currently interested in awesomeness and how to create it

Panic not quite deserved

Salesforce.com was down for ~40 minutes, and the airwaves are full of “what does this mean for the SaaS market” reflections again.

40 minutes?

Just 40 minutes?

It’s a testament to the ever-increasing capabilities of on-demand vendors that we’re talking minutes not hours or days. And if you’ve ever been a client of a “traditional” software vendor (regardless of respectability), you can certainly appreciate it. You are probably paying top dollar for 99.9% uptime, which translates to roughly 7 hours of downtime a year.

Fact: software systems fail. Sometimes they recover without the user noticing it, sometimes they crash hard. The rates of failure are different in mission-critical systems (airplanes, nuclear power plants) and other domains (both enterprise and consumer). But they all fail sometimes.

We would have to see a systemic failure to start questioning the viability of the affected platform, and since Salesforce.com is not the only SaaS platform out there, I can’t imagine what would have to happen to drive happy customers back to the on-premise cut-throats.

PS … “In fairness, Salesforce.com’s uptime record is still the envy of many IT and business decision-makers” – same column as the one mentioned above.

January 11, 2009 | Filed Under CRM, Software 

comments

One Response to “Panic not quite deserved”

  1. Jeff Grosse on January 12th, 2009

    Tomas, I’m with you. It seems like a bit of a panic for people to just take the opportunity to predict doom and gloom over something that really hasn’t happened (to this extent) for almost 3 years.

    And my guess is, just like last time, Salesforce will take corrective action to reduce the risk of the same thing happening again. Last time, they fixed their architecture and visibility of uptime, and it appears to have paid off. Reliability has been quite astounding. I’m trying to remember just when it happened, but they also ran into a bug with their Oracle database where they had to have Oracle fix it, but that only came up as a result of Oracle never being stretched to that extent.

    My friend JP wrote an article that helps put server and application availability in perspective. http://forcemonkey.blogspot.com/2009/01/no-chicken-little-sky-is-not-falling.html

Additional comments powered by BackType