If you look at the evolution of computing architecture, you can see why he makes that claim. Over time, network connections arguably have become more important as computing has gotten progressively more decentralized. In the mainframe era, wide area connections were crucial, but in-building networks generally were not.
The era of client-server computing created the need for local area networks. The era of Internet computing radically decentralized network end points, created a need for house-area networks and simultaneously boosted the vao
Some think the next era of computing will be profoundly driven by mobile computing, which will again emphasize wide area connections. The ubiquitous "radio tails" are crucial to support nomadic computing, of course, but what makes mobile computing different is ubiquity.
Some of us also would note something else: network connections have steadily become more important for ever-larger numbers of end points. "Connected life" doesn't mean much when a user loses their connections. But the connections are valuable largely because, over time, the role of third party applications and devices also has grown.
It is true that network connections are essential. It also is true that value is shifting away from the connections to the applications and devices. Some might say Kerravala focuses on the first trend, while others might focus on the second and third trends.
In other words, there is not a contradiction between arguing that network connections will be the foundation for all coming waves of computing, and also that networks increasingly are mostly "dumb pipes." That is not to say there are not applications and services embedded in the network. It is to say that, over time, more of the valuable or essential applications are provided by third parties.
No comments:
Post a Comment