Forget about Microsoft's Azure, abandon your Amazon EC2 servers, cancel your Gmail account and head for the hills: It's the end of the line for cloud computing. Why? Because Salesforce.com suffered a little service outage last week.
What a load of nonsense!
Reading some of the hysterical commentary on the Internet last week, you'd think there were many valuable and important lessons to be learned from Salesforce's downtime, and that the whole concept of cloud computing is now due a complete rethink. "This is what happens when you lose control of your vital business applications." "How can you possibly trust someone else to run them?" And so on.
Which is just a little bit ridiculous really when you consider that just about all enterprise applications are unavailable occasionally, whether they are run in a corporate data center, the cloud or an old fashioned desktop PC for that matter. In fact, Salesforce.com has a pretty good record for the availability of its applications, and you can be fairly sure that the moment things do go wrong, CEO Marc Benioff cracks the whip and unleashes a truck-load of techies to scurry around and fix things as quickly as they jolly well can. Not many enterprises have that many resources to throw at problems. If you work for a smaller organization that runs its apps in-house, you have probably had the experience of waiting for Barry from IT to come back from lunch before he can take a look at what's causing the outage.
One reason that it causes such a stir when Salesforce.com goes down is that so many people perhaps almost one million users are unable to access their applications and data. That's an awful lot of people and quite an endorsement of the whole software as a service or cloud idea as well, when you think of it. Plus, if you're a Salesforce customer and you can't do any work because the system is down there's always the reassuring knowledge that many of your competitors cannot either.
It turns out that almost nothing is immune to outages and downtime as Microsoft learned to its cost last week. In a stunning display of ineptitude and lack of foresight, its web site was brought to its knees by excited punters who flocked to download the new beta release of its Windows 7 operating system. How did a company that prides itself on being forward-looking not see that coming?
Windows 7 is intended to ensure Windows remains "at the center of people's technological solar system," said Microsoft's Steve Ballmer, in his address at the Consumer Electronics Show (CES) in Las Vegas last week. If you have any idea what that means, do please let me know.
Far from the Windows solar system there are plenty of other stars. One of them, FreeBSD, managed a release without bringing any web sites to their knees. The FreeBSD Release Engineering Team announced its latest version last week, offering copies of FreeBSD 7.1-RELEASE from servers around the world as well as via BitTorrent. Microsoft, please take note.
This is the second release from the 7-STABLE branch of the UNIX operating system, which "improves on the functionality of FreeBSD 7.0 and introduces some new features," the team modestly notes. Most interesting is the news that support for using DTrace inside the kernel has been imported from OpenSolaris.
As if that wasn't enough, Macworld 2009 wrapped up last week, marking exactly 10 years since the introduction of OS X Server. In that time, the server OS has gone precisely nowhere in terms of enterprise adoption, which begs an awkward question: "Why has it failed so dismally?"
Three likely explanations:
- It's a rubbish server OS running on rubbish hardware.
- The OS is insanely great, the hardware is insanely great, but Apple knows nothing about marketing.
- Apple couldn't really give a monkeys *** about the enterprise server market place and is just having a laugh.
Talk to OS X Server administrators (there are some), and they'll tell you that the OS is very good and the hardware is excellent. There goes theory number one.
Ever heard of the iPod, the iPhone or the MacBook? So has everyone on the planet. If there is one thing that Apple is good at it's marketing. There goes theory number two.
You can't help but be left with the distinct feeling that the explanation must be number three: Apple isn't interested in the server market. This is certainly backed up by a quick inspection of the company's web site: There's no mention of server hardware or software on the home page, and you have to dig very deep to find the Xserve or OS X Server software mentioned at all. A lack of any sort of action to back up the solution has also been noticed.
In an article in Computerworld Ken Dulaney, a Gartner analyst, said "... an enterprise-friendly organization would provide staff to go into the enterprise to support them, they'd give customers visibility into future products, they'd provide detailed lists of changes every time they released a device. That's not something Apple does today. They want to do just enough to get past the enterprise barriers involved."
Given Apple's success with its consumer gadgets and that the trend for companies of all sizes to move away from proprietary server software toward open source enterprise OSes like Linux, then perhaps when it comes to OS X Server it's not surprising that Apple is just messing around.
Paul Rubens is an IT consultant and journalist based in Marlow on Thames, England. He has been programming, tinkering and generally sitting in front of computer screens since his first encounter with a DEC PDP-11 in 1979.
This article was first published on ServerWatch.com.