Trending: Japanese aircraft carrier lost at Battle of Midway is latest find by Paul Allen’s research vessel Petrel
Charu Jain, (L) CIO of Alaska Airlines, and Janice Newell, CIO of Providence St. Joseph Health, share insights at the 2018 GeekWire Cloud Tech Summit. (GeekWire Photo / Kevin Lisota)

Most enterprise software is delivered as a service these days because paying an expert to manage a complex application just makes too much sense. But that service has to work, and based on the experiences of two prominent CIOs speaking at our recent GeekWire Cloud Tech Summit, some prominent SaaS companies still have a ways to go.

“We’ve learned to ‘trust but verify’ going forward, even with some of the large SaaS providers,” said Charu Jain, CIO at Alaska Airlines, on a panel discussion during the event that also featured Janice Newell, CIO of Providence St. Joseph Health and 451 Research analyst Nancy Gohring. Uptime guarantees and the service-level agreements that enforce them are table stakes for a lot of large SaaS deals, but despite their promises, not all vendors follow through on crucial services like disaster recovery.

Newell shared a painful tale of an unnamed SaaS provider (she ducked my request to name and shame during the Q&A session) that was providing cloud-based speech recognition services to Providence St. Joseph’s thousands of doctors. Doctors have been using voice recorders to take notes for decades, freeing their hands to examine patients or monitor equipment, and when that service went down, “it was not a pleasant day,” Newell said, in quite the understatement.

Turns out that SaaS provider had been hit with the WannaCry ransomware, which disrupted tech operations around the globe last year before it was contained. Getting hit with such an attack is bad enough, but in this case, it was even worse: “we found out that one of our SaaS providers, whom we thought had all of this great disaster-recovery capability, business-continuity capability, we found out they in fact did not have this capability,” she said.

The service remained down for an astonishing 30 days, during which Providence St. Joseph moved its speech-recognition users back onto its own hardware.

Alaska hasn’t run into anything quite that bad, but it has seen enough to require that SaaS vendors clearly demonstrate their disaster-recovery capabilities and plans for redundant services should something go wrong, Jain said. “Some of our larger SaaS providers are on their own journey” figuring out how to deliver distributed services with scale and reliability, she said.

Watch the full video of this GeekWire Cloud Tech Summit session above, and see additional coverage here.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline


Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.