Bill, you missed my point: I am accusing the dependency on user reachability information dissemination and wanted to hi-light its absurdity by this example. The determination of various policy/QoS- specific routes is another topic and can well be served, if the topolgy is well known. On the one hand side, I am confronted with non-understanding, on the other hand side I should explain policy/QoS routing prior shortest routing. Believe me, I have algorithms ipfrr doesn't get close to them. And distance vector based BGP doesn't even provide half of the possible routes. Where is time-of-day based inter-domain routing ? Where is the right and natural approach to combine the end-hosts' and the ISPs' interests in routing ? There would be a substantial improvement due to TARA, a) by better than Dijkstra routing algorithms, b) because it became affordable (no churn, no binary or hash search, no table size problem) and c) because it became feasible ( roughly: two thirds of the paths aren't even offered for selection). Heiner In einer eMail vom 21.12.2008 23:57:37 Westeuropäische Normalzeit schreibt [email protected]:
On Sun, Dec 21, 2008 at 11:17 AM, <[email protected]> wrote: > My saying since 45. I have referred to two other networks which both are > 1000 times larger: > the road network, and the postal service network. None of them has this > crazy problem because none of them does user reachability information > dissemination. Let me point to a 3rd network: the railroad network. > There a single link (=railway between two distant cities) may take 10 years > building it. Heiner, How do you select which train or flight you will take? When I select a train or flight, I have multiple considerations: 1. From the potentially viable destination stations, how convenient is the street-level route I've precomputed after acquiring a local map? If I take a cab, how much will the cab fare cost? Would it be cheaper to rent a car? 2. From the potentially viable source stations, how convenient is the street-level route from my current location? If I take a cab, how much will the cab fare cost? Would it be cheaper to pay for parking at the station? 3. Which trains or flights go to the stations in question during the period I want to travel? 4. Which train or flight is offered at a price I'm willing to pay? You'll notice that in order to take a single long trip, I've acquired a substantial subset of the routing data and applied a very complex set of heuristics in order to precompute my door-to-door route. And did I mention that the trains and aitlines are regularly bankrupt? It seems they have a helluva time managing their costs. Geographic routing for human travel may seem simple and efficient, but it only seems that way. Regards, Bill Herrin -- William D. Herrin ................ [email protected] [email protected] 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/> Falls Church, VA 22042-3004
_______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
