Saturday, March 10, 2007

Distributed vs. Centralized Entities

In terms of computing platform architectures, we have mainly two types: distributed, and centralized. Google is the classical example for centralized processing: users don't do much of computations and all data are stored in servers. On the other end of the spectrum, we have distributed computing platforms, where users will store their own data and do the computations at their own ends. This turf is dominated by the computing devices that we know and love, devices running Linux, Windows, Mac OS, etc.

The interesting point is throughout these years, computing platforms are changing from centralized (the good old mainframes with dumb terminals), to distributed (advent of PCs and domination of M$ apps). The current nascent trend seems to go back to centralized architecture with server-side processing as can be seen on the popularity of web-based applications provided by Google.

If the development of computing platforms is indeed cyclic, then my conjecture will be eventually Google apps have to be distributed logically. In other words, eventually Google has to push applications and maybe, even data storage to user-ends. Google desktop is something I presume as the first step heading the direction of distributed processing. On the other hand, companies that dominated the personal computing arena, most notably Microsoft, are scrambling to centralized, web-based services. As each of these companies move to the other end of spectrum, eventually they will meet in mid-point and that is where things start to get very interesting and highly cut-throat.

As Google becomes more and more application-oriented, it will look more like MS. While MS is pushing itself to be more web-based, it is snapping up Google's market share as well and therefore more Google-ish. The clash of titans will definitely be interesting. The winner will rewrite the rule of future computing platforms, while the loser will be left in oblivion and soon forgotten.

No comments: