Amazon Web Services Outage Shows That, In The Cloud Wars, There's Always Gain In Someone's Pain
Tuesday's Amazon Web Services outage wreaked havoc across the Internet for hours, interrupting enterprise services and stalling operations at businesses globally. But in the cloud wars, there's always joy in someone else's pain—especially when that someone is the 800-lb gorilla in the industry.
The cascading effects of the AWS S3 storage failure at its data center in Virginia, which started around 12:40 pm ET and was fully resolved some four hours later, left a trail of disgruntled customers, and customers of customers.
It also provided fodder for rivals.
[Related: The 10 Biggest Cloud Outages of 2016]
"This is exactly what Google wants," said Jamie Shepard, senior vice president at Dallas, Texas-based solution provider Lumenate, a Google partner.
While the outage may not severely hurt AWS in the long run, those types of snafus can drive customers to the competition.
"This is going to force a Google conversation. They just handed Google thousands of customers," Shepard told CRN.
Smaller rivals also stand to gain from Amazon's stumble.
"We're already seeing benefits from this outage," said Neill Reidy, director of cloud computing at Evolve IP, a boutique cloud provider.
Organizations that implemented solutions on their own with AWS, now are down and realize they need a white-glove approach rather than a hyper-scale provider, have been calling Evolve IP, he said.
There's no denying that Tuesday's outage embarrassed Amazon, which owns about one-third of the global public cloud business, Reidy told CRN.
"As an engineer, our most proud accomplishment is uptime. That's really one of your bragging rights," he told CRN. "When you look at a status page today and you compare Azure with Google Cloud and AWS, there's a really big red blotch on the AWS page that you don’t see for Azure or Google."
Time does heal some wounds, he said.
"Tomorrow is a brand new day with a brand new status page. I think an outage like this is embarrassing for a week, and then it starts to vanish away after the sting is gone," Reidy said.
But not everyone has short memories, said Simon Margolis, director of cloud platform at Google partner SADA Systems.
"In my circles, people still talk about the 2012 Netflix outage that was caused by AWS. It absolutely remains in memory. It has to do with building a brand of reliability and trust, and this is becoming something people think about when moving to the cloud," Margolis said.
A potential customer exploring proofs-of-concept for Google Cloud Platform and AWS was in his Los Angeles offices just hours after the outage was first reported, Margolis said.
"They said based on what's going on right now, I don't think we can really think about AWS," Margolis told CRN.
Marketers for Google and Microsoft partners are also trying to make the most of the situation. SADA's marketing team was trying to post promotional material on HubSpot later Tuesday, but the inbound marketing platform was also down because of the outage.
"I'm an engineer at heart so I don’t think about the marketing impact, but it's definitely there," Margolis said. "It highlights a technical difference" in how Google engineered their cloud versus the competition.
While technologies deployed behind the curtain at hyper-scale providers are highly guarded secrets, Margolis said his anecdotal experience tells him there's a fundamental difference, likely the result of Google building a cloud from the ground up, originally for its internal use, rather than in a piecemeal fashion.
"[Google] built it so that issues like this don’t happen," he told CRN. "You can't have an outage in one region that takes down the whole cloud. That's a core theme when we talk to our customers."
But hyper-scale outages also challenge the public's overall confidence in the public cloud, to the detriment of all providers. That gives ammunition to competitors that focus on private and hybrid solutions.
All IT teams can do to prepare for outages like the one that hit AWS is "duplicate and triplicate everything with other cloud providers, with the cost and operating challenges. Or stay in control and own your data," said Ariel Maislos, founder of Stratoscale, an Israeli data center operator.
Companies that rely on cloud need to consider either using multiple clouds, and doubling their costs, or implementing hybrid environments, Maislos told CRN.
"Cloud companies are technology powerhouses, but you have lost all control when you give them the keys to the castle," Maislos told CRN. "There is nothing to be done when hit with these situations."
Reidy, of Evolve IP, said outages should remind customers of the importance of architecting redundancy, failover procedures and disaster recovery solutions directly into their environments. They should also avoid the common mistake of forgetting to actually test those procedures by simulating various types of failure.
And it's important to remember the true victims of a big hyper-scale outage, Reidy added.
Ultimately, those hit hardest are the midsize companies for whom "S3 and Amazon in the East availability zone are all they had," he said. "They get blindsided, stuff's not working, and they're at the mercy of when the power goes up again. That's probably the majority of companies affected. Not Netflix and Pinterest, but mom-and-pop midmarket companies. They lost money today."