Internet組播簡(jiǎn)介

1、1 清華大學(xué)計(jì)算機(jī)系 internet組播簡(jiǎn)介 2 主要內(nèi)容 u為什么需要組播？ u組播地址 u主機(jī)和路由器的交互：igmp u組播分發(fā)樹 u組播轉(zhuǎn)發(fā) u域內(nèi)組播路由協(xié)議 u域間組播路由協(xié)議 uipv6 3 主要內(nèi)容 u為什么需要組播？ u組播地址 u主機(jī)和路由器的交互：igmp u組播分發(fā)樹 u組播轉(zhuǎn)發(fā) u域內(nèi)組播路由協(xié)議 u域間組播路由協(xié)議 uipv6 4 server router unicast server router multicast 單播和組播的比較 5 example: audio streaming all clients listening to the same 8

2、kbps audio 0 0.2 0.4 0.6 0. 8 traffic mbps 120406080100 # clients multicast unicast 組播的優(yōu)勢(shì) : controls network traffic and reduces server and cpu loads : eliminates traffic redundancy : makes multipoint applications possible 6 組播帶來(lái)的問(wèn)題 ubest effort delivery: drops are to be expected. multicast applicat

3、ions should not expect reliable delivery of data and should be designed accordingly. reliable multicast is still an area for much research uno congestion avoidance: lack of tcp windowing and “slow-start” mechanisms can result in network congestion. if possible, multicast applications should attempt

4、to detect and avoid congestion conditions 組播是基于組播是基于udp的！的！ 7 組播帶來(lái)的問(wèn)題 uduplicates: some multicast protocol mechanisms (e.g. asserts, registers and spt transitions) result in the occasional generation of duplicate packets uout of order delivery: some protocol mechanisms may also result in out of orde

5、r delivery of packets 8 組播的應(yīng)用 umultimedia streaming media, iptv training, corporate communications conferencingvideo/audio unet game uany one-to-many data push applications 9 主要內(nèi)容 u為什么需要組播？ u組播地址 u主機(jī)和路由器的交互：igmp u組播分發(fā)樹 u組播轉(zhuǎn)發(fā) u域內(nèi)組播路由協(xié)議 u域間組播路由協(xié)議 uipv6 10 uipv4 multicast group addresses 224.0.0.0239.2

6、55.255.255 class “d” address space high order bits of 1st octet = “1110” ureserved link-local addresses 224.0.0.0224.0.0.255 transmitted with ttl = 1 examples: 224.0.0.1 all systems on this subnet 224.0.0.2 all routers on this subnet 224.0.0.4 dvmrp routers 224.0.0.5 ospf routers 224.0.0.13 pimv2 ro

7、uters 組播地址 11 uadministratively scoped addresses 239.0.0.0239.255.255.255 private address space similar to rfc1918 unicast addresses not used for global internet traffic used to limit “scope” of multicast traffic same addresses may be in use at different locations for different multicast sessions ex

8、amples site-local scope: 239.253.0.0/16 organization-local scope: 239.192.0.0/14 組播地址 12 32 bits 28 bits 25 bits23 bits 48 bits 01-00-5e-7f-00-01 1110 5 bits lost 組播地址 ip multicast mac address mapping (fddi and ethernet) 239.255.0.1 13 224.1.1.1 224.129.1.1 225.1.1.1 225.129.1.1 . . . 238.1.1.1 238.

9、129.1.1 239.1.1.1 239.129.1.1 0 x0100.5e01.0101 1 - multicast mac address (fddi and ethernet) 32 - ip multicast addresses 組播地址 be aware of the 32:1 address overlap ip multicast mac address mapping (fddi then the rpf check succeeds otherwise, the rpf check fails reverse path forwarding (rpf) 34 組播轉(zhuǎn)發(fā)

10、source 151.10.3.21 example: rpf checking mcast packets 35 組播轉(zhuǎn)發(fā) rpf check fails! unicast route table network interface 151.10.0.0/16s1 198.14.32.0/24s0 204.1.16.0/24e0 a closer look: rpf check fails packet arrived on wrong interface! e0 s1 s0 s2 multicast packet from source 151.10.3.21 x discard pack

11、et! 36 組播轉(zhuǎn)發(fā) a closer look: rpf check succeeds! e0 s1 s0 s2 multicast packet from source 151.10.3.21 packet arrived on correct interface! forward out all outgoing interfaces. (i. e. down the distribution tree) 37 主要內(nèi)容 u為什么需要組播？ u組播地址 u主機(jī)和路由器的交互：igmp u組播分發(fā)樹 u組播轉(zhuǎn)發(fā) u域內(nèi)組播路由協(xié)議 u域間組播路由協(xié)議 uipv6 38 multicast

12、 routing is not unicast routing u you have to think of it differently u it is not like ospf u it is not like rip u it is not like anything you may be familiar with 組播路由和單播路由 39 組播路由協(xié)議的類型 udense-mode uses “push” model traffic flooded throughout network pruned back where it is unwanted flood it sends

13、an (s, g) prune message. source “s” receiver 1 (group “g”) e x y ab cd mrouted mrouted mroutedmrouted mroutedmrouted router d prunes interface. mrouted truncated broadcast tree based on dvmrp route metrics (s, g) multicast packet flow 51 dvmrpflood & prune final pruned state source “s” receiver 1 (g

14、roup “g”) e x y ab cd mrouted mrouted mroutedmrouted mroutedmrouted mrouted truncated broadcast tree based on dvmrp route metrics (s, g) multicast packet flow 52 receiver 2 joins group “g” receiver 2 (group “g”) router y sends a “graft (s, g)” message dvmrpgrafting source “s” receiver 1 (group “g”)

15、e x y ab cd mrouted mrouted mroutedmrouted mroutedmrouted mrouted truncated broadcast tree based on dvmrp route metrics (s, g) multicast packet flow 53 dvmrpgrafting router e responds with a “graft-ack” sends its own “graft (s, g) message receiver 2 (group “g”) source “s” receiver 1 (group “g”) e x

16、y ab cd mrouted mrouted mroutedmrouted mroutedmrouted mrouted truncated broadcast tree based on dvmrp route metrics (s, g) multicast packet flow 54 receiver 2 (group “g”) source “s” receiver 1 (group “g”) e x y ab cd mrouted mrouted mrouted mroutedmrouted dvmrpgrafting router d responds with a “graf

17、t-ack” graft-ack begins forwarding (s, g) packets mroutedmrouted truncated broadcast tree based on dvmrp route metrics (s, g) multicast packet flow 55 dvmrpevaluation uwidely used on the mbone (being phased out) usignificant scaling problems slow convergencerip-like behavior significant amount of mu

18、lticast routing state information stored in routers(s,g) everywhere no support for shared trees maximum number of hops 32 unot appropriate for large scale production networks due to flood and prune behavior due to its poor scalability 56 mospf (rfc 1584) uextension to ospf unicast routing protocol o

19、spf: routers use link state advertisements to understand all available links in the network (route messages along least-cost paths) mospf: includes multicast information in ospf link state advertisements to construct multicast distribution trees (each router maintains an up-to-date image of the topo

20、logy of the entire network) 57 mospf (rfc 1584) ugroup membership lsas are flooded throughout the ospf routing domain so mospf routers can compute outgoing interface lists uuses dijkstra algorithm to compute shortest-path tree separate calculation is required for each (snet, g) pair 58 membership ls

21、as membership lsas area 1area 2 mabr1mabr2 area 0 mospf membership lsas mb mbmamama 59 area 1area 2 (s1 , b)(s2 , a) mospf intra-area traffic mama mb mbma not receiving (s2 , a) traffic mabr1mabr2 area 0 60 mospf inter-area traffic area 1area 2 mama mb mbma wildcard receiver flag (*, *) wildcard rec

22、eiver flag (*, *) (s1 , b)(s2 , a) wildcard receivers “pull” traffic from all sources in the area. mabr1mabr2 area 0 61 mospf inter-area traffic area 1area 2 mama mb mbma(s1 , b)(s2 , a) mabr1mabr2 area 0 62 (s1 , b)(s2 , a) (ga , gb)(ga ) area 1area 2 mabr1mabr2 mospf inter-area traffic mamama mb m

23、b summarized membership lsa summarized membership lsa mabr routers inject summary membership lsas into area 0. area 0 membership lsas membership lsas 63 area 1area 2 mabr1 (s1 , b)(s2 , a) mabr2 mospf inter-area traffic mama mb mbma area 0 64 area 1area 2 mabr1 (s1 , b)(s2 , a) mabr2 mospf inter-are

24、a traffic wildcard receiver flag (*, *) wildcard receiver flag (*, *) area 0 65 (ga , gb)(ga ) area 1area 2 mabr1mabr2 mospf inter-domain traffic mamama mb mb summarized membership lsa summarized membership lsa external as masbr area 0 membership lsas membership lsas 66 mospf inter-domain traffic (s

25、2 , b) external as area 1area 2 ma mabr1 ma mb mbma mabr2 masbr (s1 , a) area 0 67 external as area 1area 2 mabr1 (s1 , b)(s2 , a) mabr2 masbr mospf inter-domain traffic wildcard receiver flag (*, *) wildcard receiver flag (*, *) unnecessary traffic may flow all the way to the masbr router! area 0 6

26、8 mospfevaluation uappropriate for use within single routing domain uflood multicast traffic everywhere to create state, uses lsas and the link-state database uprotocol dependent works only in ospf-based networks 69 mospfevaluation usignificant scaling problems dijkstra algorithm run for every multi

27、cast (snet, g) pair! dijkstra algorithm rerun when: group membership changes line-flaps does not support shared-trees unot appropriate for general purpose multicast networks where the number of senders may be quite large ip/tv(every ip/tv client is a multicast source) 70 pim-dm uprotocol independent

28、 supports all underlying unicast routing protocols including: static, rip, igrp, eigrp, is-is, bgp, and ospf uuses reverse path forwarding floods network and prunes back based on multicast group membership assert mechanism used to prune off redundant flows uappropriate for. densely distributed recei

29、vers located in close proximity to source few senders -to- many receivers (due to frequent flooding) 71 pim-dm flood & prune source initial flooding receiver multicast packets (s, g) state created in router in the network! 72 pim-dm flood & prune source pruning unwanted traffic receiver multicast pa

30、ckets messages 73 pim-dm flood & prune results after pruning source receiver multicast packets flood & prune process repeats every 3 minutes! (s, g) state still exists in every router in the network! 74 pim-dm assert mechanism e0 incoming multicast packet (successful rpf check) e0 s0 routers receive

31、 packet on an interface in their “oilist”! only one router should continue sending to avoid duplicate packets s0 routers send “pim assert” messages assert assert compare distance and metric values router with best route to source wins if metric & distance equal, highest ip adr wins losing router sto

32、ps sending (prunes interface) 75 pim-dm evaluation umost effective for large number of densely distributed receivers located in close proximity to source uadvantages: easy to configuretwo commands ip pim dense-mode simple flood and prune mechanism upotential issues. inefficient flood and prune behav

33、ior complex assert mechanism mixed control and data planes results in (s, g) state in every router in the network no support for shared trees 76 pim-sm (rfc 2362) usupports both source and shared trees assumes no hosts want multicast traffic unless they specifically ask for it uuses a rendezvous poi

34、nt (rp) senders and receivers “rendezvous” at this point to learn of each others existence senders are “registered” with rp by their first-hop router receivers are “joined” to the shared tree (rooted at the rp) by their local designated router (dr) uappropriate for wide scale deployment for both den

35、sely and sparsely populated groups in the enterprise optimal choice for all production networks regardless of size and membership density 77 pim-sm shared tree join receiver rp (*, g) join shared tree (*, g) state created only along the shared tree. 78 pim-sm sender registration receiver rp (s, g) j

36、oin source shared tree (s, g) register (unicast) source tree (s, g) state created only along the source tree.traffic flow 79 pim-sm sender registration receiver rp source shared tree source tree rp sends a register-stop back to the first-hop router to stop the register process. (s, g) register-stop

37、(unicast) traffic flow (s, g) register (unicast) (s, g) traffic begins arriving at the rp via the source tree. 80 pim-sm sender registration receiver rp source shared tree source tree traffic flow source traffic flows natively along spt to rp. from rp, traffic flows down the shared tree to receivers

38、. 81 pim-sm spt switchover receiver rp (s, g) join source source tree shared tree last-hop router joins the source tree. additional (s, g) state is created along new part of the source tree. traffic flow 82 pim-sm spt switchover receiver rp source source tree shared tree (s, g)rp-bit prune traffic b

39、egins flowing down the new branch of the source tree. additional (s, g) state is created along the shared tree to prune off (s, g) traffic. traffic flow 83 pim-sm spt switchover receiver rp source (s, g) traffic flow is now pruned off of the shared tree and is flowing to the receiver via the source

40、tree. source tree shared tree traffic flow 84 pim-sm spt switchover receiver rp source source tree shared tree (s, g) traffic flow is no longer needed by the rp so it prunes the flow of (s, g) traffic. traffic flow (s, g) prune 85 pim-sm spt switchover receiver rp source (s, g) traffic flow is now o

41、nly flowing to the receiver via a single branch of the source tree. source tree shared tree traffic flow 86 pim-sm fff uthe default behavior of pim- sm in cisco ios is that routers with directly connected members will join the shortest path tree as soon as they detect a new multicast source pim-sm f

42、requently forgotten fact 87 pim-sm evaluation ueffective for sparse or dense distribution of multicast receivers uadvantages: traffic only sent down “joined” branches can switch to optimal source-trees for high traffic sources dynamically unicast routing protocol-independent basis for inter-domain m

43、ulticast routing when used with mbgp and msdp 88 主要內(nèi)容 u為什么需要組播？ u組播地址 u主機(jī)和路由器的交互：igmp u組播分發(fā)樹 u組播轉(zhuǎn)發(fā) u域內(nèi)組播路由協(xié)議 u域間組播路由協(xié)議 uipv6 89 域間組播路由協(xié)議 ubgmp (border gateway multicast protocol) (未來(lái)) (internet-draft) umsdp (multicast source discover protocol) (rfc 3618) umbgp (multi-protocol bgp) (rfc 2283) ussm (s

44、ource specific multicast) (rfc 3569) 90 域間組播路由協(xié)議bgmp ubgmp(邊界網(wǎng)關(guān)組播協(xié)議) u基本思想 對(duì)在網(wǎng)絡(luò)中活動(dòng)的任何組播組,存在一個(gè)單單 獨(dú)的雙向共享樹獨(dú)的雙向共享樹,該共享樹可以跨越包含該 組發(fā)送者或接收者的所有域 u組播的根域應(yīng)該是這樣的一個(gè)域 它擁有一個(gè)特定的特定的組播地址范圍 91 域間組播路由協(xié)議bgmp u任何組播組都存在跨越域的雙向共享樹 u顯式加入模型 u發(fā)向根域的加入和剪枝 u每個(gè)組單根域 u需要bgp4+ 必須攜帶nlri區(qū)域中的組前綴 需要建立雙向樹 u要求嚴(yán)格的層次化的地址分配 masc被推薦為分配方式 92 bgm

45、p a host in c joins to group g domaina domaine domainc domaind domainb domainf root domain c1 a2 e1 a1a4 a3 d1 b1b2 f1 f2 join join join 93 bgmp tree constructed, data goes to c domaina domaine domainc domaind domainb domainf root domain c1 a2 e1 a1a4 a3 d1 b1b2 f1 f2 94 bgmp domain e joins to g dom

46、aina domaine domainc domaind domainb domainf join c1 a2 e1 a1a4 a3 d1 b1b2 f1 f2 join 95 bgmp tree constructed. data goes to e domaina domaine domainc domaind domainb domainf c1 a2 e1 a1a4 a3 d1 b1b2 f1 f2 96 理想與現(xiàn)實(shí) ubgmp和masc非常遙遠(yuǎn) 兩者都非常難以實(shí)施 仍處于ietf草案提案階段 uisp目前要部署組播,解決方案如何? 97 interdomain multicastca

47、mpus multicast multicast components end-to-end architectureend-to-end architecture uend stations (hosts-to-routers): igmp uswitches (layer 2 optimization): cgmp, igmp snooping or rgmp urouters (multicast forwarding protocol): pim sm or bidirectional pim umulticast routing across domains mbgp umultic

48、ast source discovery msdp with pim-sm usource specific multicast pim-ssm isp b multicast source y isp a multicast source x isp b dr rp rp dr dr igmp pim-sm bidir pim pim-ssm mvpn igmp snooping, cgmp, rgmp mbgp msdp isp a 98 部署組播的要求 u需要一個(gè)域內(nèi)的顯式加入的路由協(xié)議 提高效率 pim-sm u域間路由使用現(xiàn)有的單播模型 mbgp u需要域間組播源發(fā)現(xiàn) 不同的pim域

49、rp需要共享組播源信息 msdp 99 mbgp overview umbgp: multiprotocol bgp, not multicast bgp defined in rfc 2283 (extensions to bgp) can carry different types of routes ipv4 unicast ipv6 unicast ipv4 multicastipv6 multicast may be carried in same bgp session does not propagate multicast state info still need pim t

50、o build distribution trees same path selection and validation rules as-path, localpref, med, 10 0 mbgp overview useparate bgp tables maintained unicast bgp table (u-table) multicast bgp table (m-table) bgp nlri specifies which bgp table allows different unicast/multicast topologies or policies uunic

51、ast bgp table (u-table) contains unicast prefixes for unicast forwarding populated with bgp unicast nlri umulticast bgp table (m-table) contains unicast prefixes for rpf checking populated with bgp multicast nlri 10 1 mbgp update message uaddress family information (afi) identifies address type (see

52、 rfc1700) afi = 1 (ipv4) afi = 2 (ipv6) usub-address family information (sub-afi) sub category for afi field address family information (afi) = 1 (ipv4) sub-afi = 1 (nlri is used for unicast) sub-afi = 2 (nlri is used for multicast rpf check) sub-afi = 3 (both unicast and multicast) 10 2 route/mask,

53、 dist. (first match) (default dist. = 1) route/mask, dist. (longest match) pim rpf calculation details iif, rpf neighbor decreasing preference route/mask, dist. (best path) (ibgp def. dist.=200) (ebgp def. dist.=20) 10 3 as 321 as 123 sender 192.168.100.0/24 receiver router bgp 123 neighbor 192.168.

54、100.2 remote-as 321 . . . mbgp capability negotiation .1.2 router bgp 321 neighbor 192.168.100.1 remote-as 123 . . . 10 4 as 321 as 123 mbgp session for unicast and multicast nlri sender 192.168.100.0/24 192.192.25.0/24 receiver bgp: 192.168.100.2 open active, local address 192.168.100.1 bgp: 192.16

55、8.100.2 went from active to opensent bgp: 192.168.100.2 sending open, version 4 bgp: 192.168.100.2 open rcvd, version 4 bgp: 192.168.100.2 rcv open w/option parameter type: 2, len: 6 bgp: 192.168.100.2 open has capability code: 1, length 4 bgp: 192.168.100.2 open has mp_ext cap for afi/safi: 1/1 bgp

56、: 192.168.100.2 rcv open w/option parameter type: 2, len: 6 bgp: 192.168.100.2 open has capability code: 1, length 4 bgp: 192.168.100.2 open has mp_ext cap for afi/safi: 1/2 bgp: 192.168.100.2 went from opensent to openconfirm bgp: 192.168.100.2 went from openconfirm to established mbgp capability n

57、egotiation .1.2 10 5 mbgpnlri information unicast bgp table multicast bgp table mp_reach_nlri: 192.192.2/24 afi: 1, sub-afi: 1 (unicast) as_path: 300 200 med: next-hop: 192.168.200.2 bgp update from peer *i192.192.2.0/24 192.168.200.2 300 200 i network next-hop path *i160.10.1.0/24 192.20.2.2 i *i16

58、0.10.3.0/24 192.20.2.2 i network next-hop path *i160.10.1.0/24 192.20.2.2 i *i160.10.3.0/24 192.20.2.2 i storage of arriving nlri information depends on afi/safi fields in the update message 10 6 mbgpnlri information mp_reach_nlri: 192.192.2/24 afi: 1, sub-afi: 2 (multicast) as_path: 300 200 med: ne

59、xt-hop: 192.168.200.2 bgp update from peer *i192.192.2.0/24 192.168.200.2 300 200 i network next-hop path *i160.10.1.0/24 192.20.2.2 i *i160.10.3.0/24 192.20.2.2 i network next-hop path *i160.10.1.0/24 192.20.2.2 i *i160.10.3.0/24 192.20.2.2 i storage of arriving nlri information depends on afi/safi

60、 fields in the update message unicast bgp table only (afi=1/safi=1 or old style nlri) unicast bgp table multicast bgp table 10 7 mbgpnlri information mp_reach_nlri: 192.192.2/24 afi: 1, sub-afi: 3 (both) as_path: 300 200 med: next-hop: 192.168.200.2 bgp update from peer network next-hop path *i160.1