Published on the 07/06/2016 | Written by Beverley Head
Confirmation that cloud users need a belt and braces approach to business continuity was delivered over the weekend with a high profile AWS outage…
Reportedly triggered by a power outage to a data centre during Sydney’s massive storm on Sunday, AWS users signed up to receive a service out of the local AWS data centre found operations grind to a halt for almost two hours.
Climate change science predicts that more extreme weather events are likely – Sunday’s so called “stormageddon” is unlikely to prove isolated, with impacts possible on both communications and power supplies to enterprise computer users.
It’s not the first warning Australian enterprise has had of the need for cloud aware business continuity plans; last month Telstra’s network collapse left many people unable to connect to cloud based services for hours.
iStart asked Amazon Web Services about the outage; in particular why Sydney customers were not re-routed to other nodes of the Amazon cloud and whether AWS would have to pay penalties under service level agreements.
A spokesman for the firm responded only to say that the Service Health Dashboard had updates. “We don’t comment beyond that,” he said.
The dashboard reveals very little information apart from confirming that the EC2 Sydney cloud experienced problems over the weekend, which had knock on effects for other Amazon cloud services such as Redshift, Elastic Beanstalk, the Storage Gateway and Cloud Formation.
The company acknowledged that it was a power outage that impacted the system, which does call into question its own approach to uninterruptible power supply. After six hours the majority of issues had been resolved, and the system has been working well since.
For companies which bet the business on public cloud, Amazon’s collapse is a reminder that business continuity planning can’t be easily outsourced or abrogated.
Olive Huang, Gartner research director, said that; “People go to the public cloud very ill prepared.” Where enterprise IT users with on premise solutions crafted disaster recovery and business continuity planning, she said that very often cloud solutions were bought by business units other than IT, and less attention was paid to business continuity issues.
“Only when things happen someone needs to clean up,” said Huang.
She acknowledged that in the public cloud; “you can have redundancy – but it costs money”, which had led some companies to take an “ostrich approach”, with many believing the risks of system failure were low.
She said Gartner recommended organisations negotiate with their cloud vendors the terms of service, though she acknowledged some of the large cloud companies had limited offerings with regard to service level options.
Huang said that the awareness of the issue outside the IT function was relatively low because of the “cloud vendor marketing machine,” compounded by business people’s view that if they were buying from a large cloud vendor, even if the service did go down it would be up again quickly.
But she noted; “if a ship cuts the cable (to the cloud) we are screwed,” and that enterprises always needed a Plan B.
