Do you really trust the Cloud?

Date posted
6 September 2012
Reading time
5 Minutes
Stephen Mahood

Do you really trust the Cloud?

 

 src=

One of the main reasons why many IS experts steer clear of cloud computing is a lack of trust in its ability to deliver what the business needs. We IT professionals are accustomed to being in complete control of our own services, and often look at our creations through rose-tinted glasses (ignoring the sticky tape holding it together!). The thought of losing control of IT services to a cloud provider is often difficult to come to terms with.

Cloud has often been compared to the electricity utility. We trust our electricity providers to pipe power to us wherever and whenever we need it. We pay for what we use and for all we know (or care!) it could be generated by an army of hamsters! There are occasional power outages, but they are so few and far between that we accept them unquestioningly and in any case, they're mostly local and usually don't receive much press (national disasters like the recent power failure in India excepted!). However, when cloud providers suffer an outage, the news spreads quickly and this amplifies the fear around adoption.

And in fact, cloud providers need to earn our trust. We not only consume their service, but we also entrust them with our very precious data. There are plenty of ways to protect this data just as there are safety nets for electricity providers, like using a generator to protect against an electricity outage. For example, we can replicate our data across multiple clouds or providers to protect against failure of one provider, or we can replicate data on-premise and back it up using traditional mechanisms.

It's more difficult to do this with SaaS, but there are ways to offset the risk. One way of doing this is by engineering a very resilient solution. The Amazon S3 storage platform, for example, is designed to provide a durability of 99.999999999% - in other words, it's designed to suffer a concurrent loss of data in two facilities. It would be difficult and expensive to build such a platform on-premise. But it's nonetheless difficult to trust all our data to such a platform: we simply do not know what is going on behind the scenes, and this reduces the level of trust.

For now, the only way of overcoming this lack of trust is to rely on experience. Initially it's best to use cloud platforms for less critical solutions, while at the same time providing adequate recovery measures. As we grow in experience, our trust and confidence will also grow and we will eventually get to the point where we treat cloud computing just as we do any other utility. This might still be some way away.

About the author

Stephen Mahood