Chapter 8Routing

IntroductionLook at:Routing Basics (8.1)Address Resolution (8.2)Routing Protocols (8.3)Administrative Classification (8.4)Hierarchical Routing (8.5)

IntroductionWhen networking was first introduced, only a small number of devices were interconnected As networks grew, broadcasts and collisions caused significant issuesIn order to reduce broadcasts, Layer 3 devices, protocols, and addresses are used to segment the network

Routing BasicsReasons to segment a network at Layer 3 include:Creation of small troubleshooting areasCreation of small administrator management areasInterconnection of remote offices using WAN technologiesGrouping clients together with similar network resources

Routing BasicsA router is required to properly forward data between clients on different segments. A router is a device that forwards data based on a logical Layer 3 addressMany routers support the use of different protocols

Routing BasicsThe routing process usually occurs between physical network interfaces but can also be accomplished between logical network interfaces When a single physical network interface has more than one address assigned, it is said to have a logical interfaceA router can be any network device with the proper software to make routing decisions

Routing BasicsTwo key pieces of information are required for any device to route packets:A route to one or more networksA destination Layer 3 addressThe routes to a given network are stored in the memory of the router and are referred to as a routing table

Routing BasicsOn a Router there is an entry for each network and its corresponding network interface kept in memoryThe table allows the router to properly forward frames out the correct interface This occurs once the frame is received and the destination address has been processed

Routing BasicsIn standard routing, the route table is consulted every time aframe is received and so it plays a fundamental role in the proper delivery of dataA routing table only maintains the best possible route to a destination, not all possible routes

Routing BasicsRouting table entries have the following functions:Network destinationNetmaskGatewayInterface Metric

Routing BasicsThe task of populating the routing table is accomplished by using either dynamic routing or static routingDynamic routing uses routing protocols to build route tables automaticallyStatic routing requires manual route table entries and updates to different networks

Routing BasicsLayer 3 addressing is critical for end-to-end reachability and does not change throughout the routing process The Layer 2 address moves the packet from one hand-off point or hop to the nextThe routers Layer 2 address is the frames destination The Layer 3 address remains constant through each hop

Address ResolutionAddress resolution is the mapping of one address to another It is generally a mapping between a Layer 3 network address (logical) and a Layer 2 hardware address (physical)The reverse process is also address resolution

Address ResolutionAddress resolution is accomplished in one of the following ways: Table lookup is a rarely used method of address resolution Closed-form computation is only used in very specific networks and is time consuming to configure Dynamic message exchange is the most common and involves an exchange of information between two hosts

Address ResolutionAddress Resolution Protocol (ARP) is used when an IP host has a known destination IP address (Layer 3) and it needs to retrieve the corresponding Layer 2 MAC address from the destination host The ARP cache is used to further reduce the need for broadcasts by storing the IP-to- MAC mapping in memory for a specified duration

Address ResolutionThe ARP Process:Client A sends out an ARP broadcast All clients receive and process the broadcast frame but only Machine B respondsClient A receives the response and places Machine Bs MAC address in its ARP cache

Address ResolutionARP locates the Layer 2 address when the Layer 3 address is knownReverse Address Resolution Protocol (RARP) finds the Layer 3 address when the Layer 2 address is knownA good example of RARP is found in TCP/IP address reservations and the Boot Protocol (BootP) Using BootP, IP hosts are automatically assigned their configuration information through a BootP server

Routing ProtocolsDynamic routing using routing protocols Purpose of routing protocols is to build a routing table with the best routes Routing protocols are categorized into two types:Distance VectorLink State

Routing ProtocolsDistance vector routing protocols are simple Generally they are easy to configureThey use simple logic to determine the best path to a given destinationThe term metric refers to the method or measurement used by the routing protocol logic to determine the best path to a given network

Routing ProtocolsA distance vector routing protocol usually uses hop count as its metric A distance vector routing protocol is characterized by how it communicates with other routing devicesDistance vector routing protocols use broadcasts to advertise their entire routing table to directly connected peer routers

Routing ProtocolsConvergence is the time it takes for a given set of routers to learn routes to all networks Convergence describes the time it takes a set of routers to learn of a change in the networkDistance vector routing protocols generally take longer to converge than link state protocols because they use a periodic route advertisement schedule

Routing ProtocolsA routing loop occurs when routers get confused during update operations, causing frames to bounce back and forth between a set of interfacesTwo easy methods to identify routing loops:Tracert or traceroute TCP/IP utilitiesView the routing table and the metric associated with the network

Routing ProtocolsPrevent routing loops by using the following software based methods:Split horizon Hold-down timers Triggered updates Hop count limits Poisoning

Routing ProtocolsLink state routing protocols are more intelligent than distance vector protocols The metric used by most link state protocols is bandwidth allowing more complex routing configurations Routing protocols capable of making complex decisions use a mathematical formula or algorithm for deriving the best path or route to a given network

Routing ProtocolsSome link state protocols are capable of determining the best route to a destination network based on the following:Delay Load Reliability MTU

Routing ProtocolsWhen more than one metric is used it is referred to as a composite metricLink state protocols only send updates when changes occur, and they only send the changes, not the entire route tableLink state protocols use multicast and unicast traffic instead of broadcast traffic Link state routers also develop an overall picture of the networks available by establishing neighbor relationships

Administrative ClassificationRouting protocols are also separated by an administrative classification based on where they are used in the networking environment:Interior routing protocols or interior gateway protocolsExterior routing protocols or exterior gateway protocols

Administrative ClassificationInterior gateway protocols (IGPs) are used within a companys network infrastructure to maintain routing tables and policies set by the network administrators The two industry standard IGPs are: Routing Information ProtocolOpen Shortest Path First

Administrative ClassificationRIP is a distance vector protocol that uses hop count for its metric when determining the best route to a given network In most implementations, RIP uses split horizon, hop count limit, and poisoning for routing loop prevention RIP is a classful routing protocol

Administrative ClassificationThe shortest path as measured by Open Shortest Path First (OSPF) is actually the fastest path based on bandwidth Shortest refers to the shortest timeOSPF is used in large networks and ones requiring more intelligence than distance vector routing protocols

Administrative ClassificationOSPF communicates using unicast and multicast packetsIt only transmits changes or updates to the routing table when they occur It uses hello packets to determine the current state of a link between itself and its neighborsIt utilizes a link state database to maintain a local view of the entire routing environment

Administrative ClassificationThe configuration possibilities using OSPFAreasAutonomous system (AS) Backbone router Area border router (ABR)Autonomous system boundary router (ASBR)

Administrative ClassificationThe decision making process of EGPs is far more complex than that of internal protocols The power and routing flexibility associated with EGPs requires knowledge and understanding of the complex nature of your network and its traffic EGPs can let you influence and manage traffic only as it enters or leaves your AS

Administrative ClassificationOne member of EGPs is the Exterior Gateway Protocol (EGP)EGP was the first protocol developed that allowed isolation of autonomous systemsEPG is not used today and is replaced by the Border Gateway Protocol

Administrative ClassificationBorder Gateway Protocol (BGP) version 4 is the most widely used exterior protocol in the world BGP is a well established standard and commonly used by ISPs and in very large companies there are actually two different classifications of BGP internalexternal

Administrative ClassificationiBGP is used for internal routing eBGP is used for external routingiBGP functions under different rules than eBGPIf two routers running BGP are in the same AS, they are running an iBGP connection

Administrative ClassificationBGP communication starts by establishing peers Once the peers have been established, BGP routing information is exchanged and updated as necessary BGP is an advanced distance vector protocol that uses triggered updates for communicating changes in the routing environment Routing loops in BGP are avoided by using the AS-path attribute

Administrative ClassificationBGP uses active TCP sessions that are setup and continuously maintained. Convergence in the routing environment is very fastBGP has features that you can use to help speed the convergence of the network routes under your control

Hierarchical RoutingHierarchical routing depends on hierarchical addressing It is a routing technique originally designed to help reduce the size of the routing tables on the Internet as well as speed up the overall routing processThe concept uses an address block or blocks to represent different sections of a network

Hierarchical RoutingSummarizing routes is often referred to as supernetting networksThe process of summarization is built around the binary bit patterns just as in subnetting The difference is that rather than extending the subnet mask by adding bits, we remove bits

Hierarchical RoutingBy using summarization, you reduce the routing tables on each router To accomplish the summarization, you need to determine how many bits to unmask or un-subnet in order to make the networks appear as one big address block The routing protocol must transmit the network prefix along with the network address during route advertisements