Bad Things Good Services
Cloud Society: when bad things happen to good services
Finally, cloud computing has had the reality check it needed, says Jon Collins
Cloud computing’s been in the news, and it’s not been pretty. First Amazon’s AWS service was rendered inoperable [LINK: http://www.forbes.com/sites/reuvencohen/2012/07/02/cloud-computing-forecast-cloudy-with-a-chance-of-fail/) at the beginning of the summer, and then GoDaddy’s hosting services (along with thousands of web sites) were brought to their knees (http://www.dailymail.co.uk/sciencetech/article-1379474/Web-chaos-Amazon-cloud-failure-crashes-major-websites-Playstation-Network-goes-AGAIN.html) by those dastardly chaps at Anonymous. Both nasty, both, might say the naysayers, both illustrations that the cloud can’t be trusted.
But such examples could be exactly what was needed. Many observers (including myself) have long argued that the cloud should not be trusted, not 100%, any more than you would 100% trust a car, or a contract, or a person in authority (and that’s relevant, given that some outages with Google services have been put down to human error).
Trust should never be an absolute, and neither should it be the only factor on which to base decisions. With the best will in the world, cloud service delivery also requires reliable bandwidth (which is frequently not the case, even in fixed locations), simplicity of administration and control of costs. None of which are guaranteed, nor necessarily anyone’s fault when things go wrong.
There’s simply no such thing as the perfect service – one which can be expected to just work, to deliver on requirements in all circumstances. Architectural decisions – that is, deciding where to locate technology to give the best results – are predicated on the fact that imperfection is part and parcel of the whole technology landscape. Falur prevention should be built into the architecture – if a service cannot be trusted, some kind of fall-back needs to be in place.
Indeed, the only people who were ever saying that cloud computing is ‘The Answer’ as opposed to ‘one potential service delivery mechanism that needs to be weighed against, and integrated with other options’ were, you guessed it, marketing executives that worked for companies promoting the benefits of cloud computing. They’re not wrong that benefits exist; they are in their rights to ignore or downplay other choices. Technology buyers and their non-tech-literate bosses would be naïve to think otherwise, and should plan accordingly.
So, yes indeed, read stories about this service or that service going out of action for a while and consider them carefully, not with a sense of disappointment but relief that a veil has been removed from the eyes of the decision making process. With eyes wide open, organisations can start to make the right decisions about where and how to integrate cloud services with their own in-house kit, to deliver a set of capabilities which – even if it can’t be 100% trusted either – can most closely meet the needs and all-too-imperfect realities of the business.