Amazon, Microsoft Cloud Outages No Reason To Avoid Cloud Services
Instead, cloud outages suffered by Amazon, Microsoft and others reinforce the importance of rock solid SLAs and cloud systems that are designed for failure.
"New technology is a tough sport; I call it a full-contact sport," said Ingram Micro CIO Mario Leone. "If you're not ready to accept that you're going to have some issues along the way, you're not ready."
Amazon Web Services suffered a massive cloud outage in April, a major hiccup that knocked some Amazon cloud customers offline for days. Following Amazon's cloud outage, Microsoft's Business Productivity Online Services (BPOS) was hit with a trio of outages that caused delays and downtime for Exchange cloud e-mail.
Those outages put cloud computing environments under a microscope and called into question the reliability and performance of cloud. But several industry experts said the cloud outages shouldn't thwart solution provider cloud plans.
Gartner analyst Tiffani Bova said the fallout from Amazon's outage can be chalked up to poor planning, and is not an indication that the cloud as a whole is unreliable.
"It's not that on-prem is perfect and cloud is flawed," she said. "IT is flawed."
According to Bova, customers that were hit hardest by recent cloud outages didn't design for failure and have the systems in place to survive and outage. Many Amazon customers that were hit hardest during Amazon's cloud outage did not have failover or redundancy in place.
"If you've only got something in one place; shame on you," she said.
Leone said using the same best practices used in the past can help avoid issues if clouds go down.
"I don't think the formula's changed, but you have to apply it more aggressively," he said.
Designing For Failure Designing for failure is a must. Leone pointed out that Netflix was not affected by Amazon's cloud outage, because its cloud infrastructure had been designed for failure.
"If you have one application, on one server, on one network on the San Andreas Fault, you're going to have a problem," he said.
Leone said the cloud will suffer disruption. There will be failures, but "they shouldn't hold you back from experimenting," he said. "Don't let the news take you off of the path of talking these solutions through with customers."
Renee Bergeron, Ingram Micro vice president of managed services and cloud computing, said cloud outages and delivery issues haven't scared VARs away from adopting cloud and brining customers up to speed.
"We have not seen our resellers slow down adoption of cloud as a result of these outages," she said, adding that solution providers that were on the fence about cloud solutions may not move to cloud offerings as quickly, but it's not halting them altogether.
Bergeron added that cloud environments are as stable, if not more so, than their on-premise counterparts.
"An on-premise solution in an IT end user company goes down and it doesn't get the same publicity," she said. "It probably goes down a lot more, actually."
Recent cloud outages, however, highlight the need for strong SLAs, which Ingram Micro has in place to ensure vendor offerings perform as promised. Bergeron said if an end user suffers cloud issues, the solution provider can call Ingram Micro, which will hash out the SLA with the vendor.
"We view that as our role," she said. "That's our responsibility."