This week’s AWS Simple Storage Service (S3) outage has served as a stark reminder that all cloud services can and do fail.
Increased error rates affected AWS’s US-East-1 region, taking some well-known customers offline and slowing others down. The outage was not a one-off for the major public cloud providers. For instance, Office 365 users have been hit on several occasions over the last twelve months. The results can be costly for any businesses relying on their services.
This does not mean it’s time to ditch the public cloud. It is though, a wake-up call though to consider how to avoid putting all your cloud eggs in one basket.
The services provided by the likes of AWS and Azure are helping us all do business better online. Outages do happen whether they are caused by people, software or hardware and whether it’s your public cloud provider or your hosting company.
When building your cloud systems it is important therefore to remember the adage ‘design to fail.’ For instance, if those who were worst affected by the AWS S3 outage had spread their services across multiple AWS regions, they would have been in a degraded state but not offline.
Whatever cloud provider you are using – public, private or a hybrid combination – you should always consider spanning your services across multiple locations in order to guarantee access to and the ability to restore your data and systems in the event of the failure of one connection.
As Gartner said in its Business Continuity Management (BCM) report at the start of this year, “Businesses that fail to create an effective BCM program could risk catastrophic business disruptions.”
Investing in business continuity might cost you now but it could save your business later.