an environment like ours, long term reliability is actually more important than short term performance.

I'm not saying that Dell servers are bad, just that they aren't as reliable as HP. We've still got a bunch of PowerEdge 4200's and 4400's running as "less-than-critical" departmental servers, and we run PowerEdge 6650's for our media servers because they outperform the HP's in the same price range, but we accept the fact that the 4x00's crash more often, and we replace the media servers on a two year cycle. Dell servers do a great job for the price, as long as the datacenter admins are willing to accept a greater overall TCO and steeper maintenance costs.