Distance Vector Routing

1. Distance Vector Routing 박주호

2. Introduction. • Modern computer Network generally use Dynamic routing algorithms rather than The static one. • Two dynamic algorithm • Distance vector routing • Link state routing juhpark@cs.chonbuk.ac.kr

3. Distance vector routing • This algorithms operate by • Having each router maintain a table • The table(i.e vector) give • the best known distance to each destination and • which line to use to get there. Updated by exchanging information with the neighbors. juhpark@cs.chonbuk.ac.kr

4. History • Sometimes called by other names… • Distributed Bellman-Ford routing algorithm(1957) • Ford-Fulkerson algorithm(1962) • It was original ARPANET routing algorithm • And was used • In the Internet(RIP) • In early versions of DECnet and Novell’s IPX • AppleTalk and Cisco routers use improved distributed vector protocols. juhpark@cs.chonbuk.ac.kr

5. The Routing table • Each router maintains a routing table • Contain one entry. • Two part of one entry • The preferred outgoing line • An estimate of the time or distance • The router is assumed to know the “distance” to each of its neighbors. juhpark@cs.chonbuk.ac.kr

6. The Metric of distance • Number of hops • The distance is just one hop • Time delay in milliseconds • The router can measure… • Using special ECHO packets(timestamps) • Total number of packets queued along the path • The router simply examines each queue juhpark@cs.chonbuk.ac.kr

7. New estimated dealy from J To A I H K Line A 0 24 20 21 8 A B 12 36 31 28 20 A C 25 18 19 36 28 I D 40 27 8 24 20 H E 14 7 30 22 17 I F 23 20 19 40 30 I G 18 31 6 31 18 H H 17 20 0 19 12 H I 21 0 14 22 10 I J 9 11 7 10 0 - K 24 22 22 0 6 K L 29 33 9 9 15 K JA delay is 8 JI delay is 10 JH delay is 12 JK delay is 6 New routing table for J Vectors received from J’s four neighbors Algorithm JA = 8 JA = 8 AB = 12 JB = JA+AB = 20 (a) Receive table from neighbors Distances to each neighbor (b) juhpark@cs.chonbuk.ac.kr

8. A A B B C C D D E E The Count-to-Infinity Problem • Distance vector routing works in theory • A serious drawback in practice • Reacts rapidly to good news, • But leisurely to bad news. •    Initially 1    After 1 exchange 1 2   After 2 exchange 1 2 3  After 3 exchange 1 2 3 4 After 4 exchange 1 2 3 4 Initially 3 2 3 4 After 1 exchange 3 4 3 4 After 2 exchange 5 4 5 4 After 3 exchange 6 5 6 After 4 exchange 7 6 7 6 After 5 exchange 7 8 7 8 After 6 exchange     juhpark@cs.chonbuk.ac.kr

9. In case good news. • If longest path of subnet is N hops, • Everyone will know good news. • In case bad news. • Infinity… • Set infinity to the longest path + 1. • If metric is time delay, no well-defined upper bound. • There is a tradeoff… juhpark@cs.chonbuk.ac.kr

10. A B C D E The Split Horizon Hack • Many ad hoc solutions to the problem have been proposed… • But, complicate , useless … • One of them will be described(why fail?) • The Split Horizon Hack algorithm 1 2 3 4 Initially  2 3 4 After 1 exchange   3 4 After 2 exchange •   4 After 3 exchange     After 4 exchange juhpark@cs.chonbuk.ac.kr

11. A B C Router D Fig. 5-12. An example where split horizon fails. juhpark@cs.chonbuk.ac.kr