Advertisement

Xserve End Of Life: Some opinions and ideas about Apple's server strategy

As an Apple Consultants Network member who has installed his share of Apple Xserves, I have mixed emotions about Apple's decision today to pull the plug on the only "real" server that they've been selling. On the one hand, Apple hasn't been selling a lot of the pricey pizza boxes to big business, so it makes sense that Apple would simply re-purpose existing products -- the Mac mini and Mac Pro -- as servers. On the other hand, I think it sends mixed signals to the enterprise market about Apple's commitment to business.

Apple Senior World Product Marketing Manager Eric Zelenka stated in a post to the Xsanity forums that Apple is still committed to the server products, technologies, and devices, and that the decision to kill the Xserve has no impact on future Xsan or Mac OS X Server development. I'm wondering if that's just a lot of marketing-speak to try to calm down admins who have a huge investment in Mac OS X Server, Xsan, and Xserves. According to our own Victor Agreda, who is at MacTech 2010 this week, many Mac IT admins feel that the Mac OS X Server flavor of Mac OS X is safe for the time being. I'm not so sure -- read more of my personal thoughts on the next page.


The departure of the Xserve from the Mac server lineup got me thinking a bit about the future of the Mac OS X Server software. In the past, pre-announcements of Mac OS X updates also meant that the Apple website was updated with statements about new features in Mac OS X Server. This was true for both Mac OS X Server 10.5 and 10.6. With the recent announcement about Mac OS X 10.7 Lion, there was no corresponding site update touting the upcoming features of the server version of the OS. Perhaps the OS X Server team just hasn't made their plans public, or maybe there are no awe-inspiring new features, but it is surprising that Apple didn't at least make a nod in the direction of the server OS.

Could Apple be tossing Mac OS X Server out the door in the future? Perhaps. In an IDC survey of market share of server operating systems published earlier this year, Windows Server operating systems had about a 75 percent share by units installed, Linux about 21 percent, and Unix operating systems (including Mac OS X Server) a measly 4 percent. When you consider that the Unix OS picture includes Oracle/Sun's Solaris, which is still popular with some organizations, Mac OS X Server probably doesn't account for much in terms of the server OS picture. As such, it can't be adding that much to Apple's bottom line, either.

Perhaps, in the interest of only having to support one operating system, Apple might be considering making the server administration pieces and background processes of Mac OS X Server an "add-on app" to Mac OS X 10.7. In other words, if someone wanted to turn a Mac Pro into a server, they'd purchase apps from the upcoming Mac App Store that would add the functionality of Mac OS X Server to Lion.

Let's say that a company wants file and print services, but doesn't want to worry about mail, web server, wiki server, calendar server or any of those other services -- they could purchase a "file and print services" app and a "user administration" app and add it to the base Lion install. If they later decide to let the server provide additional services, say email and web services, the company could buy those apps and run them on the server. It's an intriguing idea, and one that could make Mac servers even more attractive and affordable to small business. Most of Mac OS X Server's raw functionality comes from the open source underpinnings of OS X and the Unix-oriented server applications it supports (Apache, Samba, Jabber/XMPP, etc.). Breaking those into install-on-demand packages, with official Apple support and maintenance, might even get more Mac servers into the field.

Some other thoughts:

  • Apple's decision on the Xserve does seem to highlight the company's commitment to small businesses, most of whom can afford a Mac mini server to handle file, print, web, mail, calendaring, and other services.

  • Enterprises are not going to shell out for either the Mac mini server, which doesn't have the "horsepower" to power enterprise apps for thousands of people, nor are they going to want to fill server rooms with Mac Pro boxes. The Xserve was the only standard 1U rack mount device made by Apple. Sure, you can purchase rack mounts for Mac minis, but as I just said, they're not powerful enough for enterprise use. This suggests to me that Apple is abandoning the server side of enterprise business and just concentrating on small business. That's not necessarily a bad idea, as there are a lot more small businesses than massive enterprises.

  • The road ahead for Apple's Xsan product line (storage area networking over Fibre Channel for video editing and other collaborative applications that need top transfer speeds) is a little bit murky. Although Zelenka's comment noted above gives some hope to Xsan installs, the lack of a rack-mount server option from Apple is going to complicate matters quite a bit. Xsan deployments need Mac OS X Server to control metadata on the massive RAID arrays that make up the SAN storage pool, and the density requirements of most server rooms will not lend themselves to replacing the existing boxes with Mac Pros when the time comes; the Mac mini, with no expansion slot to add an FC card, is not an option. Other SAN solutions qualified for Final Cut Pro use are available (like Quantum's StorNext), but they're more expensive to deploy and maintain than Xsan is. Keep an eye on the Xsanity forums and the Macenterprise.org mailing list over the next few weeks if you want to follow the conversations and options for this sector.

  • Mac OS X has always been the red-haired stepchild in enterprises, mainly because Apple hasn't (in my opinion) done enough to make the OS a good player in Windows networks. Despite Apple's assertion that OS X works well in any environment, there have been issues with binding Macs to Windows Active Directory. Things have improved with Mac OS X 10.5 and 10.6, but many admins have found enough issues with Apple's way of doing this that they are forced to resort to third-party solutions such as Thursby Software's ADmit Mac. Perhaps Apple should concede the server OS market and just concentrate on making sure that Mac OS X 10.7 can be a completely equal client in the dominant Windows environment in the enterprise market.

  • Server In The Clouds: Could Apple be looking at providing cloud services for companies? Once again, this is in line with some of the strategic moves that Apple is making. Imagine if a business could rely on Apple to provide file, web, wiki, calendar, mail, and other services for a low monthly fee, through some sort of "virtual server" service. There's no upfront cost for hardware, backups are done off-site by Apple, and there's a 24/7/365 team of experts making sure that the "server" doesn't go down. Apple's making that big investment in the new North Carolina data center for some reason, and perhaps a future "server in the clouds" service could be part of the reason that facility is being built.

As I noted earlier, these are just some thoughts that were bouncing around inside my head today as I thought about my experiences with Xserves in the past. I'd love to see how other server admins feel about Apple's decision, and what they think might be the future for Mac OS X Server. Leave your comments below.