One of Apple's best friends in the enterprise market is retrenching and re-researchging portions of the study it published Oct. 19 predicting the Mac OS will increase its market share in the enterprise by 57 percent, less than a month before Apple killed the Xserve line of rackmount servers that might have helped support that growth.
There are a couple of caveats to that. First, it's the two-year-old consortium of integrators, VARs and a vendors is the Enterprise Desktop Alliance, so they weren't all that concerned with the Xserve anyway. They were looking at laptops and desktops, and the story was mainly relevant because EDA rightly pointed out that iPhones and iPads were making at least as big a dent in the enterprise as any other Mac box.
Plus, not that many people were buying Xservers anyway, which is why Apple killed them in the first place.
And, to keep its information up to date, the EDA launched a survey right after the news broke asking how (and whether) people actually use Xservers.
Neither is currently available as a rackmount. And according to Macworld's John C. Welch, who's more qualified than I to identify the comparative weaknesses of various Mac boxes, the Mac Pro is no server. It doesn't rack mount; it doesn't hot swap; you have to open the case to work on it and the components aren't designed for 24/7 high-heat, high I/O, (comparatively) high utilization the way a data-center server would be.
I don't see data-center managers stocking up on them any time soon.
Which raises the other question in this: Yesterday I wondered if bailing on its only rackmount server made Apple a less serious contender for enterprise accounts.
Now I'm wondering if having a server is even relevant, at least to Apple.
If you can virtualize all the data and the OS and run any corporate computing environment to any handheld device (not yet practical, but VMware, Citrix and Microsoft are all pushing customers in that direction) does it make sense to choose Apple as one of your hardware vendors if all you can get from it are endpoints?
Are virtual servers and cloud computing so powerful and paradigm-shifting that it will be practical to commit to purchasing a large number of client machines with a different OS and different support requirements, even if there's no server component to make interoperability and content-serving more effective?
Can you really be a core player in the enterprise if you only live at the edge?