1. Spring 20071 Routing & Switching Umar Kalim Dept. of Communication Systems Engineering umar.kalim@niit.edu.pk http://www.niit.edu.pk/~umarkalim 17/04/2007

2. Spring 20072 Outline Link State (OSPF) Routing Metrics

3. Spring 20073 Routing in the Internet The Global Internet consists of Autonomous Systems (AS) interconnected with each other: –Stub AS: small corporation: one connection to other AS’s –Multihomed AS: large corporation (no transit): multiple connections to other AS’s –Transit AS: provider, hooking many AS’s together

4. Spring 20074 Internet AS Hierarchy Inter-AS border (exterior gateway) routers Intra-AS interior routers

5. Spring 20075 Intra-AS Routing Also known as Interior Gateway Protocols (IGP) Most common Intra-AS routing protocols: –RIP: Routing Information Protocol –OSPF: Open Shortest Path First –IGRP: Interior Gateway Routing Protocol (Cisco proprietary)

6. Spring 20076 Link State (OSPF) Reliable Flooding for dissemination –A nodes ends its link-state information out on all of its directly connected links which in turn forward it on all of their directly connected links. –Link State Packet (LSP)  ID of node creating the packet  List of directly connected neighbours + cost  A sequence number  TTL –Example  Small seq. # ~ old LSP  The fact that an LSP is not sent back on the link from which it is received allows to put an end to the flooding

7. Spring 20077 Link State (OSPF) LSPs are generated in the case of a change in topology or due to timer expiry Hello packet to test if a node is alive or not Design goals –Newest information must be flooded ASAP  Minimize overhead –How?  Seq # & TTL ensure correctness of information Route Calculation ~ Dijkstra’s Algorithm

8. Spring 20078 Link State (OSPF) Characteristics –Stabilizes quickly  Promptly Responds to topology changes –Information stored at each node is quite large  Thus potential problems with scalability Difference in DV and LS?

9. Spring 20079 OSPF (Open Shortest Path First) “open”: publicly available Uses Link State algorithm –LS packet dissemination –Topology map at each node –Route computation using Dijkstra’s algorithm OSPF advertisement carries one entry per neighbor router Advertisements disseminated to entire AS (via flooding) –Carried in OSPF messages directly over IP (rather than TCP or UDP

10. Spring 200710 Routing Metrics Assign cost of 1 to all links ~ count hops –Problems?  Differentiation b/w links on the basis of –Latency –Capacity –Current load  Calculating current load is a non-trivial task

11. Spring 200711 Routing Metrics Initially ARPANET suggested we use queue length –No consideration for latency & capacity New routing mechanism –Consider bandwidth and latency –For latency calculate delay  Delay = (DepTime – ArrTime) + TransTime + Latency  Note that (DepTime – ArrTime) relfects queuing delay  Also if the link-level ACK is not received (hence retransmission) the reliability can be calculated Worked well with light load For heavy load, results advertise very high cost thus shifting load to another link (ping-pong!.. Under heavy load one link becomes idle)

12. Spring 200712 Routing metrics Range of link values was too large –A heavily loaded 9.6 Kbps link may appear to be N times costly than a lightly loaded 56Kbps link  Thus a 100 hops of lightly loaded 56 Kbps link may be selected instead of a single hop of 9.6 Kbps

13. Spring 200713 Routing metrics Solution ~ revised ARPANET routing metric –What can be the solution? –Smoothing  Link utilization instead of delay measurements  Limit on how much the measured metric can change Characteristics –A highly loaded link never shows a cost of more than 3 times its cost when idle –The most expensive link is only seven times the cost of the least expensive link –A high speed satellite link is more attractive than a low speed terrestrial link –Cost is a function of link utilization only at moderate to high loads

14. Spring 200714 Routing Metrics Calculating edge weights ~ frequency? –Points to ponder  None of the metrics are instantaneous; we calculate averages  Whenever a metric changes, should we update –Why not update when the change is greater than a certain threshold?

15. Spring 200715 Assigned study Revised ARPANET routing metric

16. Spring 200716 Questions? That’s all for today!