I just finished reading the Robert J. Sawyer novel, "WWW:WATCH" that was a sequel to WAKE and I do anticipate reading and enjoying the next one in the series WONDER coming out in 2011.
I read the first novel "WWW:WAKE" in serial form in a science fiction magazine and did enjoy it. The idea is that a conscious entity emerges spontaneously from the Internet.
Because I do understand a lot of the science that is behind the fiction in the book and those parts are pretty much correct from what I could see, then it was a short step to 'suspend-my-sense-of-disbelief' an enjoy the story.
Having said that I can now talk a little real-world stuff:
Assuming that such a thing did come to exist there is a place in the book where the group that was trying to shutdown WebMind could have done a better job. If they were 'on-their-toes' they could have, in cooperation with the companies that supply updates to the routers send out a 'code-update' to all the routers that would cause them to instantly ignore 'mutant-packets' that are supposed to make up the entity.
It is the case that routers all over the world are responsible for sending packets along.
However the possible existence of 'mutant-packets' that have a 'time-to-die' count that prevents them from expiring is almost certainty impossible. The packets themselves are are  received,  inspected,  sent along or not sent along. The third of these three activities depends on the packet passing some basic integrity tests. Sadly, 'mutant-packets' would instantly fail the test and be ignored. However having said that I can imagine that a 'self-modifying-packet' could be added to the type of Internet packets processed and such a packet could somehow fit the bill. I expect implementation of such a packet would be a significant change to how Internet routers work.
About a sufficiently complex cellular automaton spontaneously becoming conscious: I do think that sufficiently complex system can undergo and do indeed undergo 'self-organization' events. Even very simple systems can be thought of as undergoing self organization events. However I think that for 'self-organization' events to rise to the level of a conscious object requires something more than just becoming self organized.
An example of the result of a highly self organized event is a biological cell and multi cellular life forms. So this could and did emerge spontaneously (with apologizes to the Intelligent Design folks). To get form a spontaneously organized complex object to to a conscious object requires another step. I believe that consciousness actually is the ability of an organism (or object, or cellular automaton) to take its sensory inputs and then use them as inputs into a modeling engine that then drives its outputs (movements) in such a fashion that it can negotiate its environment.
In the story the cellular automaton only gradually acquired sensory inputs. In the story the consciousness existed before it had sensory inputs. I think this is not correct. I think this is an example of the chicken-or-the-egg where the story had the wrong thing first. First there are sensory inputs into the complex system and then there emerges the ability of the system to model those inputs with the result being the ability of the system to negotiate its environment. As this model gets better and better then the amount of 'consciousness-stuff' increases.
In the story there is nothing that explains what consciousness actually is. I think any talk of consciousness requires such a thing because I think that you can not contemplate this without thinking about what it actually is.
The story can certainly get away with this part. So in the end the emerging consciousness is described as this completely unexpected thing. No discussion of how it came about just that it did. And when you think about self organizing systems in general it is not clear how they self organize. I do expect that once we get a handle on how this occurs from first principals we will have at that time a much deeper understanding of things than we do now.