ROUTING AND FORWARDING

Transcription

1 ROUTING AND FORWARDING Mario Baldi Routing - 1

2 Copyright Notice This set of transparencies, hereinafter referred to as slides, is protected by copyright laws and provisions of International Treaties. The title and copyright regarding the slides (including, but not limited to, each and every image, photography, animation, video, audio, music and text) are property of the authors specified on page 1. The slides may be reproduced and used freely by research institutes, schools and Universities for non-profit, institutional purposes. In such cases, no authorization is requested. Any total or partial use or reproduction (including, but not limited to, reproduction on magnetic media, computer networks, and printed reproduction) is forbidden, unless explicitly authorized by the authors by means of written license. Information included in these slides is deemed as accurate at the date of publication. Such information is supplied for merely educational purposes and may not be used in designing systems, products, networks, etc. In any case, these slides are subject to changes without any previous notice. The authors do not assume any responsibility for the contents of these slides (including, but not limited to, accuracy, completeness, enforceability, updated-ness of information hereinafter provided). In any case, accordance with information hereinafter included must not be declared. In any case, this copyright notice must never be removed and must be reported even in partial uses. Routing - 2

3 Routing Determine a path through the network for packets Routing - 3

4 Forwarding Advance packets through the network Includes a routing decision Routing - 4

5 Forwarding and Routing B B 1. Routing (proactive) A B,D A,C,E,F B 2. Forwarding (on the fly routing) A,B C C,D,F B,D A,C,E,F C,D,F A,B D B F C,A,B,D E F E,A,B F Routing - 5

6 Proactive Routing Independent of actual traffic Determine reachable destinations Compute best route Commonly referred to as routing Routing - 6

7 On-the-fly Routing Realized when handling each packet Based on local information Routing/forwarding table Output of proactive routing or signaling A.k.a. route Routing - 7

8 On-the-fly Routing Algorithms Routing by Network Address Label Swapping Source Routing Each protocol architecture adopts one or more Routing - 8

9 Forwarding Phases Routing (on-the-fly) Output port selection Possibly next-hop selection Switching: transfer to output port Transmission Routing - 9

10 A Proactive Routing Algorithm Classification Non-adaptive algorithms (static) Adaptive algorithms (dynamic) Routing - 10

11 NON-ADAPTIVE ROUTING Routing - 11

12 Non-adaptive Algorithms Fixed Directory routing AKA static routing Manual configuration (Selective) flooding and derivates Routing - 12

13 A Fixed Directory Routing Routing Table of B A,E C,D B C D Routing Table of D A,B,C,E Routing Table of A B,D E,C E Routing Table of E A,B C,D Routing Table of C A,B D E Routing - 13

14 Pros and Cons Administrator has full control Error prone It does not adapt to topology changes Routing - 14

15 Backup Route Load Balancing Routing Table of B A,E,C,D C,D,E,A Routing Table of D A,B,C,E A B C D Routing Table of A B,D,C,E E,C,B,D E Routing Table of E A,B C,D,B Routing Table of C A,B D E,A,B Routing - 15

16 Static vs. Dynamic The only way to the rest of network Multiple alternative routes Core Dynamic routing Edge Static routing Small routing tables Routing - 16

17 DYNAMIC ROUTING Routing - 17

18 Adaptive Algorithms Centralized Routing Isolated Routing Distributed Routing Distance Vector Link State Routing - 18

19 Centralized Routing Routing Control Center (RCC) Calculates and distributes routing tables Needs information from all nodes Routing - 19

20 Centralized Routing Optimizes performance Simplifies troubleshooting Significant network load in proximity of RCC Routing - 20

21 Centralized Routing RCC is single point of failure RCC is bottleneck Not suitable for highly dynamic networks Routing - 21

22 Isolated Routing Each node decides independently No exchange of information E.g., Backward Learning IEEE 8O2.1D bridges Routing - 22

23 Distributed Routing Combines advantages of isolated and centralized routing Routers co-operate in exchanging connectivity information Each router decides independently, but coherently Routing - 23

24 DISTANCE VECTOR ROUTING ALGORITHM Routing - 24

25 Basic Principle 2. A at dist 0 and E at 1 (from A1) 3. A at 0 and E at 1 1. E at dist 0 A1 C1 C2 A C 4. C at 0, A at 1 (from C1), F at 1 (from C2), and E at 2 (from C1 and C2, same dist) 2. F at 0 and E at 1 (from F1) E F1 F 3. F at 0 and E at 1 Routing - 25

26 Distance Vector List of reachable destinations (all!) Distance from announcing router Generated by each router Received from neighbors Routing - 26

27 Sample Scenario A D B E Loc (A) A, 0 C Routing information stored by A DV (B) A, 1 B, 0 C, 1 D, 1 E, 1 DV (D) A, 1 B, 1 C, 2 D, 0 E, 1 Routing - 27

28 A A2 A1 B C Distance Vector Merging and Generation D E Received from line A1 Received from line A2 Loc (A) DV (B) A, 0 DV (D) ROUT. TABLE (A) A, 1 B, 0 A, 1 B, 1 A, local, 0 B, A1, 1 C, 1 D, 1 C, 2 D, 0 C, A1, 2 D, A2, 1 E, 1 E, 1 E, A2, 2 DV (A) A, 0 B, 1 C, 2 D, 1 E, 2 Routing - 28

29 A A2 A1 B C Topology Change D E Loc (A) DV (B) DV (D) ROUT. TABLE (A) A, 0 A, 1 B, 0 C, 1 D, 1 E, 1 E, 1 A, 1 A, local, 0 B, 1 B, A2, 2 C, 2 C, A2, 3 D, 0 D, A2, 1 E, A2, 2 DV (A) A, 0 B, 2 C, 3 D, 1 E, 2 Routing - 29

30 Example: Cold Start RT (A) A,loc,0 RT (A) A,loc,0 RT (A) A,loc,0 B,A1, 1 D,A2, 1 RT (A) A,loc,0 B,A1, 1 C,A1, 2 D,A2, 1 E,A2, 2 RT (B) B,loc,0 RT (B) A,B1, 1 B,loc,0 RT (B) A,B1, 1 B,loc,0 D,B2, 1 RT (B) A,B1, 1 B,loc,0 C,B4, 1 D,B2, 1 E,B3, 1 RT (C) C,loc,0 A sends its DV RT (C) C,loc,0 B and D send their DVs RT (C) A,C1,2 B,C1,1 C,loc,0 RT (D) D,loc,0 RT (D) A,D1, 1 D,loc,0 RT (D) A,D1, 1 B,D2, 1 D,loc,0 All send their DVs... RT (C) A,C1,2 B,C1,1 C,loc,0 D,C2, 2 E,C2, 1 RT (D) A,D1, 1 B,D2, 1 C,D2, 2 D,loc,0 E,D3, 1 RT (E) E,loc,0 RT (E) E,loc,0 RT (E) A,E2, 2 B,E2, 1 D,E1, 1 E,loc,0 RT (E) A,E2, 2 B,E2, 1 C,E3, 1 D,E1, 1 E,loc,0 A D A2 D1 A1 D2 D3 B1 B2 E1 B3 E2 B E B4 E3 C1 C C2 Routing - 30

31 Issues Several problems Black Hole Count to infinity Bouncing Effect (loop) Instability Routing - 31

32 Count to Infinity A A1 B1 B2 C1 B C IS B: Loc (B) B, 0 DV (A) A, 0 B, 1 C, 2 DV (C) A, 2 B, 1 C, 0 RT (B) A,B2,3 B,loc,0 C,B2,1 IS C: Loc (C) C, 0 DV (B) A, 1 B, 0 C, 1 RT (C) A,C1,2 B,C1,1 C,loc,0 B sends DV Loc (C) C, 0 DV (B) A, 3 B, 0 C, 1 RT (C) A,C1,4 B,C1,1 C,loc,0 C sends DV Loc (B) B, 0 DV (C) A, 4 B, 1 C, 0 RT (B) A,B2,5 B,loc,0 C,B2,1 Count to Infinity! Routing - 32

33 Bouncing Effect A A1 B1 B2 C1 B C IS B: Loc (B) B, 0 DV (A) A, 0 B, 1 C, 2 DV (C) A, 2 B, 1 C, 0 RT (B) A,B2,3 B,loc,0 C,B2,1 IS C: Loc (C) C, 0 DV (B) A, 1 B, 0 C, 1 RT (C) A,C1,2 B,C1,1 C,loc,0 B sends its DV Loc (C) C, 0 DV (B) A, 3 B, 0 C, 1 RT (C) A,C1,4 B,C1,1 C,loc,0 Routing - 33

34 Issues Partial solutions Split Horizon Path Hold Down Route Poisoning Routing - 34

35 Split Horizon If C reaches destination A through B, it is useless for B trying to reach A through C A A1 B1 B2 C1 B C C2 D1 D IS B: Loc (B) B, 0 DV (A) A, 0 DV (C) C, 0 D, 1 RT (B) B,loc,0 C,B2,1 D,B2, 2 DV (C C1) C, 0 D, 1 DV (C C2) A, 2 B, 1 C, 0 Routing - 35

36 Split Horizon Routing - 36 Prevents loops between two nodes Speeds up convergence Personalized DVs to neighbors DV of C to B does not contain destinations reached through B In actual implementations, route has to expire

37 Split Horizon on Mesh IS B: A Loc (B) B, 0 IS C: Loc (C) C, 0 A1 D1 D DV (A) A, 0 B1 B3 D2 B DV (C) C, 0 D, 1 B2 C2 B sends its DV DV (B) B, 0 D, 1 DV (D) A, 2 B, 1 D, 0 C1 C DV (D) C, 1 D, 0 RT (C) A,C2, 3 B,C1, 1 C,loc,0 D,C2, 1 RT (B) B,loc,0 C,B2,1 D,B3, 1 IS C: Loc (C) C, 0 IS D: Loc (D) D, 0 IS D: Loc (D) D, 0 DV (B) A, 1 B, 0 D, 1 DV (B) A, 1 B, 0 C, 1 DV (B) B, 0 C, 1 DV (D) A, 2 B, 1 D, 0 DV (C) A, 2 B, 1 C, 0 DV (C) A, 2 B, 1 C, 0 RT (C) A,C1, 2 B,C1, 1 C,loc,0 D,C2, 1 RT (D) A,D1, 2 B,D1, 1 C,D2, 1 D,loc,0 RT (D) A,D2, 3 B,D1, 1 C,D2, 1 D,loc,0 Routing - 37

38 Split Horizon on Mesh IS C: Loc (C) C, 0 (from the previous slide) DV (B) B, 0 D, 1 DV (D) A, 2 B, 1 D, 0 RT (C) A,C2, 3 B,C1, 1 C,loc,0 D,C2, 1 IS D: Loc (D) D, 0 DV (B) B, 0 C, 1 DV (C) A, 2 B, 1 C, 0 RT (D) A,D2, 3 B,D1, 1 C,D2, 1 D,loc,0 C and D send their DVs IS B: Loc (B) B, 0 IS C: Loc (C) C, 0 DV (C) A, 3 C, 0 D, 1 DV (B) B, 0 D, 1 DV (D) A, 3 C, 1 D, 0 DV (D) B, 1 D, 0 RT (B) A,B3, 4 B,loc,0 C,B2,1 D,B3, 1 RT (C) B,C1, 1 C,loc,0 D,C2, 1 IS D: Loc (D) D, 0 A A1 DV (B) B, 0 C, 1 D1 D B1 B3 D2 DV (C) B, 1 C, 0 B B2 C2 C1 C RT (D) B,D1, 1 C,D2, 1 D,loc,0 Routing - 38

39 Path Hold Down If link L fails, all destinations reachable through link L are considered unreachable for a certain period of time I.e., no routes to them are computed A A1 B1 B2 C1 B C C2 D1 D IS B: Loc (B) B, 0 DV (A) A, 0 B, 1 C, 2 D, 3 Quarantine! DV (C) A, 2 B, 1 C, 0 D, 1 RT (B) B,loc,0 C,B2,1 D,B2, 2 IS C: Loc (C) C, 0 DV (B) B, 0 C, 1 D, 2 DV (D) A, 3 B, 2 C, 1 D, 0 Count to Infinity! RT (C) A,C2, 4 B,C1, 1 C,loc,0 D,C2, 1 Routing - 39

40 Path Hold Down High convergence time for the examined node (even with an alternative path) The router that noted the fault may not participate to any loop at least until the timeout of Hold Down timer Routing - 40

41 React on Cost Increase Routing loops happen when routes have increasing costs Routing - 41 Cost-increasing routes in DVs are not used Two subsequent advertisements show a cost increase Possibly with Path Hold Down Might block routes with legitimate cost increase

42 Route Poisoning Routing - 42 An invalid route is advertised at infinite distance Instead of just omitting it It would have to expire Faster convergence time E.g., when link fails or cost increases It can substitute or complement Path Hold Down

43 Split Horizon with Poisonous Reverse More aggressive No theoretical advantages Practically, no need to wait for route expiration A A1 B1 B2 C1 B C C2 D1 D IS B: Loc (B) B, 0 Routing - 43 DV (A) A, 0 B, inf C, inf D, inf DV (C) A, inf B, inf C, 0 D, 1 RT (B) B,loc,0 C,B2,1 D,B2, 2 DV (C C1) A, inf B, inf C, 0 D, 1 DV (C C2) A, 2 B, 1 C, 0 D, inf

44 The Bottom Line Routers do not know the network topology Based on distance vectors B cannot distinguish A B C A B C Routing - 44

45 Advantages Simple to implement Protocols simple to deploy Very little configuration Routing - 45

46 Shortcomings Exponential worst case complexity and convergence time O(n 2 ) to O(n 3 ) Routing - 46

47 Shortcomings Convergence limited by slower links and routers set pace Complex tuning Routing - 47

48 Shortcomings Complex troubleshooting Large routing traffic (and storage) Not suitable for large complex networks Routing - 48

49 Path Vector Eliminates routing loops A A1 B1 B2 C1 B C C2 D1 D A2 IS A: Loc (A) A, 0 PV (B) A, 1, [B] B, 0, [-] C, 1, [B] D, 2, [B,C] PV (C) A, 1 [C] B, 1, [B] C, 0, [-] D, 1, [D] RT (A) A, loc, 0 B, A1, 1 C, A2, 1 D, A2, 2 PV (A) A, 0, [-] B, 1, [A] C, 1, [A] D, 2, [A,C] Routing - 49

50 LINK STATE ROUTING ALGORITHM Routing - 50

51 Basic Principles 2. I, node A, connect to E and C A A E F A C E F 1. I, node E, connect to A and F C A C E F E F 3. I, node F, connect to E and C Routing - 51

52 Basic Principles Information on the state of each link Link state A local map Sent by each node to all other nodes Selective flooding Routing - 52

53 Basic Principles Nodes build a network map The same on all nodes Each node computes routes on the map Dijkstra (shortest path first) algorithm Routing - 53

54 Link State Database A B C D LS (A) B, 1 D, 1 E LS (B) A, 1 C, 1 D, 1 E, 1 Stored by each router LS (C) B, 1 E, 1 LS (D) A, 1 B, 1 E, 1 LS (E) B, 1 C, 1 D, 1 Routing - 54

55 Dijkstra Algorithm Low complexity L log (N) L: number of links N: number of nodes Shortest Path First Routing - 55

56 Shortest Path First The next node nearest to the root is identified Its path is inserted into the routing table Routing - 56

57 Routing - 57 Example B A C D E root PATH TEMP F 2 2) B A C D E F 3) B A C D E F 4) B A C D E F 5) B A C D E F 6) B A C D E F B A C D E F 7) B A C D E F 8) B A C D E F 9) B A C D E F 10) 1)

58 Rapid Convergence LSs spread quickly No intermediate processing Routing - 58

59 Limited Routing Traffic and Storage Link states are small Fast and efficient neighbor greeting Routing - 59

60 Other Advantages It rarely generates loops Simple to understand and troubleshoot All nodes have identical databases Routing - 60

61 Shortcoming High implementation complexity Selective flooding First implementation took several years Protocols with complex configuration Routing - 61

62 Link State Generation In principle: when there is a topology change Actual protocols: LS are generated periodically Increased reliability Routing - 62