The cloud, and cloud technologies, came about not from a sinister attempt by big companies to take over the world, but from our collective demand for computing resources. Without even looking for statistics we know, as consumers or business, that we are demanding a lot more of our computing resources. The amount of data that we want collected and stored is going through the roof – be that transactional data for data mining or the storage of HD videos of the kids. The trend is similar with the amount of processing needed, backups, bandwidth, screen resolution or just about any computing resource that we can think of.
Obviously someone has to build the machines, power them up and keep them going to satisfy our voracious appetite for computing resources, whether that be your corporate data centre, Googleplexes or even your own home network for backups and streaming of media to your Xbox. A lot of the cloud technologies have come about because the provisioning and effort required to provide the demanded resources has traditionally been high and needs to be optimised – not to save costs as much as to make it humanly and technically possible. The days of having hardware engineers being able to hand configure, install and maintain every server are over – we would simply run out of engineers, similar to the concern last century, before the automatic telephone exchange became mainstream, that manual telephone exchanges would employ most women in the world – such was the growth in demand for making phone calls.
The demand is also non-linear and unpredictable and you can no longer count the number of expected concurrent users and plan accordingly. Sometimes undesirable events will affect load and other times the peaks in demand may be of your own creation. Could you sit in a planning meeting and predict the load of your systems by a video of two guys playing with diet coke and mentos? Would you want to spend the money on the off chance that someone may find that video interesting?
So we have a set of technologies for storage, virtualization, commodity hardware, load balancing, firewalls, routing, intrusion detection, backup, failover, energy efficiency, cooling and a whole load of really difficult and complicated bits needed to supply the demand – to our own internals users, external customers or people out in the world that we don’t have much to do with. What the big cloud players have done is take these technologies and their implementations and tried to find a market for them and they are, as they are perfectly entitled to do, trying to sell all that stuff so that they can make as much money as possible.
Those big players are creating the marketing pull for their services and have wrapped up the whole package in a simple phrase ‘The Cloud’. No acronyms, no complicated, copyrighted, patentable, trademarked or protected names – something simple that quite easily (apparently) conveys what they do. The latecomers to the market and those that feel threatened by the big players are making lots of noise about how the cloud doesn’t work – it’s more expensive, it is unsecure, it has compliance problems. Which is just a lot of whinging about the practical issues that we have to deal with in modern computing that are good sticks to poke in the eyes of the big players – while they play catch up. They are missing the point that they should just get stuck in and market their products to the real demands – not to try and position themselves against the hype.
The debate about cloud interoperability is stillborn – it is just plain ‘ol interoperability that has plagued IT for decades. Security warnings, while real, exist in the cloud as well as outside of it (wherever that may be). Location of data, who owns it and which court has jurisdiction is a little bit fuzzy in the modern world of large multinationals anyway – so the problem has to be dealt with cloud or not. SLA’s are just as difficult to establish, cost and mediate in corporate IT as they are anywhere else. Enforcing compliance in a large corporate environment is no trivial matter. Virtually any criticism of the cloud can be turned around back at non-cloud solutions because it is not about the cloud – it is about the unbelievable (but surmountable) difficulty in satisfying the relentless growth in demand for computing resources.
The Cloud, as defined by the technologies, organizations and products to support this demand is here to stay and will continue to be discussed, argued, hyped and marketed. Whether you buy the computing utility from a cloud provider or whether you run multiple data centres in your corporate environment doesn’t matter – and in most cases you will probably use a combination anyway and, even more likely, you do already.
What is important about the cloud and why I spend a lot of time talking about it is because our ability to respond to this increasing is important and imperative. I don’t know many developers who could build a booking system for a one-night-only Michael Jackson concert in Wembley Stadium where booking starts online at a predetermined time. I know that there must be thousands of organizations out there that are turning down business requests for functionality because their nightly batch runs are already twelve hours long and they don’t know how to get more done. We all know the Twitter ‘Fail Whale’ that nearly sunk the organization before it gained traction.
The development of modern IT systems that can respond to the demand is hard. It requires approaches, architectures and skills that are mostly new and sometimes relearned from forgotten lessons. It requires attention – attention from CEOs, vendors, developers, architects, support staff and a whole host of other people – including air-conditioning engineers. If the hype surrounding ‘The Cloud’ is what creates that attention then that is fine with me (it is at least easier than SOA). I don’t own a big data centre that is going to take over the world so I don’t have a lot riding on the success of the cloud as brand – but I do want to help people architect for the cloud.
If you are not thinking about the cloud you are not ready to respond to the demands – so why not join the discussion?
Simon Munro
@simonmunro