With the relation to my work at the Global Brain Institute and particularly with ChallProp, I started to research more deeply the area of distributed systems. The idea is to see the borderline between distributed system and distributed intelligence and how the former morphs into another. There is an implicit assumption that distributed intelligence is a kind of special case of the distributed system. I have started with YouTube first looking for presentations and lectures around the topic. Here interesting things + my notes and thoughts that I was able to dig out so far:
Distributed Computing Theory by N.Lynch
Distributed Computing Systems (university course)
A presentation about Distributed Computing Theory, focused more on practical applications / algorithms and mobile networks. The most interesting part for me was something called “Computing in Dynamic Graph Networks” (0:35).
- Distributed Computing Theory (book);
- Atomic Memory and sharing memory robustly in message -passing systems (article); Quite interestingly, this is more or less the same algorithm that I use in ChallProp …
- Computing in Dynamic Graph Networks. Lynch’s PhD student, Rotem Oshman has been studying that (PhD Thesis: Distributed Computation in Wireless and Dynamic Networks; article: Distributed Computation in Dynamic Graphs);
- The problem that is being studied by the field of computation in dynamic graphs is that dynamic graphs are changing during computation.. This is the problem very similar to the one we address with the challenge propagation model and in general trying to model the emergence of the distributed intelligence (whether in Internet in the form of the Global Brain, or elsewhere..).
A course from Prof. Keshav, University of Waterloo, Winter 2012. In this course he is focusing on the Internet as a distributed computing system.
- The trend in the world is converting Atoms to Bits (see Nicholas Negroponte, Being Digital).
- Internet can be looked at / described from three main perspectives:
- service view (transporting bits from here to there);
- topology view: the global conservativeness;
- formal definition: the set of all reachable IP addresses;
- There is no one Internet in the sense, that every access point (or every server computer / user) sees Internet differently. I guess what you can infer from that is that Internet (the global one) is the come sort of the amalgam of these different views;
- The protocol (definition): a way to formulate relationship and interaction [between agents /nodes, whatever];
- Kind of the definition of the Internet (came at the end of the lecture, after quite a long discussion of basic topology of the Internet): the network of autonomous systems. My try to formulate this idea a little more generally: Internet is the recursive network of autonomous systems, where the autonomous system can be network itself. Consider association with a framework for scalable cognition.