Amazon cloud offering hit by 47-minute service disruption

According to a log of service interruptions published by the company, the problems were experienced during a 47 minute period on Tuesday.

"Log-in functionality for the AWS Management Console was impaired between 11:45 AM and 12:32 PM PDT (Pacific Daylight Time) [on Tuesday]," the status update said. "As a result, customers trying to log in to the AWS Management Console during that time experienced errors. AWS customers who were already logged in to the Console were able to use the Console normally, and AWS's core services (e.g Amazon EC2, Amazon S3, Amazon RDS, etc.) largely continued to operate normally."

According to a report by TechWeek Europe, Amazon's main shopping websites for US and Canadian shoppers also experienced downtime lasting approximately 25 minutes.

The New York Times, Microsoft and Google are among the other organisations to have experienced recent system outages. In the case of Google, a five minute outage last week saw an approximate 40% drop in anticipated internet traffic for the period, real-time analytics company GoSquared calculated, according to the TechWeek Europe report.

Most IT outages (60%) occur during summer months, John Gentry, vice president of marketing at Virtual Instruments, said last month. Gentry said that the reasons for this could be put down to factors such as IT experts taking holidays, the stress on mobile and internet networks caused by increased travel during the summer, and businesses electing to migrate from data centres to cloud infrastructure services during the summer months 

At a roundtable discussion event on IT outages hosted by Pinsent Masons, the law firm behind Out-Law.com, in June, expert in IT contract negotiations David Isaac of Pinsent Masons said that he believed "human error and frailty" were largely to blame for IT outages happening. Isaac said "poor governance, probably bad planning, and bad communications" led to outages.

IT expert Iain Monaghan of Pinsent Masons has said that, in an IT outsourcing context, improving IT governance may mean conducting "a wholesale review of existing relationships and contracts". He said that although this can represent "a massive and complex task" it can help companies, especially regulated businesses such as those in financial services, escape regulatory penalties.

"When IT failures emerge in any industry it is always tempting to look for a simple, single explanation, but while the subsequent investigation might find one proximate cause – a coding error or aberrant process – its focus often turns to the context in which the error occurred: why did the agreed procedures not prevent the introduction of poorly-written code or trigger a fail-safe routine that would contain the effects of the aberrant process? Why did the business not foresee the risk and adequately guard against it? This approach is reflected in the attitude of regulators," Monaghan said.

"If governance has been implemented effectively, the failure can be seen to be caused, as it often is, by an inadvertent human error or lapse in concentration against which no practical preventative measures could have been put in place without disproportionate expense. Assuming there was some external involvement in the failure, the question then becomes one of contract rather than governance: who agreed to take the risk of such errors and how was financial responsibility apportioned?" he said.

"Contracts should be regularly reviewed to ensure they remain fit for purpose," the expert said. "If current performance has drifted away from what was originally intended, there is a choice: change the contract so that it reflects current practice or, if current practice is not acceptable, agree an updated contract that reflects the parties' intentions. Both these solutions require the agreement of suppliers but maintaining a constructive dialogue with suppliers is one of the keys to successful governance. Distressed contracts should be tackled as quickly as possible."