Lecture 6 (Networks) PDF

Systems and Architecture WAN Technologies and Routing G51SYS Objectives  To explain the basic functions of a router  To explain the fundamental concept of routing and show how routing is used in such network.  To explain the following routing techniques: Adaptive ro...

Systems and Architecture WAN Technologies and Routing G51SYS Objectives  To explain the basic functions of a router  To explain the fundamental concept of routing and show how routing is used in such network.  To explain the following routing techniques: Adaptive routing and Fixed routing Centralized routing Distributed routing Isolated routing Flooding Dijkstra’s least cost algorithm SYS.2 Router  What’s inside a router input ports, switching, output ports buffer management, scheduling  IP: the Internet Protocol datagram format addressing network address translation IPv6 SYS.3 Router architecture overview high-level view of generic router architecture: routing, management routing control plane (software) processor operates in millisecond time frame forwarding data plane (hardware) operates in nanosecond timeframe router input ports router output ports SYS.4 Network-layer services and protocols  transport segment from sending to receiving host sender: encapsulates segments into datagrams, passes to link layer receiver: delivers segments to transport mobile network layer protocol national or global ISP  network layer protocols in every Internet applicati on device: hosts, routers transpor t network link netwo netwo physical rk  routers: rk link link physic physic al al examines header fields in all IP datagrams netwo passing through it rk link netwo rk physic link netwo moves datagrams from input ports to al physic rk datacenter al link network output ports to transfer datagrams along physic al end-end path applicati on transpor enterprise t network network link physical SYS.5 Two key network-layer functions network-layer functions: forwarding: move packets from a router’s input link to analogy: taking a trip appropriate router output link  forwarding: process of getting through single interchange  routing: process of planning trip from source to destination  routing: determine route taken by packets from source to destination routing algorithms forwarding routing SYS.6 Network layer: data plane, control plane Data plane: Control plane local,per-router network-wide logic function determines how determines how datagram is routed datagram arriving on among routers along router input port is end-end path from forwarded to router source host to output port destination host  two control-plane approaches: values in arriving packet header traditional routing algorithms: implemented in routers 0111 1 software-defined networking 2 3 (SDN): implemented in (remote) servers SYS.7 Per-router control plane Individual routing algorithm components in each and every router interact in the control plane Routing Algorithm control plane data plane values in arriving packet header 0111 1 2 3 SYS.8 Software-Defined Networking (SDN) control plane Remote controller computes, installs forwarding tables in routers Remote Controller control plane data plane CA CA CA CA CA values in arriving packet header 0111 1 2 3 SYS.9 Routing In A WAN  Routing is the process of discovering network paths Decide what to optimize (e.g., fairness vs efficiency) Update routes for changes in topology (e.g., failures) Forwarding is the process of sending of packets along a path SYS.10 Routing protocols Routing protocol goal: determine “good” paths (equivalently, routes), mobile network from sending hosts to receiving host, national or global ISP through network of routers applicati path: sequence of routers packets on transpor traverse from given initial source host t network to final destination host link netwo netwo physical rk rk link link physic physic al al netwo “good”: least “cost”, “fastest”, “least rk link netwo rk congested” physic link netwo al physic rk datacenter al link network physic routing: a “top-10” networking al challenge! applicati on transpor enterprise t network network link physical SYS.11 Routing in a WAN A WAN can be easily modeled as a graph. SYS.12 Types of Routing Static routing (Fig 6.4)  Same route used from source to destination  Route is fixed  Data may not arrive if path is blocked Dynamic (adaptive) routing  Route changes based on conditions on the network  The network software selects the best route Broadcast routing  Transmit to all nodes in the network (flooding) SYS.13 Static routing  In static routing, routing tables remain the same when network changes occur.  Maintaining static routing tables is simpler than maintaining adaptive routing tables.  Nodes do not have to share information with other nodes, since the routing tables are not going to change. SYS.14 SYS.15 Adaptive routing  A dynamic system in which routing tables react to network fluctuations, such as congestion and node/link failure.  As a problem occurs in the network, the appropriate information is transmitted to the routing tables, and new routes that avoid the problem areas are created. SYS.16 Advantages and disadvantages  Static routing The chief advantages of static routing are simplicity and low overhead The chief disadvantage is inflexibility -- static routes cannot be changed easily. Most NW s use dynamic routing due to of their advantages.  Dynamic (Adaptive) routing Advantages: improved performance aid congestion control but since is a complex system, may not realize theoretical benefits Disadvantages: decisions more complex tradeoff between quality of network info and overhead reacting too quickly can cause oscillation reacting too slowly means info may be irrelevant SYS.17 Routing Techniques  Flooding  Centralized routing and Distributed routing  Isolated routing SYS.18  Flooding  A packet is sent by a source node to every one of its neighbors.  At each node, an incoming packet is retransmitted on all outgoing links except for the link on which it arrived.  Eventually, a number of copies of the packet will arrive at the destination.  The packet must have some unique identifier (e.g., source node and sequence number, or virtual circuit number and sequence number) so that the destination knows to discard all but the first copy.  The number of packets in circulation just from a single source packet grows without bound.  Prevented by identifying the duplicate copies of the packet arrive, they are discarded.  A simpler technique is to include a hop count field with each packet. Flooding Example SYS.19 Centralized The routing information generated from the least cost algorithm is stored at a central location within the network. Any node wishing to transmit a packet to another node consults this centralized site and inquires as to the best route on which to transmit the data. The central site might maintain a least cost routing table such as that shown in table. This table was created by performing the least cost algorithm seven times using each node in the network as the starting node. SYS.20 Distributed routing Allows each node to maintain its own routing table. When a data packet enters the network at node x, that node consults its own routing table to determine the next node to receive the packet. Each node needs routing information only for its own locale. SYS.21  Isolated routing Each node uses only local information to create its own routing table. One technique a node can use to gather routing information is to observe the status information of incoming packets. Suppose all nodes maintain a clock system and each packet is time-stamped as it leaves a node. Then, when a packet arrives at the receiving node, the receiver can compare the time stamp of the packet with the current time and calculate the time the packet spent in transit. SYS.22  Routing algorithm classification global: all routers have complete topology, link cost info How “link state” algorithms fast do routes dynamic: routes change? static: routes change more quickly change slowly over periodic updates or time in response to link cost changes decentralized: iterative process of computation, exchange of info with neighbors routers initially only know link costs to attached neighbors “distance vector” algorithms global or decentralized information? SYS.23 Shortest Path Computation in a Graph  Dijkstra's Algorithm (DA) finds the distance along a shortest path from a single source node to each of the other nodes in the graph; a next-hop RT is constructed during the computation of shortest paths  The algorithm must be run once for each RT  Distance represented by weights on edges.  Distance between two nodes to be the sum of the weights along a path between the nodes SYS.24 Distance Vector Routing & Link-State Routing Distance Vector Routing (DVR) One of the best-known algorithms for distributed route computation. Messages sent from one PS (Packet Switch) to another contain pairs of values. Each specify a destination and a distance to that destination. Each link in the network is assigned a weight, and the distance to a destination between two packet switches is defined to be the sum of weights along the path between the two. In DVR scheme, a router send a complete list of destinations and the current cost of reaching each network that can be reached. A PS periodically sends routing information across the network to neighbors. Each message contains pairs of ( destination, distance ) SYS.25 Distance Vector Routing & Link-State Routing Link-State Routing (LSR) Uses Dijkstra’s algorithm: Packet Switches periodically send messages across the network that carries the status of a link between two packet switches. Each status message is broadcast to all switches. Every switch runs software that collects incoming status messages and uses them to build a graph of the network. Inaddition, OSPF has the advantage that all computations can be carried out simultaneously After the status of a link changes, all PS receive a status message and each switch begins computing its RT In contrast, a DVA it requires a PS to update its RT, before sending a message to another PS SYS.26 Reference  Chapters 13 and 18. Computer Networks and Internets, Douglas E. Comer, Prentice Hall. 5th Edition.  Chapter 10. Data and Computer Communications 8th Edition by William Stallings SYS.27