Download presentation
Presentation is loading. Please wait.
1
Dynamic Interior Routing Information Mechanisms
Chapter 5 Part 2 Dynamic Interior Routing Information Mechanisms
2
Dynamic Routing Using Distance Vector
Distance-vector algorithms pass along a message Describes the cost to varying destinations Each neighbor add its additional cost to costs of reachability Each node calculates the cost to adjacent or neighbor nodes Each node then transmits table of costs to all other nodes First- and second- generation transmit routing tables at fixed times. RIP broadcasts every 30 seconds. IGRP-distance vector but also transmit on triggered updates. Designing Routing and Switching Architectures Howard C. Berkowitz
3
Dynamic Routing using Distance Vector
Initially the router knows about adjacent entities. All routers know about all others after the third iteration of routing updates. Covergence Loop avoidance features will slow convergence routes. Designing Routing and Switching Architectures Howard C. Berkowitz
4
Designing Routing and Switching Architectures
Howard C. Berkowitz
5
Dynamic Routing Using Distance Vector
Loop Detection and Prevention Distance-vector depends on routers notifying other routers that a path is no longer available. Designing Routing and Switching Architectures Howard C. Berkowitz
6
Dynamic Routing Using Distance Vector
First-Generation Distance Vector:RIP ver 1 and 2 Widely implemented in hardware and software. UNIX routed Hops-count of intermediate systems between two endpoints as its routing metric Limited to networks of moderate size No more than 15 hops between two distances No heirarchy of routers. Large amount of routing traffic Designing Routing and Switching Architectures Howard C. Berkowitz
7
Dynamic Routing Using Distance Vector
RIP uses the following procedure: Routing tables have an entry for every possible routing destination. Next hop router and count to destination. Routers periodically send their routing tables to all adjacent routers. When new update arrives , the receiving router adds a 1 to each metric. If the resulting cost to any destination is less than what is in the current routing table, a cheaper cost replaces the old. Designing Routing and Switching Architectures Howard C. Berkowitz
8
Dynamic Routing Using Distance Vector
Special cases of updating If a new update arrives to a router it can be replaced if it has a greater cost-if the underlying router is in a better position to know the true cost. Not usually implemented in RIP. RIP ver 1 is widely implemented. RIP ver 2 Variable length subnet mask Optional authentication Operation over demand media Capability to multicast as well as broadcast updates. Uses hop count for metric and loop prevention Designing Routing and Switching Architectures Howard C. Berkowitz
9
Dynamic Routing Using Distance Vector
RIP Disadvantages Go thru slow link because it has a lower hop count For routers that do not keep information on potential routes and use holddown timers. When a failure is detected, a holddown timer is started. Holddown timer has value of 3 times the update interval. After the holddown timer has completed, 90 seconds. It takes at least 90 – 129 seconds for router to converge(find out about new route) 90 seconds for failure detection, 90 seconds holddown and 30 seconds waiting for update. Designing Routing and Switching Architectures Howard C. Berkowitz
10
Dynamic Routing Using Distance Vector
Second-Generation Distance Vector:IGRP Does not use hop count Uses complex metric primarily based on bandwidth. IGRP use hop counts for the purpose of loop detection. Uses periodic updates for new usable routes It sends triggered updates as soon it detects a failure IGRP is improvement over RIP Uses triggered updates and split horizon to prevent loops. IGRP knows the difference between links of different speed and generates less routing update overhead than RIP. Designing Routing and Switching Architectures Howard C. Berkowitz
11
Dynamic Routing Using Distance Vector
IGRP Updates, Timers, and Loop Control Assumed if hop count on a given route begins to increase steadily, or metric increases by more than 10%, it is a warning of a loop. Distance-vector adopts neew routes easily Does not purge old ones from the system. IGRP/EIGRP Metric Complex metric algorithm Primarily based on bandwidth, lesser extent on delay. Designing Routing and Switching Architectures Howard C. Berkowitz
12
Dynamic Routing Using Distance Vector
Third-Generation Distance Vector/DUAL:EIGRP Internally different protocol than IGRP Same configuration commands as IGRP Uses the same metric computation. Has different separate hello subprotocol Reliable update mechanisms that do away with most looping problems. Algorithm is called the diffusing update algorithm(DUAL). Does not use hop count Must reliably detect dead routers. Designing Routing and Switching Architectures Howard C. Berkowitz
13
Dynamic Routing Using Distance Vector
EIGRP forms adjacencies with all of its neighbors as long as there are no errors. EIGRP runs directly over IP EIGRP updates are fairly small than link-state updates in OSPF and they contain metric components rather than the computed metric. Uses a formal hello subprotocol. Routing Table Initialization Has neighbor table Also has a topology table which holds information about potential routes. Designing Routing and Switching Architectures Howard C. Berkowitz
14
Dynamic Routing Using Distance Vector
Learning about Non-adjacent routes Learns about routes from its neighbors Calculates its cost to reach the neighbor to the cost calcluated from the metric factors in updates it receives. The update carries factors from which the metric can be calculated. Learning Better Routes When a successor is found, it is moved from the topology table to the routing table. Designing Routing and Switching Architectures Howard C. Berkowitz
15
Dynamic Routing Using Distance Vector
Failure Recovery Topology table has two entries Active- router is searching for a way to reach the destination because it cannot find a route to the destination from its locally stored information. Passive-normal case. If a router sees that one of its routes have failed(infinite cost), it checks to see if one of its neighbors has a feasible route(feasible successor) Designing Routing and Switching Architectures Howard C. Berkowitz
16
Dynamic Routing Using Link-State Routing Protocols
Also called shortest path first (SPF) Each node determines which other nodes are adjacent to it and, sends information about the links to which it is directly connected. Link is not a route Router links with directly connected media Media links with directly connected router interface Designing Routing and Switching Architectures Howard C. Berkowitz
17
Dynamic Routing Using Link-State Routing Protocols
OSPF and IS-IS Different addresses used Different terminology OSPF- link state advertisements IS-IS-link state packets Similar Use link state algorithms Allow multiple levels of default route Use designated routers Hello protocols Designing Routing and Switching Architectures Howard C. Berkowitz
18
Dynamic Routing Using Link-State Routing Protocols
Distance vector protocols use less CPU than link state Link state routers normally do not advertise routes from a routing table. They advertise link states Algorithm is quick to adapt and does not produce loops but is resource intensive. Each router receives link states from all other routers in an OSPF area. Designing Routing and Switching Architectures Howard C. Berkowitz
19
Dynamic Routing Using Link-State Routing Protocols
OSPF Interior routing protocol intended to remove some of the limitations of RIP. Designed to scale for use in very large networks. Numerous mechanisms intended to reduce routing table size and volume of routing updates. No hop count-uses arbitrary metric defined by the system administrator. Mixed bandwidth. Designing Routing and Switching Architectures Howard C. Berkowitz
20
Dynamic Routing Using Link-State Routing Protocols
Uses link state algorithm As routers intialize, they flood the common medium within their area about the paths connected to them. Designing Routing and Switching Architectures Howard C. Berkowitz
21
Dynamic Routing Using Link-State Routing Protocols
Reduces traffic by specializing routers. Four functions Backbone routers (BRs)-routers in Should be reserved for inter-area transit and network management servers. Autonomous system border routers (ASBRs)-connect to other routing domains. Means of exchanging paths to another routing process or to accept static routes for redistribution. Most often in area. Area border routers (ABRs)-backbone is linked to non-backbone areas. Serve hosts within that area Only knows how to reach other routers in its area including area boundary router that connects the area to the backbone. Interior routers (IRs)-routers with interfaces in a single nonzero area. Designing Routing and Switching Architectures Howard C. Berkowitz
22
Designing Routing and Switching Architectures
Howard C. Berkowitz
23
Dynamic Routing Using Link-State Routing Protocols
OSPF Protocol Messages Designing Routing and Switching Architectures Howard C. Berkowitz
24
Dynamic Routing Using Link-State Routing Protocols
OSPF Packet Header Have a reserved multicast address as their destination. When packets are sent, IP TTL field in their IP headers are set to one, so they only have a link-local scope. Not intended to go beyond a single subnet Designing Routing and Switching Architectures Howard C. Berkowitz
25
Dynamic Routing Using Link-State Routing Protocols
Address and Neighbor Discovery As OSPF router initializes , it determines a unique router ID Then searches for other OSPF routers on media to which it is connected. After OSPF establishes interfaces that have been configured, it builds interface machines for each interface. When a router hears a hello message on an interface, it copies the router ID it has heard into the next hello packet. When a hello message is received with one’s own router ID in it, it goes through validity checks, such as making sure timers used by both routers match. Designing Routing and Switching Architectures Howard C. Berkowitz
26
Designing Routing and Switching Architectures
Howard C. Berkowitz
27
Dynamic Routing Using Link-State Routing Protocols
Designated Routers Interface on a router Designated router interface and backup designated router. DR election is part of the neighbor/adjacency discovery process. Most common for first router that intializes Second router that initializes-backup. Designing Routing and Switching Architectures Howard C. Berkowitz
28
Dynamic Routing Using Link-State Routing Protocols
OSPF Database Synchronization Master Role Slave Role Routers enter Exchange state and exchange database information. Master sends slave DD(Database Description)-contents of the master’s topological database. If master better than slave Slave installs If slaves better than master it sends an update to master which the master acks When no more changes occur the database is now sync’d Designing Routing and Switching Architectures Howard C. Berkowitz
29
Dynamic Routing Using Link-State Routing Protocols
OSPF Updating Topological information in an OSPF routing domain is considered to be owned by the router that initially advertises it. LSAs Data structures rather than explicit messages. Multiple LSAs in single protocol message. Designing Routing and Switching Architectures Howard C. Berkowitz
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.