The wife and I watched The Matrix with Keanu Reeves last night. Great SciFi action, but it will cost me a couple of ChickFlix over the next month.
Anyway, in the movie, Reeves goes through a rigorous training program to increase his reflexes and physical abilities. As his training level matures, he enters the matrix and begins dodging bullets, jumping between buildings, etc. We find ourselves in a high tech action flick and lay back for the ride. But about 1.5 hours into it, we meet a spiritual black woman who runs what appears to be a school for the ESP and mental telekinesis-enhanced children of the matrix. Let’s skip over the how of this motley crew and just ask one question … what is she teaching and what are they learning? … they show us blocks floating up in the air as they “play” with them. Keanu is quite taken by what he sees and leaves to continue his mission.
Later, as he assimilates this new learning, he is able not only to dodge the bullets, but to first slow them down, then stop them in mid-air, and finally to have them fall to the ground.
How did he do that????
Now, I know just what you are thinking. Hey! It’s only a movie, chill out.
But that is not the point. If we are following the basic premise of this movie, the world is now trashed, people live out their physical lives in nutrient pods, and we exist as programs in some pan-computing matrix cluster of processors and memory.
So the way Keanu was able to slow, stop, and drop the bullets was by “on the fly” coding of new software to be executed by the computing matrix to alter the flow of information surrounding him as the bullets were approaching! No superman gig, just a very good programmer. So, should we be issuing capes to our computer science graduates instead of caps and gowns? Maybe.
What happens today?
Today there are many different times when our computers take action on our behalf or that we “write” programs that try to do things to protect us. Some of this is very simplistic compared to dropping bullets, but still, perhaps, directly related.
One example is anti-virus software. In this case, if you have it installed, the software constantly looks for incoming virus (bullets) that it is trained to recognize and it stops and drops them into the bit bucket trashcan. This software is configurable, and to the extent we give it enough connectivity, it can keep itself current and examine all of the I/O to our local computing space. Some of the newest algorithms even look for “virus-like” activity and flag that. In the future, perhaps the software will shunt aside that file (quarantine) and then communicate with a virus-tracking center to see if its internal list is out of date, whether this “suspicious” file can be identified as a problem.
The point is that even today our software is enhancing itself.
And the impact on Gb/s data networks?
As we evolve, our intelligent networks and our increasingly man/machine interdependent society, we are now finding that data communications (i.e. communications that are generated and received by computers) is now dominate over the early leader – voice. Moreover, data is in an early phase of its exponential growth while voice traffic growth is linear and only slowly growing.
Interestingly enough, voice traffic may find itself assimilated into the data at some very near point since the data network will have so much capacity that voice will only occupy a negligible fraction of that systems bandwidth within 10 years (here comes voice over IP, VOIP services).
But from all we have just seen, the growth in data communications may be just beginning. If computers begin communicating between themselves for internally generated purposes it is possible that this traffic will do to current data traffic what it is doing today to voice traffic. Today nearly all computer data communications results from the direct action of a human operator – request a web page, copy a file, send an email, etc.
So, the networks we are building today, even in their grandest vision, typically have their plans based on extrapolations of today’s types of applications and the requests/messages that they generate. The next wave of computer generated and received and acted upon messages will likely require new thinking again.
And it won’t be just today’s case for more bandwidth. Gb/s will be mandatory, WDM too, fully switched. But we may find that the protocol of the network must change as well to allow the communications to occur in ways that are native for the sender and receiver. Computers don’t really operate in terms of IP packets, fibre channel frames, etc. They deal in terms of various organization of memory – cache lines, virtual memory pages.
The next generation of networks will likely be dominated not only by bandwidth but also by communications originated by computers, received and acted upon by computers as a part of their normal operation. The information will be sent and received in ways that are native to that population and therefore in ways that are most optimal for that intelligent entity.
(This essay was originally published in September, 2000)