It had to happen. After a couple of years where every IT and communications company on the planet has been carefully weaving “cloud computing” into all of their marketing materials, mission statements and in some cases even renaming their companies to jump on the bandwagon, we are now seeing analysis decrying the term as meaningless.
The IT industry in general of course loves this kind of thing, our history is littered with phrases invented by clever people to explain fairly intangible concepts to the mere mortals who own the cheque books and determine company valuations. Once enough commentators and analysts use a new term, momentum builds around it until it acquires a life of its own. This is the point it starts to evolve fairly randomly, becoming more and more divorced from reality with each marketing $ spent by aspirants trying to position their company or technology as the ultimate embodiment of the buzzword even if they started out as something else entirely. Amongst cynical technologists at the implementation coal face, the place where any reality gap is painfully visible, these phrases are often dismissed as memes fairly early in the cycle.
Whether or not you personally regard the phrase cloud computing as a mantra or just another buzzword bingo candidate, there has for many years been an inexorable progression from isolated shrink-wrap applications running on local hardware to delivery of applications and data over the network from centralised facilities.
There are a number of key drivers to the process of divorcing the place where programs physically run and the data lives from actual delivery of the application as a service:
- Applications are becoming ever more complex which means that application development, QA and support in a traditional shrink-wrap environment gets more resource intensive with each release. This is having an exponential negative impact on development cycle times and customer deployment cost.
- Network bandwidth in most parts of the world continues to become exponentially cheaper per bit to the point where there are very small marginal costs in delivering new applications remotely. Cost effective network reliability has also improved to the point where the risk of service outages due to communications failures are now similar or lower than those associated with system failures in local deployments.
- Power and cooling for local deployments is becoming more expensive
- Wide availability of open standards based endpoints (e.g. DHTML browsers, SIP audio & video devices) has removed any barriers to universal deployment. Actually this is more of a consequence of technology developing to deliver solutions in this economic space but it is the big key enabler for cloud deployment.
There are also a couple of factors that mean that even if costs and service levels are the same, both application providers and users benefit from their delivery as a service:
- Particularly in the current climate there are good cashflow, tax and business agility reasons to buy services (OPEX) instead of investing scarce capital in building facilities (CAPEX).
- Once application vendors saturate their market with one-off application purchases, it becomes very difficult to continue to generate the year on year revenue growth that their stakeholders demand just by producing more and more expensive to develop follow on products (look at Microsoft’s share price over the last decade for a classic illustration). The application as a service business model breaks this cycle of ever decreasing return on development cost by locking in recurring revenue based on the number of current application users.
None of this is new, in fact what we now call cloud computing has been a more or less continuous progression since at least the mid 90s when deployment of centralised services like web and e-mail on high bandwidth public data connections started to gain momentum.
There are some technology areas that are currently experiencing rapid change, particularly around server virtualisation to deliver infrastructure itself as a service and also big leaps forward over the last 5 years in deployment of open communications protocols, which are starting to drag all of the islands of on-site TDM telephony into the Cloud. These are however all just part of an obvious progression towards using open network technology to provide better resource utilisation when delivering ever more complex applications.
Whatever we call it, I don’t think that this shift in technology consumption is going to decelerate or go away any time soon. I would love to hear what other people think though: cloud computing, is it a mantra or meme in your organisation? And what are the areas where you actually see it having most impact on the ground?