With the amazing speed & success of Google, cloud computing has emerged as the leading edge paradigm of computing architectures. The "newness" (or "fogginess", as the cynical banker is apt to think!) of this concept can be gauged from the fact that a Google search "define: cloud computing" results in zero results!

Google's cloud is a network of a huge number (some say, 1 million) cheap servers, each, more or less, as powerful as the average PC. This cloud stores staggering amounts of data which, in turn, facilitates providing search results in a "fraction of a second". Since the cloud is made of "normal" PCs, every time one of them "conks off" or becomes outdated, the Google engineers simply replace it with a newer, faster, box

So, what attraction does this hold for the earliest adaptors of mainframe computing?! The concept is appealing for two reasons-

1. From a business perspective, the promise of hiding the complexities of IT and being concerned only with the services being demanded/delivered
2. From an IT perspective, the flexibility to easily write and run applications that process vast amounts of data

Having been currently involved in a credit card implementation-cum-migration project, I am painfully aware of the complexities in using mainframes/mid-ranges architecture-based solutions as the core technology stack of banks of all sizes. Being a newbie to the MF/MR world, I regularly observe the business and the IT folks sweating out at ungodly hours transmitting humongous data files from one DC (data center) to another and verifying the accuracy of the transmissions

If, instead, cloud computing was the dominant architecture of the banks, they might not be forced to count MIPS usage and get into convoluted ways of "performing an open heart surgery" - the data might be residing in any part of the world on any type of technology platform (most likely cheap, off-the-shelf, hardware & open source software) and the users will transparently access the various services they need. With cloud storage, the need for dedicated data centers will be minimized (if not altogether eliminated) and the IT staff can also breathe easy

The veterans may find some similarities with the original utility computing model of IBM - it will host the data on its servers and let you use it for a fee. However, the crucial difference between the IBM model & the Google model is the open source nature of the latter against the proprietary nature of former.

Google is not the first or the only player on the scene; Amazon launched Elastic Compute Cloud (EC2) in the summer of 2006 and this service is now available to paying customers. Yahoo is trying to catch up by patronizing Hadoop (derived from Google MapReduce and Google File System). The grand daddy of business computing, IBM, has decided to go along with Google and the two are collaborating to offer the technology to universities around the world. On the other hand, the 800lb gorilla of PC-based computing is still rooting for a PC (with Windows & Office) on every desk of the world!

I am sure that some major financial institution will decide to give cloud computing a chance in the very near future and, in all probability, it will one of the public-sector ones - private players are too bogged down with their legacy systems to think outside the box. Hope that I am proven wrong!!