Benioff Blames Data Center for Salesforce.com Outages

Salesforce.com CEO Marc Benioff said that ending the outages that have hit users of its Web-based CRM (customer relationship management) application service is a matter of “shaking out” the companys new $50 million data center.

Speaking at a Merrill Lynch & Co. Inc. IT services and software conference in Las Vegas, Benioff said that Salesforce.com was “going to do everything we can” to get its system availability “as close to 100 percent as possible.”

“No company has 100 percent uptime. No utility has 100 percent uptime. But still we are doing it better than anyone else in our industry,” Benioff claimed.

Customers experienced a major outage lasting more than five hours on Dec. 20 and what the company called a “minor interruption” of service on Jan 5.

The most recent outage occurred Jan. 30 when Salesforce.com servers were down between about 10:00 a.m. and 2:15 p.m. EST, according to customers.

Several Salesforce.com customers suggested that the outages were caused by a problem with the API that allows them to connect their back office sales reporting applications to Salesforce.coms CRM system.

“That means that any kind of applications you have interfacing with Salesforce, using their API, you cant connect to them,” said Melissa Caylor, director of IT at First New England Mortgage, who was one of the customers affected by the Jan. 30 outage.

/zimages/6/28571.gifClick hereto read why customers are concerned about Salesforce.coms recent performance problems.

“We have a lot of stuff integrated with them, so we cant get new leads into our system; we integrate our loan origination system, so we cant get to that either. It is inconvenient,” Caylor said.

However, Benioff denied that there was any problem with the Salesforce.com API in an e-mail to eWEEK. “We do not have API issues to my knowledge, and this is not accurate. In fact, that is the first I have heard anything like that,” Benioff wrote in the e-mail.

Bruce Francis, Salesforce.com vice president of strategy, also discounted the theory that API problems were causing the “intermittent access” problems that the company acknowledged, indicating that it was a more fundamental issue.

“If the system is unavailable, so is the API,” Francis wrote in an e-mail to eWEEK.com

Benioff told the audience at the Merrill Lynch conference that the Dec. 20 outage brought down Salesforce.coms system availability to 99.5 percent for the month of December. It had been running at about 99.7 percent for the 2005 fourth calendar quarter, Benioff said.

/zimages/6/28571.gifRead morehereabout the Salesforce.coms explanation of what caused the Dec. 20 service outage.

“Thats still not where we want to be. We have very high expectations for ourselves,” Benioff said.

Benioff suggested that Salesforce.com customers are a little different from eBay or Amazon.com customers who expect service availability to approach 100 percent. “We are committed to that and we are going to do everything we can to get it as close to 100 percent as possible,” he said.

Benioff said that the company will achieve higher availability rates after it has completely broken in its new data center.

“We have a brand new data center that just went in basically on Nov. 1. And the two brief outages that we had were around shaking out our data center and figuring out some of this new hardware that we have and how our system is running,” he said.

The data center represents a “great new architecture and we feel very committed to the success of that architecture,” he said.

Editors Note: eWEEK Senior Writer Renee Boucher Ferguson contributed to this story.

/zimages/6/28571.gifCheck out eWEEK.coms for the latest news, reviews and analysis about customer relationship management solutions.