Computer Networks Architectures and Protocols ShortestPath Routing and

Computer Networks Architectures and Protocols Shortest-Path Routing and IP OSPF Routing Departamento de Informática da FCT/UNL 1

Lecture Outline • What is routing, shortest-path routing • One routing protocol example – Link state: Open Shortest Path First (OSPF) 2

Why Does Routing Matter? • End-to-end performance – Quality of the path affects user performance – Propagation delay, throughput, and packet loss • Transient disruptions during changes – Failures, maintenance, and load balancing – Limiting packet loss and delay during changes • Use of network resources – Balance of the traffic over the routers and links – Avoiding congestion by directing traffic to lightlyloaded links 3

Some Routing Facets • Static versus dynamic routing – Static routing: when routing decisions are seldom or never changed – Dynamic routing: when routing decisions are evaluated periodically or when a network modification takes place • Criteria for routing – Shortest-path routing – Network usage maximisation (load distribution, traffic engineering, . . . ) – Policy routing (e. g. Inter-AS routing, BGP, routing based on the origin of packets, . . . ) 4

Forwarding versus Routing • Forwarding: data plane – Directing a data packet to an outgoing link – Individual router using a forwarding table • Routing: control plane – Computing paths the packets will follow – Routers talking amongst themselves – Individual router creating a forwarding table 5

Forwarding versus Routing 6

Static Routing ! ! hostname R 1 10. 1. 2. 0/24 ! 10. 1. 2. 254 ip route 10. 1. 5. 0 R 1 10. 1. 4. 254 10. 1. 3. 254 255. 0 10. 1. 3. 253 ! ! 10. 1. 3. 0/24 10. 1. 3. 253 10. 1. 4. 0/24 R 2 10. 1. 5. 254 10. 1. 5. 0/24 R 3 This kind of configuration is only used to enter specific and static routes in the routing table. Entering a default route is an example of such usage: ip route 0. 0 10. 1. 2. 100 7

Open Shortest-Path Protocol 8

Modelling a Network as a Graph 9

Dijkstra Algorithm: Shortest-Paths Tree 10

Link-State Routing • Each router keeps track of its incident links – Whether the link is up or down – The cost on the link • Each router broadcasts the link state (LSA) – To give every router a complete view of the graph • Each router runs Dijkstra’s algorithm – To compute the shortest paths – … and constructs the forwarding table • Example protocols – Open Shortest Path First (OSPF) – Intermediate System (IS-IS) 11

Detecting Topology Changes 12

Broadcasting the Link State • Reliable flooding –Ensure all nodes receive link-state information –… and that they use the latest version • Challenges –Packet loss –Out-of-order arrival –Crash and quickly rebooting a router • Solutions –Acknowledgments and retransmissions –Sequence numbers –Time-to-live for each packet 13

When to Initiate Flooding • Topology change – Link or node failure – Link or node recovery • Configuration change – Link cost change • Periodically – Refresh the link-state information – Typically (say) 30 minutes – Corrects for possible corruption of the data 14

Convergence • Getting consistent routing information to all nodes – E. g. , all nodes having the same link-state database • Consistent forwarding after convergence – All nodes have the same link-state database – All nodes forward packets on shortest paths – The next router on the path forwards to the next hop 15

Transient Disruptions • Detection delay – A node does not detect a failed link immediately – … and forwards data packets into a “blackhole” – Depends on timeout for detecting lost hellos – Some routers know about failure before others – The shortest paths are no longer consistent – Can cause transient forwarding loops 16

Inconsistency Example 17

Convergence Delay • Sources of convergence delay – Detection latency – Flooding of link-state information – Shortest-path computation – Creating the forwarding table • Performance during convergence period – Lost packets due to blackholes and TTL expiry – Looping packets consuming resources – Out-of-order packets reaching the destination • Very bad for Vo. IP, online gaming, and video 18

Reducing Convergence Delay • Faster detection – Smaller hello timers – Link-layer technologies that can detect failures • Faster flooding – Flooding immediately – Sending link-state packets with high-priority • Faster computation – Faster processors on the routers – Incremental Dijkstra algorithm • Faster forwarding-table update – Data structures supporting incremental updates 19

Paper P. François, C. Filsfils, J. Evans and O. Bonaventure, “Achieving Sub-Second IGP Convergence in Large IP Networks, ” in Computer Communications Review, Volume 35, Number 2, July 2005, pp. s 35 - 44 20

Hierarchical OSPF and OSPF Areas 21

Hierarchical OSPF • Two level hierarchy: local areas and backbone area • Link-state announcements are only flooded in one area • Other areas are not known in detail but their list of networks. • Area border routers: announce network summaries from one area to the other • Backbone routers: these are the routers of area 0 • Boundary routers: are connected to other Ass • Therefore, besides link-state announcements, there also network summaries announcements 22

OSPF Route Types Intra-area route — a route to a network of the same area Inter-area route — a route to a network of another area; its cost includes all link costs up to a area border router, and from there up to the destination External route — a route leading to a network external to the ospf domain 23

Designated Router • Designated Router — if a certain number of different OSPF routers are connected to the same broadcast network, one is in charge of all the locally reachable networks • The Designated Router is elected by the way of a manually set priority – ip ospf priority number • To avoid partitions, there is a Backup Designated Router announcing the same networks 24

OSPF Database Exchanges • OSPF “database state exchanges” take place among adjacent OSPF routers • These are known as LSAs or “Link State Announcements” • The header of these packets has the following information: sequence number, router ID of the sender router, type, announcement ID, authentication information, . . . • The exchange transaction begins with an exchange proposal. If the receiver does not know the announcement ID, or has one with an older sequence number, it sends a link-state request, otherwise it declines the update • The answer is a link-state update which must be acknowledge by a link-state acknowledgement 25

Route Aggregation Summaries • To avoid the explosion of route propagation information, routes should be aggregated as much as possible — this allows shorter route announcements • Routers can try to perform automatic route aggregation which is a computationally intensive task. Another way consists in doing aggregation by hand, which can be error prone • Aggregation efficiency can be improved by carefully choosing network numbers — for example all “totallystubby areas” (areas only connecting to the backbone area) should have, as much as possible, networks sharing the same prefix • This is yet another example of the interactions of routing scalability and network addressing 26

Link flapping • If an interface goes up and down too frequently, this will constantly flood the network with new link-state announcements, and will require, a new computation of the new shortest paths (SPF algorithm) per each link-state announcement (LS) received • The command: show ip ospf shows the number of times the algorithm has been executed recently • By default, the SPF algorithm is only executed 5 seconds after a LS announcement arrives, and at least 10 seconds should separate each SPF execution • The command ospf timers allows the modification of these values router ospf timers spf <schedule delay in seconds> <hold-time in seconds 27

Conclusions • Routing may be treated by a distributed algorithm – React to changes in the topology – Compute the shortest paths • Two main shortest-path algorithms – Dijkstra → link-state routing (e. g. , OSPF and IS-IS) –Bellman-Ford → distance vector routing (e. g. , RIP) • Convergence process – Changing from one topology to another – Transient periods of inconsistency across routers – Big problem if the convergence process is slow 28

Written assignment Each student must deliver a written report of at most 5 pages (12 dots, double spaced) on the subject of the following paper: – P. François, C. Filsfils, J. Evans and O. Bonaventure, “Achieving Sub-Second IGP Convergence in Large IP Networks, ” in Computer Communications Review, Volume 35, Number 2, July 2005, pp. s 35 - 44 Deliver date: 5 th of April 2018 29