Salesforce.com servers suffer 38-minute outage

CRM Software-as-a-Service (SaaS) provider Salesforce.com was down for about half an hour on Tuesday, January 6.

Salesforce.com representatives declined to comment on the failure but directed inquiries to the site trust.salesforce.com, which provides information on the security and system status of Salesforce.com. According to that site, “Starting at 01/06/2009 20:39 UTC [Coordinated Universal Time] a core network device failed due to memory allocation errors… This resulted in a full service failure for all instances.”

The memory errors were also present on the “failover system,” meaning the Salesforce.com team had to institute manual recovery steps. Most services were restored by UTC 21:17.

While a number of real-time “tweets” and blog posts bemoaned the system failure, it's not expected to hurt Salesforce.com in the long run.

“When you build an infrastructure that's that large and complex, it will never be 100% uptime initially,” explained Denis Pombriant, managing principle of Beagle Research Group. “Outages like this don't happen very often, and I regard them as learning opportunities that companies like Salesforce need to leverage to improve their infrastructure. Is it acceptable? Probably not, but I have to compliment them on the relative transparency they have offered for customers and providing insight into the workings of the infrastructure.”

Salesforce.com updates trust.salesforce.com regularly with all the steps that the company is taking to fix issues.

“If there were a vendor that was perfect, I suspect that it would have more of an effect, but there isn't a vendor that's perfect, and no competitor is perfect,” Pombriant continued. “Viewed in the context that it's an evolving technology, it appears that Salesforce resolved the problem effectively.”

Salesforce's clients include Alltel, Dr Pepper Snapple Group and Avis Budget Group.