Hi folks. Long time no see. I've been swamped with work in recent weeks without a moment to switch into blogging mode but a couple recent events of highly publicised "cloud failures" have prompted me to briefly take to the soap box.
The events I'm specifically referring to are the disruptions to Gmail and AWS services. What do these news reports mean to the the market? Have we suddenly stumbled on to the realisation that the cloud is too risky for enterprises to adopt? Will I now be advising my clients to avoid the cloud route?
Well my answers are "not much", "no" and "no - not at all". The reports of service failure by media channels are dramatised to draw readers, not necessarily reflecting a balanced assessment of the actual maturity of today's offerings. What the reports do serve to remind us, however, is that the adoption of cloud does not promise a perfect utopian world. Even well-known providers will have service interruptions and enterprises need to have plans in place for when they do. As Amazon architect Jinesh Varia is well known for pointing out - "design for failure and nothing will fail". Being a pessimist about the reliability of the cloud services that you combine and integrate and ensuring a good level of exception and error handling will help shore up the vulnerable points in your architecture. Furthermore, on the non-technical side, business continuity planning should not be seen as an optional step when it comes to deploying cloud services into critical process areas of your business.
Of course services fail and they do so at even higher rates in a local IT department near you today.