(updated)
Little did Jeff Clavier or Brad Feld know just how timely their posts on “Shared Nothing Architecture” would become in days now that the granddaddy of all on-demand software, Salesforce.com was partially knocked out for almost a day.

The Typepad outage that prompted Brad and Jeff write their piece was just storm in a teacup; this is the real thing, the Perfect Storm. Real business customers could not conduct their business for a day. That something like this would happen was inevitable, but didnt’ we all expect it in the form of a major Internet outage? After all, on-demand vendors are likely to do everything in their power to avoid such outages – or do they? In the case of Salesforce.com, the answer is probably a yes: Earlier this year, Salesforce.com announced it would spend US$50 million to set up redundant East Coast and West Coast data centers with rapid data replication and failover capabilities, an initiative it dubbed “MirrorForce.” (source: IDG).
That’s exactly the kind of commitment Brad and Jeff are asking for, and not all (smaller) providers can afford it. Not that they all should… their core competency being in developing innvative software, not running data centers, which should be outsourced to the “pros” like Vinnie Mirchandani pointed it out numerous times.

Back to our “Perfect Storm”, it will have an effect on the entire on-demand industry, since Salesforce.com is such an icon for this segment. SAP, Oracle etc… will no doubt refer to this “vulnerability” in their sales pitches. Rival NetSuite will not brag about it on their homepage, but their salesforce will likely be trained to point out to prospects why this could never happen to them …

What exactly happened is still unknown – which in itself is quite a customer communications fiasco on Salesforce.com’s part. I bet it will soon be fixed though: the company will come forward with an explanation of what happened, what they do to avoid it in the future, and what they do to accomodate their customers who suffered from the outage. My bet is on Marc Benioff – he will somehow manage to turn this fiasco into a PR victory.

Talk about communication, I am amazed the blogosphere is not abuzz with this story – in fact it’s hardly being mentioned, in sharp contrast to the recent Typepad outage. Isn’t this the type of imbalance Chris Selland and Brad Feld just complained about? Or is everyone out Christmas shopping? Ohh… stores close soon .. gotta run now:-)

P.S. Salesforceless.com is a valid site – I just bought it. (not that I know what to do with it… )

Happy Holidays!

Update (12/21): Others on the subject:

• Jeff Nolan
• ZDNet
• TheStreet.com
• Chris Selland
• NevOn
• Digging In
• News.com
• Silicon.com
• The Stalwart
• NevOn
• Extranet Evolution

Update (12/23): Unlike Salesforce(less).com, TechCrunch is not mission critical software, just an extremely popular blog, yet when they have an outage, Mike finds it important enough to go public right-away. Way to go!

Update (12/31): Reuters talks about Web Services outages, citing Typepad, del.icio.us … etc, not even mentioning Salesforce(less).com. Funny… Nice-to-have services appear to be more important than mission critical business applications?

Tags: 24sevenoffice, Availability, crm, DataCenter, netsuite, On-Demand, oracle, Outsourcing, SaaS, salesforce.com, salesforce.com outage, salesforceless.com, sap, techcrunch