07 April 2009

Cloud Computing 2.0

While even the Wall Street Journal has a hard time defining it, basically, Cloud Computing refers to a computing environment when important parts of what you need to run your computer, often including both applications and data, are stored remotely from the computer that you are working at (i.e. in a cloud).

Hotmail, blogger software, and Google docs are operate on cloud computer paradigm. The computer from which you access the software is a dumb terminal, and all your important data and applications are accessed over the Internet.

This is the second time that the dumb terminal became dominant. In the last 1960s through the early 1980s, serious computing was done through dumb terminals that time shared the resources of a mainframe computer (before then, there were punch cards). The economics driving the dumb terminal system at that point was a shortage of computing power. Only a mainframe computer had the processing capacity to handle reasonable tasks.

When computer chips got cheaper and the technology for storing data became more compact, we had a PC revolution that began in the late 1970s and early 1980s. A PC provided nearly the computing power available at a dumb terminal sharing time on a mainframe, but you had that processing power and storage capacity all to yourself. Applications (i.e. computer programs) and data were both stored upon and run from your hard drive.

By the early 1990s, local area networks became the standard way to run a small office and at the University of Michigan used a similar system across its vast university. Applications remained on desktop PCs (typically with a standard installation at each of the machines in the system), but data was shared on a server on a hardwired (and later wireless network) and accessed from any machine on the system.

This year, I have come full circle. My new law firm operates in a Cloud Computing environment. All data and applications are accessed via the Internet. Local PCs act basically as dumb terminals, while the heavy lifting is done in some distant hurricane proof bunker. I can access all my documents, e-mails and applications from any computer, anywhere in the world, with the dumb terminal software downloaded on it, and can download that with the proper passwords onto any computer anywhere with internet access.

The transition process is painful, time consuming and requires considerable retraining and expertise, but it comes with a big payoff. One payoff for my firm, which has many offices in Florida, is that the system Hurricane proofs their computer systems. And, it provides a practical way to coordinate everyone in a sprawling multi-office law firm with 500 attorneys, many of whom travel a great deal for business, and need to be on call for clients, even when they are at home or on vacation.

Unlike the first round of dumb terminal computing, call it Cloud Computing 1.0, the current round is not driven by hardware limitations. Fast internet connections enable the change, but don't precisely drive it. The computers we use as dumb terminals in our Cloud Computing environment are probably capable of designing space rockets and nuclear bombs, editing video and managing the nation's airline scheduling system. Instead, in addition to the other benefits, Cloud Computing 2.0 is driven by technical support issues.

In any decent sized office with networked computers, the payroll of the in house or outside IT consultants dwarfs the costs of buying the hardware and software itself. A PC environment introduces hosts of variations in how systems are configured which is hard to support efficiently, and makes it much harder to have functional cyber security and reliable backup practices. Cloud computing gives the IT department the power to limit the ability of regular users to alter components of the system that matter from a technical support perspective, and to maintain high levels of security and good computing practice. For example, if an employee loses a laptop, that user's password can be changed in an instant and the laptop itself has no sensitive data on it.

Similarly, important security updates in e-mail programs can be handled at the IT department level, rather than being a personal responsibility that hundreds of individuals must all remember to carry out. Neverending upgrades of software can be done quickly and easily in a single transaction, rather than in a sprawling, never quite completed process, that must go to every single desk.

The end result is that there is far less back office IT labor required, and the system is cheaper to support, while offering more functionality and coordination.

If Cloud Computing becomes a dominant model of business computing, as it seems likely to be (and perhaps for personal computing as well), then we may see some minor but important changes in consumer PCs as well. Large hard drives and super powerful processors will be less important. Wireless connections will need to be universally good. Backup devices like external hard drives, CD drives, memory sticks and memory cards will become less important, as data is consolidated in internet accessable servers rather than locally. The reduced hardware demands and ease of travel that cloud computing facilitates will encourage smaller PCs that are easier to carry, and/or, widespread internet cafe style shared terminals. Big screens and speaker systems, possibly as part of a home theater, will replace the computer itself as the important pieces of hardware.

1 comment:

Anonymous said...

Very informative. Elaborated on some ideas about "the cloud" in a very precise way that I had just begun to think about. Minor quibble: the typos can be a little distyracting :-)