On Mon, 2005-11-07 at 11:29 +0100, Neal H. Walfield wrote: > At Sat, 05 Nov 2005 12:49:46 -0500, > Jonathan S. Shapiro wrote: > > 4. Because market-based mechanisms rely on global communication, they > > constitute a global communication channel. > > Can you clarify this a bit more. How do market-based mechanisms rely > on global communication? My understanding of the power of markets is > that they are useful becuase they precisely don't require global > communication to find acceptable solutions. Scott H. Clearwater > states in the introduction to "Market-Based Control: A Paradigm for > Distributed Resource Allocation:"
Clearwater doesn't grok markets very well if he believes this. What Clearwater neglects is that the market itself is a communicating agent. Therefore, between any two parties A and B in a market, there is a communication channel involving at most two steps. To illustrate: by varying my demand, given knowledge of the market resolution algorithm, I can predictably induce rebalancing requests that will be delivered to you. Two parties collaborating in this way have a strikingly effective means of communication. shap _______________________________________________ L4-hurd mailing list [email protected] http://lists.gnu.org/mailman/listinfo/l4-hurd
