Chapter 4: Network Layer. Chapter 4 Network Layer. Chapter 4: Network Layer. Key Network-Layer Functions. Network layer.

Transcription

1 Chapter 4 Network Laer note on the use of these ppt slides: We re making these slides freel available to all (facult, students, readers). The re in PowerPoint form so ou can add, modif, and delete slides (including this one) and slide content to suit our needs. The obviousl represent a lot of work on our part. In return for use, we onl ask the following: If ou use these slides (e.g., in a class) in substantiall unaltered form, that ou mention their source (after all, we d like people to use our book!) If ou post an slides in substantiall unaltered form on a www site, that ou note that the are adapted from (or perhaps identical to) our slides, and note our copright of this material. Thanks and enjo! JFK/KWR ll material copright J.F Kurose and K.W. Ross, ll Rights Reserved Computer Networking: Top Down pproach Featuring the, rd edition. Jim Kurose, Keith Ross ddison-wesle, Jul 004. Network Laer 4- Chapter goals: understand principles behind laer services: routing (path selection) dealing with scale how a works advanced topics: IPv6, mobilit instantiation and implementation in the Network Laer 4-4. Introduction 4. Virtual circuit and 4.4 IP: Network laer transport segment from sending to receiving host on sending side encapsulates segments into datagrams on rcving side, delivers segments to transport laer laer protocols in ever host, Router eamines header fields in all IP datagrams passing through it application transport data link data link data link data link data link data link data link data link data link application transport data link Network Laer 4- Network Laer 4-4 Ke Network-Laer Functions Interpla between routing and forwarding forwarding: move packets from s input to appropriate output routing: determine route taken b packets from source to dest. Routing algorithms analog: routing: process of planning trip from source to dest forwarding: process of getting through single interchange value in arriving packet s header routing algorithm local forwarding table header value output link Network Laer 4-5 Network Laer 4-6

2 Connection setup rd important function in some architectures: TM, frame rela, X.5 Before datagrams flow, two hosts and intervening s establish virtual connection Routers get involved Network and transport laer cnctn service: Network: between two hosts Transport: between two processes Network service model Q: What service model for channel transporting datagrams from sender to rcvr? Eample services for individual datagrams: guaranteed deliver Guaranteed deliver with less than 40 msec dela Eample services for a flow of datagrams: In-order datagram deliver Guaranteed minimum bandwidth to flow Restrictions on changes in interpacket spacing Network Laer 4-7 Network Laer 4-8 Network laer service models: Network rchitecture TM TM TM TM Service Model best effort CBR VBR BR UBR Guarantees? Bandwidth Loss Order Timing none constant rate guaranteed rate guaranteed minimum none no es es no no no es es es es no es es no no Congestion feedback no (inferred via loss) no congestion no congestion es no 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-9 Network Laer 4-0 Network laer connection and connection-less service Datagram provides - laer connectionless service VC provides - laer connection service nalogous to the transport- laer services, but: Service: host-to-host No choice: provides one or the other Implementation: in the core Virtual circuits source-to-dest path behaves much like telephone circuit performance-wise actions along source-to-dest path call setup, teardown for each call before data can flow each packet carries VC identifier (not destination host address) ever on source-dest path maintains state for each passing connection link, resources (bandwidth, buffers) ma be allocated to VC Network Laer 4- Network Laer 4-

3 VC implementation Forwarding table VC number VC consists of:. Path from source to destination. VC numbers, one number for each link along path. Entries in forwarding tables in s along path Packet belonging to VC carries a VC number. VC number must be changed on each link. New VC number comes from forwarding table Network Laer 4- Forwarding table in northwest : interface number Incoming interface Incoming VC # Outgoing interface Outgoing VC # Routers maintain connection state information! Network Laer 4-4 Virtual circuits: signaling protocols used to setup, maintain teardown VC used in TM, frame-rela, X.5 not used in toda s Datagram s no call setup at laer s: no state about end-to-end connections no -level concept of connection packets forwarded using destination host address packets between same source-dest pair ma take different paths application transport data link 5. Data flow begins 6. Receive data 4. Call connected. ccept call. Initiate call. incoming call application transport data link application transport data link application transport. Send data. Receive data data link Network Laer 4-5 Network Laer 4-6 Forwarding table 4 billion possible entries Longest prefi matching Destination ddress Range through Link Interface Prefi Match Link Interface otherwise through through otherwise Eamples D: D: Which interface? Which interface? Network Laer 4-7 Network Laer 4-8

4 Datagram or VC : wh? TM data echange among evolved from telephon computers human conversation: elastic service, no strict strict timing, reliabilit timing req. requirements smart end sstems need for guaranteed (computers) service can adapt, perform dumb end sstems control, error recover telephones simple inside, compleit at edge compleit inside man link tpes different characteristics uniform service difficult Network Laer Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-0 Router rchitecture Overview Two ke functions: run routing algorithms/protocol (RIP, OSPF, BGP) forwarding datagrams from incoming to outgoing link Input Port Functions Phsical laer: bit-level reception Data link laer: e.g., Ethernet see chapter 5 Decentralied switching: given datagram dest., lookup output port using forwarding table in input port memor goal: complete input port processing at line speed queuing: if datagrams arrive faster than forwarding rate into switch fabric Network Laer 4- Network Laer 4- Three tpes of switching fabrics Switching Via Memor First generation s: traditional computers with switching under direct control of CPU packet copied to sstem s memor speed limited b memor bandwidth ( bus crossings per datagram) Input Port Memor Output Port Sstem Bus Network Laer 4- Network Laer 4-4

5 Switching Via a Bus Switching Via n Interconnection Network datagram from input port memor to output port memor via a shared bus bus contention: switching speed limited b bus bandwidth Gbps bus, Cisco 900: sufficient speed for access and enterprise s (not regional or backbone) overcome bus bandwidth limitations Banan s, other interconnection nets initiall developed to connect processors in multiprocessor dvanced design: fragmenting datagram into fied length cells, switch cells through the fabric. Cisco 000: switches Gbps through the interconnection Network Laer 4-5 Network Laer 4-6 Output Ports Output port queueing Buffering required when datagrams arrive from fabric faster than the transmission rate Scheduling discipline chooses among queued datagrams for transmission Network Laer 4-7 buffering when arrival rate via switch eceeds output line speed queueing (dela) and loss due to output port buffer overflow! Network Laer 4-8 Input Port Queuing Fabric slower than input ports combined -> queueing ma occur at input queues Head-of-the-Line (HOL) blocking: queued datagram at front of queue prevents others in queue from moving forward queueing dela and loss due to input buffer overflow! 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-9 Network Laer 4-0

6 The Network laer Host, laer functions: Network laer Routing protocols path selection RIP, OSPF, BGP Transport laer: TCP, UDP forwarding table Link laer laer IP protocol addressing conventions datagram format packet handling conventions ICMP protocol error reporting signaling 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4- Network Laer 4- IP datagram format IP protocol version number header length (btes) tpe of data ma number remaining hops (decremented at each ) upper laer protocol to deliver paload to how much overhead with TCP? 0 btes of TCP 0 btes of IP = 40 btes + app laer overhead bits ver head. tpe of len service length 6-bit identifier flgs fragment offset time to upper live laer checksum bit source IP address bit destination IP address Options (if an) data (variable length, tpicall a TCP or UDP segment) total datagram length (btes) for fragmentation/ reassembl E.g. timestamp, record route taken, specif list of s to visit. Network Laer 4- IP Fragmentation & Reassembl links have MTU (ma.transfer sie) - largest possible link-level frame. different link tpes, different MTUs large IP datagram divided ( fragmented ) within net one datagram becomes several datagrams reassembled onl at final destination IP header bits used to identif, order related fragments reassembl fragmentation: in: one large datagram out: smaller datagrams Network Laer 4-4 IP Fragmentation and Reassembl Eample 4000 bte datagram MTU = 500 btes 480 btes in data field offset = 480/8 length =4000 ID = length =500 length =500 length =040 fragflag =0 ID = ID = ID = offset =0 One large datagram becomes several smaller datagrams fragflag = fragflag = fragflag =0 offset =0 offset =85 offset =70 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-5 Network Laer 4-6

7 IP ddressing: introduction IP address: -bit identifier for host, interface interface: connection between host/ and link s tpicall have multiple interfaces host ma have multiple interfaces IP addresses associated with each interface = Subnets IP address: subnet part (high order bits) host part (low order bits) What s a subnet? device interfaces with same subnet part of IP address can l reach each other without intervening LN... consisting of subnets Network Laer 4-7 Network Laer 4-8 Subnets...0/4...0/4 Subnets... Recipe To determine the subnets, detach each interface from its host or, creating islands of isolated s. Each isolated is called a subnet....0/4 Subnet mask: /4 How man? Network Laer 4-9 Network Laer 4-40 IP addressing: CIDR CIDR: Classless InterDomain Routing subnet portion of address of arbitrar length address format: a.b.c.d/, where is # bits in subnet portion of address subnet host part part / Network Laer 4-4 IP addresses: how to get one? Q: How does host get IP address? hard-coded b sstem admin in a file Wintel: control-panel->->configuration- >tcp/ip->properties UNIX: /etc/rc.config DHCP: Dnamic Host Configuration : dnamicall get address from as server plug-and-pla (more in net chapter) Network Laer 4-4

8 IP addresses: how to get one? Q: How does get subnet part of IP addr? : gets allocated portion of its provider ISP s address space ISP's block /0 Organiation / Organiation / Organiation / Organiation / Hierarchical addressing: route aggregation Hierarchical addressing allows efficient advertisement of routing information: Organiation / Organiation / Organiation / Organiation /.. Fl-B-Night-ISP ISPs-R-Us Send me anthing with addresses beginning /0 Send me anthing with addresses beginning /6 Network Laer 4-4 Network Laer 4-44 Hierarchical addressing: more specific routes ISPs-R-Us has a more specific route to Organiation Organiation / Organiation / Organiation /.... Organiation / Fl-B-Night-ISP ISPs-R-Us Send me anthing with addresses beginning /0 Send me anthing with addresses beginning /6 or / IP addressing: the last word... Q: How does an ISP get block of addresses? : ICNN: Corporation for ssigned Names and Numbers allocates addresses manages DNS assigns domain names, resolves disputes Network Laer 4-45 Network Laer 4-46 NT: Network ddress Translation NT: Network ddress Translation rest of ll datagrams leaving local have same single source NT IP address: , different source port numbers local (e.g., home ) 0.0.0/4 Datagrams with source or destination in this have 0.0.0/4 address for source, destination (as usual) Motivation: local uses just one IP address as far as outside word is concerned: no need to be allocated range of addresses from ISP: - just one IP address is used for all devices can change addresses of devices in local without notifing outside world can change ISP without changing addresses of devices in local devices inside local net not eplicitl addressable, visible b outside world (a securit plus). Network Laer 4-47 Network Laer 4-48

9 NT: Network ddress Translation Implementation: NT must: outgoing datagrams: replace (source IP address, port #) of ever outgoing datagram to (NT IP address, new port #)... remote clients/servers will respond using (NT IP address, new port #) as destination addr. remember (in NT translation table) ever (source IP address, port #) to (NT IP address, new port #) translation pair incoming datagrams: replace (NT IP address, new port #) in dest fields of ever incoming datagram with corresponding (source IP address, port #) stored in NT table Network Laer 4-49 NT: Network ddress Translation : NT changes datagram source addr from , 45 to , 500, updates table NT translation table WN side addr LN side addr , , 45 S: , 500 D: , S: , 80 D: , 500 : Repl arrives dest. address: , S: , 45 D: , 80 S: , 80 D: , 45 4 : host sends datagram to , : NT changes datagram dest addr from , 500 to , 45 Network Laer 4-50 NT: Network ddress Translation 6- bit port- number field: 60,000 simultaneous connections with a single LN-side address! NT is controversial: s should onl process up to laer violates end-to-end argument NT possibilit must be taken into account b app designers, eg, PP applications address shortage should instead be solved b IPv6 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-5 Network Laer 4-5 ICMP: Control Message Traceroute and ICMP used b hosts & s to communicate -level information error reporting: unreachable host,, port, protocol echo request/repl (used b ping) -laer above IP: msgs carried in IP datagrams ICMP message: tpe, code plus first 8 btes of IP datagram causing error Tpe Code description 0 0 echo repl (ping) 0 dest. unreachable dest host unreachable dest protocol unreachable dest port unreachable 6 dest unknown 7 dest host unknown 4 0 source quench (congestion control - not used) 8 0 echo request (ping) 9 0 route advertisement 0 0 discover 0 TTL epired 0 bad IP header Source sends series of UDP segments to dest First has TTL = Second has TTL=, etc. Unlikel port number When nth datagram arrives to nth : Router discards datagram nd sends to source an ICMP message (tpe, code 0) Message includes name of & IP address When ICMP message arrives, source calculates RTT Traceroute does this times Stopping criterion UDP segment eventuall arrives at destination host Destination returns ICMP host unreachable packet (tpe, code ) When source gets this ICMP, stops. Network Laer 4-5 Network Laer 4-54

10 IPv6 4. Introduction 4. Virtual circuit and 4.4 IP: Initial motivation: - bit address space soon to be completel allocated. dditional motivation: header format helps speed processing/forwarding header changes to facilitate QoS IPv6 datagram format: fied-length 40 bte header no fragmentation allowed Network Laer 4-55 Network Laer 4-56 IPv6 Header (Cont) Priorit: identif priorit among datagrams in flow Flow Label: identif datagrams in same flow. (concept of flow not well defined). Net header: identif upper laer protocol for data Other Changes from IPv4 Checksum: removed entirel to reduce processing time at each hop Options: allowed, but outside of header, indicated b Net Header field ICMPv6: new version of ICMP additional message tpes, e.g. Packet Too Big multicast group management functions Network Laer 4-57 Network Laer 4-58 Transition From IPv4 To IPv6 Not all s can be upgraded simultaneous no flag das How will the operate with mied IPv4 and IPv6 s? Tunneling: IPv6 carried as paload in IPv4 datagram among IPv4 s Tunneling Logical view: Phsical view: B E F tunnel IPv6 IPv6 IPv6 IPv6 B C D E F IPv6 IPv6 IPv4 IPv4 IPv6 IPv6 Flow: X Src: Dest: F data Src:B Dest: E Flow: X Src: Dest: F Src:B Dest: E Flow: X Src: Dest: F Flow: X Src: Dest: F data data data Network Laer to-b: IPv6 B-to-C: IPv6 inside IPv4 B-to-C: IPv6 inside IPv4 E-to-F: IPv6 Network Laer 4-60

11 Interpla between routing and forwarding 4. Introduction 4. Virtual circuit and 4.4 IP: value in arriving packet s header routing algorithm local forwarding table header value output link Network Laer 4-6 Network Laer 4-6 Graph abstraction u Graph: G = (N,E) 5 v w 5 N = set of s = { u, v, w,,, } E = set of links ={ (u,v), (u,), (v,), (v,w), (,w), (,), (w,), (w,), (,) } Remark: Graph abstraction is useful in other contets Eample: PP, where N is set of peers and E is set of TCP connections Graph abstraction: costs u 5 v w 5 c(, ) = cost of link (, ) -e.g., c(w,) = 5 cost could alwas be, or inversel related to bandwidth, or inversel related to congestion Cost of path (,,,, p ) = c(, ) + c(, ) + + c( p-, p ) Question: What s the least-cost path between u and? Routing algorithm: algorithm that finds least-cost path Network Laer 4-6 Network Laer 4-64 Routing lgorithm classification Global or decentralied information? Global: all s have complete topolog, link cost info link state algorithms Decentralied: knows lconnected neighbors, link costs to neighbors iterative process of computation, echange of info with neighbors distance vector algorithms Static or dnamic? Static: routes change slowl over time Dnamic: routes change more quickl periodic update in response to link cost changes 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-65 Network Laer 4-66

12 Link-State Routing lgorithm Dijsktra s lgorithm Dijkstra s algorithm net topolog, link costs known to all nodes accomplished via link state broadcast all nodes have same info computes least cost paths from one node ( source ) to all other nodes gives forwarding table for that node iterative: after k iterations, know least cost path to k dest. s Notation: c(,): link cost from node to ; = if not direct neighbors D(v): current value of cost of path from source to dest. v p(v): predecessor node along path from source to v N': set of nodes whose least cost path definitivel known Network Laer 4-67 Initialiation: N' = {u} for all nodes v 4 if v adjacent to u 5 then D(v) = c(u,v) 6 else D(v) = 7 8 Loop 9 find w not in N' such that D(w) is a minimum 0 add w to N' update D(v) for all v adjacent to w and not in N' : D(v) = min( D(v), D(w) + c(w,v) ) /* new cost to v is either old cost to v or known 4 shortest path cost to w plus cost from w to v */ 5 until all nodes in N' Network Laer 4-68 Dijkstra s algorithm: eample Dijkstra s algorithm, discussion Step N' u u u uv uvw uvw D(v),p(v),u,u,u D(w),p(w) 5,u 4,,, D(),p(),u D(),p(), D(),p() 4, 4, 4, lgorithm compleit: n nodes each iteration: need to check all nodes, w, not in N n(n+)/ comparisons: O(n ) more efficient implementations possible: O(nlogn) Oscillations possible: e.g., link cost = amount of carried traffic u 5 v w 5 +e D 0 0 B 0 e C e initiall +e 0 D +e B 0 C 0 recompute routing 0 +e D 0 0 B C +e recompute +e 0 D B 0 +e e C recompute Network Laer 4-69 Network Laer 4-70 Distance Vector lgorithm () 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-7 Bellman- Ford Equation (dnamic programming) Define d () := cost of least- cost path from to Then d () = min {c(,v) + d v () } where min is taken over all neighbors of Network Laer 4-7

13 Bellman-Ford eample () u 5 v w 5 Clearl, d v () = 5, d () =, d w () = B-F equation sas: d u () = min { c(u,v) + d v (), c(u,) + d (), c(u,w) + d w () } = min { + 5, +, 5 + } = 4 Node that achieves minimum is net hop in shortest path forwarding table Network Laer 4-7 Distance Vector lgorithm () D () = estimate of least cost from to Distance vector: D = [D (): є N ] Node knows cost to each neighbor v: c(,v) Node maintains D = [D (): є N ] Node also maintains its neighbors distance vectors For each neighbor v, maintains D v = [D v (): є N ] Network Laer 4-74 Distance vector algorithm (4) Distance Vector lgorithm (5) Basic idea: Each node periodicall sends its own distance vector estimate to neighbors When node a node receives new DV estimate from neighbor, it updates its own DV using B-F equation: D () min v {c(,v) + D v ()} for each node N Under minor, natural conditions, the estimate D () converge the actual least cost d () Iterative, asnchronous: each local iteration caused b: local link cost change DV update message from neighbor Distributed: each node notifies neighbors onl when its DV changes neighbors then notif their neighbors if necessar Each node: wait for (change in local link cost of msg from neighbor) recompute estimates if DV to an dest has changed, notif neighbors Network Laer 4-75 Network Laer 4-76 node table cost to 0 7 node table cost to from from from 0 node table cost to 7 0 D () = min{c(,) + D (), c(,) + D ()} = min{+0, 7+} = from from from cost to cost to cost to from from from cost to cost to cost to time D () = min{c(,) + D (), c(,) + D ()} = min{+, 7+0} = 7 Network Laer 4-77 Distance Vector: link cost changes Link cost changes: node detects local link cost change updates routing info, recalculates distance vector if DV changes, notif neighbors good news travels fast 4 50 t time t 0, detects the link-cost change, updates its DV, and informs its neighbors. t time t, receives the update from and updates its table. It computes a new least cost to and sends its neighbors its DV. t time t, receives s update and updates its distance table. s least costs do not change and hence does not send an message to. Network Laer 4-78

14 Distance Vector: link cost changes Comparison of LS and DV algorithms Link cost changes: good news travels fast bad news travels slow - count to infinit problem! 44 iterations before algorithm stabilies: see tet Poissoned reverse: If Z routes through Y to get to X : Z tells Y its (Z s) distance to X is infinite (so Y won t route to X via Z) will this completel solve count to infinit problem? Network Laer 4-79 Message compleit LS: with n nodes, E links, O(nE) msgs sent DV: echange between neighbors onl convergence time varies Speed of Convergence LS: O(n ) algorithm requires O(nE) msgs ma have oscillations DV: convergence time varies ma be routing loops count-to-infinit problem Robustness: what happens if malfunctions? LS: node can advertise incorrect link cost each node computes onl its own table DV: DV node can advertise incorrect path cost each node s table used b others error propagate thru Network Laer 4-80 Hierarchical Routing 4. Introduction 4. Virtual circuit and 4.4 IP: Our routing stud thus far - idealiation all s identical flat not true in practice scale: with 00 million destinations: can t store all dest s in routing tables! routing table echange would swamp links! administrative autonom internet = of s each admin ma want to control routing in its own Network Laer 4-8 Network Laer 4-8 Hierarchical Routing Interconnected Ses aggregate s into regions, autonomous sstems (S) s in same S run same routing protocol intra-s routing protocol s in different S can run different intra- S routing protocol Gatewa Direct link to in another S c a b S a c d b Intra-S Routing algorithm S Forwarding table Inter-S Routing algorithm c a b S Forwarding table is configured b both intra- and inter-s routing algorithm Intra-S sets entries for internal dests Inter-S & Intra-s sets entries for eternal dests Network Laer 4-8 Network Laer 4-84

15 Inter-S tasks Suppose in S receives datagram for which dest is outside of S Router should forward packet towards on of the gatewa s, but which one? c a b S a c d b S needs:. to learn which dests are reachable through S and which through S. to propagate this reachabilit info to all s in S Job of inter-s routing! S c a b S Network Laer 4-85 Eample: Setting forwarding table in d Suppose S learns from the inter- S protocol that subnet is reachable from S (gatewa c) but not from S. Inter- S protocol propagates reachabilit info to all internal s. Router d determines from intra- S routing info that its interface I is on the least cost path to c. Puts in forwarding table entr (,I). Network Laer 4-86 Eample: Choosing among multiple Ses Now suppose S learns from the inter-s protocol that subnet is reachable from S and from S. To configure forwarding table, d must determine towards which gatewa it should forward packets for dest. This is also the job on inter-s routing protocol! Hot potato routing: send packet towards closest of two s. Learn from inter-s protocol that subnet is reachable via multiple gatewas Use routing info from intra-s protocol to determine costs of least-cost paths to each of the gatewas Hot potato routing: Choose the gatewa that has the smallest least cost Determine from forwarding table the interface I that leads to least-cost gatewa. Enter (,I) in forwarding table 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-87 Network Laer 4-88 Intra-S Routing lso known as Interior Gatewa s (IGP) Most common Intra-S routing protocols: : Routing Information : Open Shortest Path First IGRP: Interior Gatewa Routing (Cisco proprietar) 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-89 Network Laer 4-90

16 RIP ( Routing Information ) Distance vector algorithm Included in BSD-UNIX Distribution in 98 Distance metric: # of hops (ma = 5 hops) u C B D v w destination hops u v w RIP advertisements Distance vectors: echanged among neighbors ever 0 sec via Response Message (also called advertisement) Each advertisement: list of up to 5 destination nets within S Network Laer 4-9 Network Laer 4-9 RIP: Eample w D B C Destination Network Net Router Num. of hops to dest. w B B Routing table in D Network Laer 4-9 RIP: Eample Dest Net hops w C dvertisement from to D w D B C Destination Network Net Router Num. of hops to dest. w B B Routing table in D Network Laer 4-94 RIP: Link Failure and Recover If no advertisement heard after 80 sec --> neighbor/link declared dead routes via neighbor invalidated new advertisements sent to neighbors neighbors in turn send out new advertisements (if tables changed) link failure info quickl propagates to entire net poison reverse used to prevent ping-pong loops (infinite distance = 6 hops) Network Laer 4-95 RIP Table processing RIP routing tables managed b application-level process called route-d (daemon) advertisements sent in UDP packets, periodicall repeated routed Transprt (UDP) forwarding (IP) table link forwarding table routed Transprt (UDP) (IP) link Network Laer 4-96

17 OSPF (Open Shortest Path First) 4. Introduction 4. Virtual circuit and 4.4 IP: open : publicl available Uses Link State algorithm LS packet dissemination Topolog map at each node Route computation using Dijkstra s algorithm OSPF advertisement carries one entr per neighbor dvertisements disseminated to entire S (via flooding) Carried in OSPF messages directl over IP (rather than TCP or UDP Network Laer 4-97 Network Laer 4-98 OSPF advanced features (not in RIP) Hierarchical OSPF Securit: all OSPF messages authenticated (to prevent malicious intrusion) Multiple same-cost paths allowed (onl one path in RIP) For each link, multiple cost metrics for different TOS (e.g., satellite link cost set low for best effort; high for real time) Integrated uni- and multicast support: Multicast OSPF (MOSPF) uses same topolog data base as OSPF Hierarchical OSPF in large domains. Network Laer 4-99 Network Laer 4-00 Hierarchical OSPF Two-level hierarch: local area, backbone. Link-state advertisements onl in area each nodes has detailed area topolog; onl know direction (shortest path) to nets in other areas. rea border s: summarie distances to nets in own area, advertise to other rea Border s. Backbone s: run OSPF routing limited to backbone. Boundar s: connect to other S s. 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4-0 Network Laer 4-0

18 inter-s routing: BGP BGP (Border Gatewa ): the de facto standard BGP provides each S a means to:. Obtain subnet reachabilit information from neighboring Ss.. Propagate the reachabilit information to all s internal to the S.. Determine good routes to subnets based on reachabilit information and polic. llows a subnet to advertise its eistence to rest of the : I am here Network Laer 4-0 BGP basics Pairs of s (BGP peers) echange routing info over semipermanent TCP conctns: BGP sessions Note that BGP sessions do not correspond to links. When S advertises a prefi to S, S is promising it will forward an datagrams destined to that prefi towards the prefi. S can aggregate prefies in its advertisement c a b S a S c d b c a b S ebgp session ibgp session Network Laer 4-04 Distributing reachabilit info With ebgp session between a and c, S sends prefi reachabilit info to S. c can then use ibgp do distribute this new prefi reach info to all s in S b can then re-advertise the new reach info to S over the b-to-a ebgp session When learns about a new prefi, it creates an entr for the prefi in its forwarding table. c a b S a S c d b c a b S ebgp session ibgp session Network Laer 4-05 Path attributes & BGP routes When advertising a prefi, advert includes BGP attributes. prefi + attributes = route Two important attributes: S-PTH: contains the Ss through which the advert for the prefi passed: S 67 S 7 NEXT-HOP: Indicates the specific internal-s to net-hop S. (There ma be multiple links from current S to net-hop-s.) When gatewa receives route advert, uses import polic to accept/decline. Network Laer 4-06 BGP route selection Router ma learn about more than route to some prefi. Router must select route. Elimination rules:. Local preference value attribute: polic decision. Shortest S-PTH. Closest NEXT-HOP : hot potato routing 4. dditional criteria BGP messages BGP messages echanged using TCP. BGP messages: OPEN: opens TCP connection to peer and authenticates sender UPDTE: advertises new path (or withdraws old) KEEPLIVE keeps connection alive in absence of UPDTES; also CKs OPEN request NOTIFICTION: reports errors in previous msg; also used to close connection Network Laer 4-07 Network Laer 4-08

19 BGP routing polic BGP routing polic () W B C X legend: provider customer : W B C X legend: provider customer : Y Y Figure 4.5-BGPnew: a simple BGP scenario,b,c are provider s X,W,Y are customer (of provider s) X is dual-homed: attached to two s X does not want to route from B via X to C.. so X will not advertise to B a route to C Network Laer 4-09 Figure 4.5-BGPnew: a simple BGP scenario advertises to B the path W B advertises to X the path BW Should B advertise to C the path BW? No wa! B gets no revenue for routing CBW since neither W nor C are B s customers B wants to force C to route to w via B wants to route onl to/from its customers! Network Laer 4-0 Wh different Intra- and Inter- S routing? Polic: Inter-S: admin wants control over how its traffic routed, who routes through its net. Intra-S: single admin, so no polic decisions needed Scale: hierarchical routing saves table sie, reduced update traffic Performance: Intra-S: can focus on performance Inter-S: polic ma dominate over performance 4. Introduction 4. Virtual circuit and 4.4 IP: Network Laer 4- Network Laer 4- duplicate R duplicate creation/transmission R R R duplicate c B R R4 R R4 F E D (a) (b) G Figure 4.9 Source-duplication versus in- duplication. (a) source duplication, (b) in- duplication Figure 4.40: Reverse path forwarding Network Laer 4- Network Laer 4-4

20 c B c B c B c B F E D G F E D G F 4 E D 5 G F E D G (a) Broadcast initiated at (b) Broadcast initiated at D (a) Stepwise construction of spanning tree (b) Constructed spanning tree Figure 4.4: Broadcast along a spanning tree Figure 4.4: Center-based construction of a spanning tree Network Laer 4-5 Network Laer 4-6 Multicast Routing: Problem Statement Goal: find a tree (or trees) connecting s having local mcast group members tree: not all paths between s used source-based: different tree from each sender to rcvrs shared-tree: same tree used b all group members pproaches for building mcast trees pproaches: source- based tree: one tree per source shortest path trees reverse path forwarding group- shared tree: group uses one tree minimal spanning (Steiner) center-based trees we first look at basic approaches, then specific protocols adopting these approaches Shared tree Source-based trees Shortest Path Tree Reverse Path Forwarding mcast forwarding tree: tree of shortest path routes from source to all receivers Dijkstra s algorithm S: source R R R 4 R6 R4 6 R7 5 R5 LEGEND i with attached group member with no attached group member link used for forwarding, i indicates order link added b algorithm rel on s knowledge of unicast shortest path from it to sender each has simple forwarding behavior: if (mcast datagram received on incoming link on shortest path back to center) then flood datagram onto all outgoing links else ignore datagram

21 Reverse Path Forwarding: eample S: source R R R R6 R4 R7 R5 LEGEND with attached group member with no attached group member datagram will be forwarded datagram will not be forwarded result is a source-specific reverse SPT ma be a bad choice with asmmetric links Reverse Path Forwarding: pruning forwarding tree contains subtrees with no mcast group members no need to forward datagrams down subtree prune msgs sent upstream b with no downstream group members S: source R R R R6 P R4 P R7 R5 LEGEND P with attached group member with no attached group member prune message links with multicast forwarding Shared-Tree: Steiner Tree Steiner Tree: minimum cost tree connecting all s with attached group members problem is NP- complete ecellent heuristics eists not used in practice: computational compleit information about entire needed monolithic: rerun whenever a needs to join/leave Center-based trees single deliver tree shared b all one identified as center of tree to join: edge sends unicast join-msg addressed to center join-msg processed b intermediate s and forwarded towards center join-msg either hits eisting tree branch for this center, or arrives at center path taken b join-msg becomes new branch of tree for this Center-based trees: an eample Suppose R6 chosen as center: R R R R6 R4 R7 R5 LEGEND with attached group member with no attached group member path order in which join messages generated Multicasting Routing: DVMRP DVMRP: distance vector protocol, RFC075 flood and prune: reverse path forwarding, source- based tree RPF tree based on DVMRP s own routing tables constructed b communicating DVMRP s no assumptions about underling unicast initial datagram to mcast group flooded everwhere via RPF s not wanting group: send upstream prune msgs

22 DVMRP: continued soft state: DVMRP periodicall ( min.) forgets branches are pruned: mcast data again flows down unpruned branch downstream : reprune or else continue to receive data s can quickl regraft to tree following IGMP join at leaf odds and ends commonl implemented in commercial s Mbone routing done using DVMRP Tunneling Q: How to connect islands of multicast s in a sea of unicast s? topolog logical topolog mcast datagram encapsulated inside normal (non-multicastaddressed) datagram normal IP datagram sent thru tunnel via regular IP unicast to receiving mcast receiving mcast unencapsulates to get mcast datagram PIM: Independent Multicast not dependent on an specific underling unicast routing algorithm (works with all) two different multicast distribution scenarios : Dense: group members densel packed, in close proimit. bandwidth more plentiful Sparse: # s with group members small wrt # interconnected s group members widel dispersed bandwidth not plentiful Consequences of Sparse-Dense Dichotom: Dense group membership b s assumed until s eplicitl prune data-driven construction on mcast tree (e.g., RPF) bandwidth and nongroup- processing profligate Sparse: no membership until s eplicitl join receiver- driven construction of mcast tree (e.g., center-based) bandwidth and non-group processing conservative PIM- Dense Mode PIM - Sparse Mode flood- and- prune RPF, similar to DVMRP but underling unicast protocol provides RPF info for incoming datagram less complicated (less efficient) downstream flood than DVMRP reduces reliance on underling routing algorithm has protocol mechanism for to detect it is a leaf-node center-based approach sends join msg to rendevous point (RP) intermediate s update state and forward join after joining via RP, can switch to source-specific tree increased performance: less concentration, shorter paths R R R join join all data multicast from rendevous point R6 join R4 R5 R7 rendevous point

23 PIM - Sparse Mode Network Laer: summar sender(s): unicast data to RP, which distributes down RP-rooted tree RP can etend mcast tree upstream to source RP can send stop msg if no attached receivers no one is listening! R R R join join all data multicast from rendevous point R6 join R4 R5 R7 rendevous point What we ve covered: laer services routing principles: link state and distance vector hierarchical routing IP routing protocols RIP, OSPF, BGP what s inside a? IPv6 Net stop: the Data link laer! Network Laer 4-4