Cloudbursting - the key issues with cloud mobility
At the present time, it is fair to say that I am not a fan of cloud. Let me be more specific here: the dense, brooding, deep-grey-on-black cloud that the BBC website assures me will deliver a monsoon upon my wedding this coming Saturday. What’s that? Not that kind of cloud, you say. Ah, but definitions of cloud are frequently so vague, so all-encompassing, so – well, nebulous – that if you’re not careful you can begin a conversation talking about the opportunities for network operators as PaaS providers and end it with a discussion on weather fronts in the south of England, having passed through each and every web-based service en route. And that is the primary reason why those in the enterprise space have tended to sigh in exasperation whenever a new “consumer cloud service” is launched, a sigh which essentially translates to: look, we know what cloud services are; we’ve had that debate. So why are you coming along and trying to muddy the waters by calling every wretched online service a cloud service? They’re right, of course, which is one of the reasons that, in our recently published report on cloud services, we attempted to provide a more rigorous definition that would hopefully satisfy most parties and exclude those products which are essentially just streamed services and don’t offer storage. (For the record, we feel that it is also sufficiently rigorous to exclude weather fronts.) In the course of writing the report, it became apparent that there were a number of fundamental concerns as to the efficacy of a cloud-based solution, particularly – though not exclusively – from an enterprise perspective. Two of the foremost were security and connectivity. I would argue that the first of these needs to be understood and explained in relation to the alternatives on offer; when viewed in this light, the concerns may be alleviated somewhat. But the second – oh, we’ll come to that in a moment. From an enterprise perspective, data security is paramount: the thought of placing your data in the hands of a third-party, in a server remote from your premises, may create a feeling of profound unease. However, what the enterprise needs to ask itself is “Are the security procedures that I have in place in-house any more secure than those offered by the cloud service provider? How satisfied am I with those existing arrangements? Do I really have full cogniscance as to the flow of data within my corporation?” The answer to this last question will frequently be no, or if it is yes, then the person answering it is not in full possession of the facts, because it will subsequently transpire that much of the data is sitting on a laptop which has been mislaid on a train at Birmingham New Street. Cloud service providers are anxious to demonstrate the extent of the security arrangements; the extent to which they have mechanisms in place which dynamically determines – according to your specifications - which corporate data can go to the cloud, who can access it and when they can access it. This is essentially because they need to show you that their own data loss protection (DLP) measures are better than everyone else’s and that therefore you will choose it. In this respect, competition is a wonderful thing. Conversely, your in-house IT solution – historically a de facto closed shop – may not have the same imperative. So far, so good. We then come to connectivity. As any fule kno, you need connectivity to be able to seamless access your synched, cloud-based content. However, even leaving the much vaunted issue of network coverage out of the equation (and I have vaunted it pretty much in this blog in times past), there is the blunt fact that your network has moments (moments which can last for several hours or even several days) when coverage disappears. This can be due to natural causes; to light-fingered scallywags breaking into the operator’s office and stealing essential equipment; or it can simply be “unexplained” (a catch-all response embracing both “we don’t know” and “we don’t want to tell you”). By a delicious irony, while I was typing up the text on data outages for the report, and compiling a list of recent data outages by UK and US networks, I was (unknowingly at first) subjected to one. The twin results of this data outage were that, firstly I bought a pair of trousers online having failed to receive the text message from my wife-to-be that she had bought precisely the same pair in-store, and secondly, that that evening I spent £15 on taxis to and from a cricket ground having failed to receive a text from the skipper than our game had been postponed. (I also failed to receive a text that would have told me that the team had all convened down the pub instead.) I am still in ignorance as to the cause of this particular outage - all their advisors were currently busy, and I had better things to do than listen on hold to Madonna for more than ten minutes – but the fact remains: the outage occurred, and had I been dependent upon connectivity for said network for a cloud-based service, then I would have been in a spot of bother. Hence my suggestion to the network operators that if we are to pursue this migration to the cloud, then a little more redundancy in the network might not go amiss. On which note, I shall return to my funk about the weather forecast.