Recently I talked to several corporate executives who claimed last month's Amazon cloud service outage gave them pause. While none were threatening to build in-house infrastructure, the blinders were off regarding the halo that had been wrapped around cloud-based strategies.
They maintain that cloud services offer cost advantages that are just unattainable through on-site data centers, but say now they realize they didn't invest enough in risk mitigation. For instance, one CEO/CFO said he plans to put host his main database with two different cloud services, just in case one goes down.
Such disaster recovery planning is not unlike the traditional enterprise where companies run multiple connectivity lines from different service providers into the data center, in case one suffers a backhoe cut or some other failure. Several executives said this same logic needs to apply to your most critical applications being hosted in the cloud.
Does this affect the attractive cost model of cloud-based services? Perhaps a little. But a multi-day outage would most likely have a more negative impact than a slight rise in operational expenses. This is where the CFO comes in. The ball is in your court, even if you have an IT staff, to plan what-if scenarios that fit within your risk tolerance profile. For some businesses, the multi-day outage that some Amazon customers suffered would not be a hardship; for others, an hour of downtime would be too much. This has to be considered even with a well-established service provider.
The executives said they figured the Amazon name would be synonymous with uptime. But as one wise executive said, the brand may be established, but the service is new -- and nothing should be taken for granted.
Meanwhile, this article has gotten quite a bit of attention, and Twitter reactions are mixed -- with some tweets chastising CFOs for not being ready with an outage game plan.