The name of the game in the Wintel world for servers today is quick and cheap.
HP used to build a product that would last, with support but they were always last to market. This equated to a well priced servers that was well engineered and had a consistent logistical service and spare parts support. Sound good ? Not good enough.
Dell was the fast and cheap guy, great sales poor support. They won!
I buy 30 servers a year and won't even deal with HP anymore.
The key to survival in the fast and cheap world is
spare parts replacement, onsite turn around.
You can depend on a high percentage of parts like hard disks and power supplies going bad after 2 to 36 months so always price the replacement support into the cost of the server. After 36 to 48 months price the entire server replacement cost in. Most groups I know use virtualization and ESX VMware to survive bad hardware on any large scale operations today.
Unix Data Centers
Not my field of expertise but Sun, IBM seems to control this space, supplemented with storage by NetAPP, Hitachi, EMC....