As discussed in my previous post I wrote an essay, a summary and commentary on Nicholas G. Carr's article "The End of Corporate Computing", for my Information Systems course at the University of Cape Town.
The following is the essay, it can be downloaded as a PDF here. The essay is published under the Creative Commons Attribution-Noncommercial-No Derivative Works 2.5 South Africa licence.
SUMMARY AND COMMENTARY ON "THE END OF CORPORATE COMPUTING" BY NICHOLAS G. CARR

In Carr’s article “The End of Corporate Computing” the former executive editor of the Harvard Business Review outlines an industry-internal prophecy of the end of computing as it exists today.

Carr compares the current computing industry, i.e. the use of desktop and laptop computers, to the electricity industry of the early 20th century. According to Carr the use of isolated and separated workstations with redundant resources and large fixed costs can be directly compared to the early ages of electricity production, in which companies with large factories would have their own internal power plants to generate electricity. But as time progressed and alternating current was discovered the centralization of power supply became evermore worthwhile. Initially the technology was unable to cater for such a scenario as producers were unable to transport electricity more than a mile using direct current.

Companies therefore had to buy expensive power production equipment, employ electrical engineers and maintain these internal plants, creating massive redundancies and overcapacities. But in the beginning of the 20th century Samuel Insull realized the potential for centralization, saving companies massive fixed costs and creating a new industry. He went to found the first large power utility, Commonwealth Edison, which would go on to provide electricity to large companies all over. This shift from electricity being an asset the company generated to a service it purchased meant power went from being a fixed cost to being a variable cost. Companies would now pay for what they used and no more.

Another advantage of this centralization was the economies of scale. With alternating current there was virtually no limit to the expansion of power utilities, as they could provide electricity over vast distances. The limit was purely their production and selling capacity.

What Carr is playing to in his reference to this paradigm shift in the electricity industry is the rising of a new industry in the computing world. Software as a service, cloud computing and worldwide high-speed telecommunications infrastructures are allowing for a possible future centralization of computing power.

As I have discussed in my own essay, written in 2008, “The Future of Personal Computing”, a worldwide shift to centralized computing is indisputable.  We are facing an industry of computing suppliers. Since Carr’s publication in 2005 Amazon has introduced it’s cloud storage solution, Google has introduced Google Apps, allowing the use of word processing, presentation and spreadsheet software on the web and is also developing an open-source operating system based solely on web-based services (Chrome OS).

Carr predicts that computing “utilities” will own huge data centres which will run virtualized processes for their clients, who access these from pure terminals. All that will be necessary to access are the basic input devices, a screen, keyboard and mouse, while all the actual processing hardware will be based in the remote data centres. This allows for the distribution and maximization of processing power, elimination of redundant data and capacities, as well as huge cost savings on behalf of the companies, who will no longer be required to buy desktop units, servers, employ maintenance staff and constantly upgrade this infrastructure. The “utility” will own the terminals, which will be leased by the company. Utility engineers will resolve most issues remotely or alternatively an engineer will visit the company to resolve hardware issues with the terminals or communications infrastructure.  For companies computing will no longer be an asset and painful distraction from daily business but a service purchased similar to electricity. They will henceforth only be paying for actually used capacity

I largely agree with this prognosis but actually go a step further as to predict a world in which we will face unified terminals available everywhere from which we log onto our online presences, completely sidelining existing operating system manufacturers such as Microsoft, Apple, Cannonicle, etc. Of course the data centres we run our entire computing world of will be running on operating systems and there will undoubtedly be various “utilities” providing these services. As Carr mentions possible candidates as leading centralized computing suppliers are Google, IBM, Microsoft and Amazon, amongst others. Personally I see Google and Microsoft as the ideal candidates as a market leader, but there will be many competitors once the trend begins, including many new startups.

For this centralization the new “utility” will require massive capacities: large state-of-the-art data centres, huge software engineering departments, hardware production and distribution infrastructure, as well as high-speed connectivity to internet backbones. These prerequisites basically put existing, established Software as a Service (SaS) and Cloud Computing providers in optimal positions to put up shop as the future computing providers. These are precisely Amazon, Google and Microsoft.

Another issue to consider is what impact all this will have on private computing and especially high-end computing needs, such as video editing, games and other processor-heavy tasks. Will cloud computing solutions be able to offer intensive graphics and processing power for these uses and just as importantly will the infrastructure be able to support these data streams?

Every sign in the industry is indicating ventures in the direction of massive centralized computing. Likely large companies will take the lead by beginning to outsource their computing as described by Carr, paying as they go for the resources used and eliminating fixed costs. Eventually small to medium enterprises (SMEs) will follow and then private users. But until Windows, Mac OS, Ubunutu and all the other operating systems will disappear, until hard drives, motherboards, CPUs and graphics cards become completely discarded from offices and homes another decade or more is likely to pass. Not only are we (South Africans) currently missing the necessary infrastructure to support the connection uptime and stability, as well as the bandwidth required, but the public mindset is not set for such a massive shift. People will not be able to comprehend the elimination of desktop computing overnight, nor will companies be able to implement this. As we stand at the moment no solution is in sight that can service all the needs as Carr describes.

Carr does not outline a timeframe for his vision, but it is certain that while many requirements must still be met this is the future of computing.

References:

  1. Nicholas G. Carr: “The End of Corporate Computing” (MIT Sloan Management Review, 2005).
  2. Michael Hubbard: “The Future of Personal Computing” (http://mikehub.net/show.php?id=82, written 2008, published April 2010).