Course 8 3. Routing Classification Properties Routing Protocols 1/39
Routing Algorithms Types Static versus dynamic Single-path versus multipath Flat versus hierarchical Host-intelligent versus router-intelligent Intradomain versus interdomain Link-state versus distance vector 2/39
Routing algorithms- depending on how the routes are calculated. STATIC ADAPTIVE Static algorithms (Non adaptive algorithms), do not base their routing decisions on the estimates of current traffic and topology. Instead the route is pre-computed and fed into the routers offline. Adaptive algorithms (Dynamic algorithms) on the other hand change their routing decisions to reflect the changes in the topology and usually in traffic as well. The various adaptive algorithms differ in : where they get their information (from adjacent routers, or from all routers), when they change the routes (when the load changes or when the topology changes) what metric is used for optimization (distance, number of hops, residual bandwidth). 3/39
Static routing The administrator is responsible for discovering and propagating routes through the network using an algorithm (which can be a link-state or a distance vector based one) as ex. Dijkstra or Belmann-Ford algorithm. After a device has been configured, it simply forwards packets out the predetermined ports. Relatively simple to administer in small networks with minimal redundancy Several disadvantages for maintaining IP routing tables: require a considerable amount of coordination and maintenance in non-trivial network environments. cannot dynamically adapt to the current network operational state. Traffic continues to be forwarded toward a destination that became unreachable (the routes pointing to that network remain in the routing table) unless the network administrator updates the configuration. Traffic is unable to use any alternate paths that may exist 4/39
Static routing can be attractive for example To manually define a default route. This route is used to forward traffic when the routing table does not contain a more specific route to the destination. To define a route that is not automatically advertised within a network. When complex routing policies are required. For example, static routes can be used to guarantee that traffic destined for a specific host traverses a designated network path. 5/39
To provide a more secure network environment. The administrator is aware of all subnetworks defined in the environment. The administrator specifically authorizes all communication permitted between these subnetworks. To provide more efficient resource utilization. This method of routing table management requires no network bandwidth to advertise routes between neighboring devices. It also uses less processor memory and CPU cycles to calculate network paths. As a conclusion the main advantage of static routing is its simplicity, and it should work well in a reliable network with a stable load. 6/39
Eg in figure below a central routing matrix is created, to be stored perhaps at a network control center. The matrix shows, for each source-destination pair of nodes, the identity of the next node on the route. Note that it is not necessary to store the complete route for each possible pair of nodes. It is sufficient to know, for each pair of nodes, the identity of the first node on the route( least cost route); In our example, the route from node 1 to node 6 begins by going through node 4. Consulting the matrix, the route from node 4 to node 6 goes through node 5. Finally, the route from node 5 to node 6 is a direct link to node 6. The complete route from node 1 to node 6 is 1-4-5-6. 7/39
8/39
With static routing, there is no difference between routing for datagrams and virtual circuits. All packets from a given source to a given destination follow the same route. A refinement to fixed routing that would accommodate link and node outages would be to supply the nodes with an alternate next node for each destination. For example, the alternate next nodes in the node 1 directory might be 4, 3, 2, 3, 3. 9/39
Adaptive Routing The routing decisions change as conditions on the network modify: Failure. When a node or trunk fails, it can no longer be used as part of a route. Congestion. When a particular portion of the network is heavily congested, it is desirable to route packets around, rather than through, the area of congestion. Information about the network state must be exchanged among the nodes. There is a tradeoff here between the quality of the information and the amount of overhead. The more information that is exchanged, and the more frequently it is exchanged, the better will be the routing decisions that each node makes. This information is itself a load on the network, causing a performance degradation. 10/39
Drawbacks of adaptive routing The routing decision is more complex; the processing load on nodes increases. In most cases, adaptive strategies depend on status information that is collected at one place but used at another; the network traffic load increases. An adaptive strategy may react too quickly, causing congestion-producing oscillation; if it reacts too slowly, the strategy will be irrelevant. 11/39
Adaptive routing strategies are the most prevalent, for two reasons: An adaptive routing strategy can improve performance, as seen by the network user. An adaptive routing strategy can aid in congestion control, as discussed later. These benefits may or may not be realized, depending on the soundness of the design and the nature of the load. Most major packet-switching networks, such as ARPANET and its successors, TYMNET, and those developed by IBM and DEC, have endured at least one major overhaul of their routing strategy. 12/39
Classification of adaptive routing strategies on the basis of information source: local e.g. a node routes each packet to the outgoing link with the shortest queue length, Q. This would have the effect of balancing the load on outgoing links adjacent nodes all nodes. As an example- the status of node 4 in next figure, above at a certain point in time. Node 4 has links to four other nodes. Decision is done by taking into account preferred direction - each link has for each destination i a bias B. 13/39
14/39
A number of packets have been arriving and a backlog has built up, with a queue of packets waiting for each of the outgoing links, as it can be seen. Improvement of above e.g- For each incoming packet is chosen the outgoing link that minimizes Q + B. Thus, a node tends to send packets in the right direction, with a concession made to current traffic delays A packet arrives from node 1 destined for node 6. With the biases and queues from figure below, the chosen route will be towards node 3, with the minimum value Q+B 6 15/39
Single Path Versus Multipath Some sophisticated routing protocols support multiple paths to the same destination. Unlike single-path algorithms, these multipath algorithms permit traffic multiplexing over multiple lines. The advantages of multipath algorithms -substantially better throughput reliability. This is generally called load sharing. 16/39
Examples Routing Scenarios Unicast- one source to one destination when a router receives a packet it forwards the packet only through one of its ports (the one belonging to the optimum path). Multicast- one source delivers a message to a group of destinations that have expressed interest in receiving the message; Broadcast -delivers a message to all nodes in the network; Anycast delivers a message to any one out of a group of nodes, typically the one nearest to the source. 17/39
Flat Versus Hierarchical In a flat routing system, the routers are peers of all others. In a hierarchical routing system (HS), some routers form a routing backbone. Packets from nonbackbone routers travel to the backbone routers, where they are sent through the backbone until they reach the general area of the destination. At this point, they travel from the last backbone router through one or more nonbackbone routers to the final destination. In HS some routers in a domain can communicate with routers in other domains, while others can communicate only with routers within their domain. In very large networks, additional hierarchical levels may exist, with routers at the highest hierarchical level forming the routing backbone. 18/39
The primary advantage of hierarchical routing is that it mimics the organization of most companies and therefore supports their traffic patterns well. Most network communication occurs within small company groups (domains). Because intradomain routers need to know only about other routers within their domain, their routing algorithms can be simplified and depending on the routing algorithm being used, routing update traffic can be reduced accordingly. 19/39
Routers 4, 5, 6, 10, 11, and 12 make up the backbone. If Host H1 in Area 3 wants to send a packet to Host H2 in Area 2, the packet is sent to Router 13, which forwards the packet to Router 12, which sends the packet to Router 11. Router 11 then forwards the packet along the backbone to Router 10, which sends the packet through two intra-area routers (Router 9 and Router 7) to be forwarded to Host H2 20/39
Intradomain Versus Interdomain For a large network a single routing protocol cannot handle the task of updating the routing table of all the routers. Interior routing- inside an autonomous system Exterior routing -between autonomous systems Some routing algorithms work only within domains; Others work within and between domains. The nature of these two algorithm types is different An optimal intradomain-routing algorithm would not necessarily be an optimal interdomain-routing algorithm 21/39
Application of IRP and ERP 22/39
Interior Router Protocol (IRP) Passes routing information between routers within AS The protocol used within the autonomous system does not need to be implemented outside of the system. This flexibility allows IRPs to be custom-tailored to specific applications and requirements. As e.g in previous figure, all of the LANs at a site, such as an office complex or campus, could be linked by routers to form an autonomous system. This system might be linked through a widearea network to other autonomous systems. In this case, the routing algorithm and routing tables used by routers in different autonomous systems may differ. 23/39
The routers in one autonomous system need a minimal level of information concerning networks that can be reached outside the system. The protocol used to pass routing information between routers in different autonomous systems is referred to as an exterior router protocol (ERP). IRP needs detailed model ERP supports summary information on reachability If a datagram is to be transferred from a host in one autonomous system to a host in another autonomous system, a router in the first system need only determine the target autonomous system and devise a route to get into that system. Once the datagram enters the target autonomous system, the routers there can cooperate to finally deliver the datagram. 24/39
Host Intelligent Versus Router Intelligent Some routing algorithms assume that the source end node will determine the entire route -usually referred to as source routing. In source-routing systems, routers merely act as storeand-forward devices, sending the packet to the next stop. Other algorithms assume that hosts know nothing about routes. Routers determine the path through the internetwork based on their own calculations. In the first system, the hosts have the routing intelligence. In the latter system, routers have the routing intelligence. 25/39
Dynamic routing Dynamic (adaptive) routing algorithms allow routers to automatically discover and maintain awareness of the paths through the network. Classification dependent on the way they discover and calculate new routes to destination networks. Distance vector protocols:eg. RIP, IGRP Link state protocols: OSPF Path vector protocols: BGP Hybrid protocols 26/39
Link State Versus Distance Vector Link-state algorithms (also known as shortest path first algorithms or Dijsktra alg.) flood routing information to all nodes in the internetwork. Each router, however, sends only the portion of the routing table that describes the state of its own links. In link-state algorithms, each router builds a picture of the entire network in its routing tables. Distance vector algorithms (also known as Bellman-Ford algorithms) call for each router to send all or some portion of its routing table, but only to its neighbors. In essence, link-state algorithms send small updates everywhere, while distance vector algorithms send larger updates only to neighboring routers. Distance vector algorithms know only about their neighbors. 27/39
Comparison between link state algorithms and distance vector algorithms converge more quickly less prone to routing loops than distance vector algorithms. require more CPU power and memory than distance vector algorithms. can be more expensive to implement and support. are generally more scalable than distance vector protocols. (Scalability is the ability to handle a growing amount of work, to be enlarged to accommodate this growth) 28/39
Path vector routing is somewhat similar to the distance vector algorithm in the sense that each border router advertises its reachable destinations to its neighboring router. (discussed in RFC 1322) It uses destination addresses and path descriptions to reach those destinations A route is defined as a pairing between a destination and the attributes of the path to that destination (the name path vector routing comes from it) A special path attribute records the sequence of routing domains through which the reachability information has passed. The preferred path to reach the destination is represented by the smallest number of domains 29/39
Several advantages most important flexibility The computational complexity is smaller than that of the link state protocol. The path vector computation consists of evaluating a newly arrived route and comparing it with the existing one, Path vector routing does not require domains to have homogeneous policies for route selection; The support for heterogeneous route selection policies has serious implications for the computational complexity. Each domain makes its route selection autonomously, based only on local policies. 30/39
Only the domains whose routes are affected by the changes have to recompute. Suppression of routing loops is implemented through the path attribute (in contrast to link state and distance vector, which use a globally-defined monotonically thereby increasing metric for route selection). Route computation precedes routing information dissemination. Only routing information associated with the routes selected by a domain is distributed to adjacent domains. Path vector routing has the ability to selectively hide information. 31/39
Disadvantages Because of the inclusion of full path information with each distance vector, the effect of a topology change can propagate farther than in traditional distance vector algorithms. Unless the network topology is fully meshed or is able to appear so, routing loops can become an issue. BGP is a popular example of a path vector routing protocol. 32/39
Hybrid routing Hybrid protocols attempt to combine the positive attributes of both distance vector and link state protocols. Like distance vector, hybrid protocols use metrics to assign a preference to a route. However, the metrics are more accurate than conventional distance vector protocols. Like link state algorithms, routing updates in hybrid protocols are event driven rather than periodic. Networks using hybrid protocols tend to converge more quickly than networks using distance vector protocols. Potentially reduce the costs of link state updates and distance vector advertisements. almost exclusively associated with the proprietary eg: Enhanced Interior Gateway Routing Protocol was developed by Cisco Systems, Inc. as an evolution of IGRP a distance vector algorithm used for Internet 33/39
Routing Metrics Routing algorithms have used many different metrics to determine the best route. Sophisticated routing algorithms can base route selection on multiple metrics, combining them in a single (hybrid) metric. All the following metrics have been used: Path length Reliability Delay Bandwidth Load Communication cost 34/39
Path length is the most common routing metric. 1. It is assigned a cost to each network link. The path length is the sum of the costs associated with each link traversed. 2. It is defined a hop count, a metric that specifies the number of passes through internetworking products, (e.g.routers) that a packet must take en route from a source to a destination. 35/39
Reliability refers to the dependability, usually described in terms of the bit-error rate, of each network link. Some network links might go down more often than others. After a network fails, certain network links might be repaired more easily or more quickly than other links. Any reliability factors can be taken into account in the assignment of the reliability ratings, which are arbitrary numeric values usually assigned to network links by network administrators. 36/39
Routing delay refers to the length of time required to move a packet from source to destination through the internetwork. It is a common and useful metric Delay is a conglomeration of many variables: the bandwidth of intermediate network links the port queues at each router along the way, network congestion on all intermediate network links the physical distance to be traveled 37/39
Bandwidth refers to the available traffic capacity of a link. All other things being equal, a 10-Mbps Ethernet link would be preferable to a 64-kbps leased line. Although bandwidth is a rating of the maximum attainable throughput on a link, routes through links with greater bandwidth do not necessarily provide better routes than routes through slower links. For example, if a faster link is busier, the actual time required to send a packet to the destination could be greater. 38/39
Load refers to the degree to which a network resource, such as a router, is busy. Load can be calculated in a variety of ways: CPU utilization packets processed per second. Monitoring these parameters on a continual basis can be resource-intensive itself. Communication cost is another important metric, especially because some companies may not care about performance as much as they care about operating expenditures. Although line delay may be longer, they will send packets over their own lines rather than through the public lines that cost money for usage time. 39/39