Yesterday's catastrophic outage at Amazon Web Services (AWS) has left the industry pondering whether the world has become too reliant on the giant, which accounts for approaching half of the public cloud market.
AWS blamed the mega outage, which lasted five hours, on glitches with its S3 storage service in its Northern Virginia datacentre.
The failure took out vast chunks of the internet and even left Amazon unable to update its own AWS status dashboard. Adobe's cloud services, Amazon's Twitch Docker, GitHub, iFixit, Kickstarter, Slack and Yahoo Mail were all known to be affected, as well as various Internet of Things devices.
Although the outage should not be taken as a black market against public cloud, it will invite questions over whether AWS customers are too reliant on the giant, which accounts for 40 pence of every £1 spent on public cloud, said Bob Tarzey, services director at analyst house Quocirca.
"The first thing to say is this is news because it's not that common a thing to happen at Amazon," Tarzey (pictured) said.
"Something serious has happened, and it's bad news. But we can over-react to stories around public cloud outages, forgetting that these issues are occurring on a daily basis at enterprises and that this is a relatively rare occurrence. It'll only be a gift to Azure and Google until they have a similar outage in a month or six months."
"But we are becoming more and more reliant on a single vendor, and that is worth scrutinising," Tarzey added. "If there are services on the internet that have gone down because they are completely reliant on Amazon, you have to ask two questions. Firstly, were they spreading themselves broadly enough across Amazon's infrastructure? And secondly, should you be reliant on a single cloud provider, or should you fail over to a second public cloud platform or use a hybrid mix of in-house and public cloud?
"We should never assume public cloud providers will never go down, and maybe this will make people take stock and ask whether we are expecting too much of a single cloud provider."
Shawn Moore, CTO of web experience platform Solodev, agreed that AWS customers should spread their risk even within AWS, claiming that although the failure impacted an estimated 20 per cent of the internet, many businesses weren't affected.
"The difference is that the ones who have fully embraced Amazon's design philosophy to have their website data distributed across multiple regions were prepared. This is a wakeup call for those hosted on AWS and other providers to take a deeper look at how their infrastructure is set up and emphasises the need for redundancy - a capability that AWS offers, but it's now being revealed how few were actually using," Moore said.
Chris Cox, AWS alliance lead at AWS partner KCOM, said: "AWS has 16 regions, comprising of 42 Availability Zones running over 70 services. Admittedly S3 is an important, core service to lose, but considering the vast scale of AWS, I would say that rather than be a cause for concern, the short duration of this incident should be seen in the context of the otherwise incredible resilience of the AWS platform and their ability to resolve issues in very short timescales. Our advice to customers is to always follow AWS best practice guidelines in architecting resilience into their solutions."
Security firm set to become part of acquisitive Shearwater Group
Distributor merges three northern sites into one new hub in Warrington
Activist investor puts forward five director candidates as turmoil continues at security giant
Nima Green asks what is driving public cloud uptake in Germany