rfc1245.ps

著名的RFC文档,其中有一些文档是已经翻译成中文的的.

(oadcast multi-access network support.) 122.63 676 T
0 F
( OSPF treats these networks \050e.g., X.25 ) 318.51 676 T
-0.01 (PDNs\051 pretty much as if they were LANs \050i.e., a DR is elected, and a network LSA is gener-) 85.54 662 P
-0.29 (ated\051. Additional con\336guration information is needed however for routers attached to these net-) 85.54 648 P
(work to initially \336nd each other) 85.54 634 T
(.) 236.45 634 T
(\245) 72 614 T
4 F
(OSPF ar) 85.54 614 T
(eas) 130.29 614 T
0 F
(. OSPF allows the Autonomous Systems to be broken up into regions call areas. ) 146.28 614 T
(This is useful for several reasons. First, it provides an extra level of ) 85.54 600 T
4 F
(r) 411.64 600 T
(outing pr) 416.75 600 T
(otection) 464.18 600 T
0 F
(: rout-) 504.81 600 T
-0.29 (ing within an area is protected from all information external to the area. Second, by splitting an ) 85.54 586 P
-0.3 (Autonomous System into areas the ) 85.54 572 P
4 F
-0.3 (cost of the Dijkstra pr) 254.27 572 P
-0.3 (ocedur) 365.44 572 P
-0.3 (e ) 400.53 572 P
0 F
-0.3 (\050in terms of CPU cycles\051 is ) 408.55 572 P
(reduced.) 85.54 558 T
(\245) 72 538 T
4 F
(Flexible import of external r) 85.54 538 T
(outing information.) 230.55 538 T
0 F
( In OSPF) 330.5 538 T
(, ) 374.19 538 T
4 F
(each external r) 380.19 538 T
(oute) 456.58 538 T
0 F
( is imported ) 478.56 538 T
(into the Autonomous System in ) 85.54 524 T
4 F
(a separate LSA) 240.47 524 T
0 F
(. This reduces the amount of \337ooding traf) 319.08 524 T
(\336c ) 518.07 524 T
(\050since external routes change often, and you want to only \337ood the changes\051. It also enables ) 85.54 510 T
4 F
-0.43 (partial r) 85.54 496 P
-0.43 (outing table updates) 127.86 496 P
0 F
-0.43 ( when only a single external route changes. OSPF external LSAs ) 230.96 496 P
(also provide the following features. A ) 85.54 482 T
4 F
(forwarding addr) 270.4 482 T
(ess) 355.81 482 T
0 F
( can be included in the external ) 370.46 482 T
(LSA, eliminating extra-hops at the edge of the Autonomous System. There are two levels of ) 85.54 468 T
(external metrics that can be speci\336ed, ) 85.54 454 T
4 F
(type 1) 269.06 454 T
0 F
( and ) 300.04 454 T
4 F
(type 2) 323.35 454 T
0 F
(. Also, external routes can be tagged ) 354.33 454 T
(with a 32-bit number \050the ) 85.54 440 T
4 F
(external r) 211.12 440 T
(oute tag) 261.19 440 T
0 F
(; commonly used as an AS number of the route\325) 302.16 440 T
(s ) 531.68 440 T
(origin\051, simplifying external route management in a transit Autonomous System.) 85.54 426 T
(\245) 72 406 T
4 F
(Four level r) 85.54 406 T
(outing hierar) 145.27 406 T
(chy) 212.69 406 T
0 F
(. OSPF has a four level routing hierarchy) 229.9 406 T
(, or trust model: ) 426.32 406 T
4 F
(intra-) 505.94 406 T
(ar) 85.54 392 T
(ea) 96.64 392 T
0 F
(, ) 107.96 392 T
4 F
(inter) 113.96 392 T
(-ar) 138.16 392 T
(ea) 153.26 392 T
0 F
(, ) 164.59 392 T
4 F
(external type 1) 170.58 392 T
0 F
( and ) 246.52 392 T
4 F
(external type 2) 269.84 392 T
0 F
( routes. This enables multiple levels of ) 345.78 392 T
(routing protection, and simpli\336es routing management in an Autonomous System.) 85.54 378 T
(\245) 72 358 T
4 F
(V) 85.54 358 T
(irtual links) 93.75 358 T
0 F
(. By allowing the con\336guration of virtual links, OSPF ) 150.07 358 T
4 F
(r) 410.94 358 T
(emoves topological ) 416.05 358 T
(r) 85.54 344 T
(estrictions) 90.64 344 T
0 F
( on area layout in an Autonomous System.) 143.27 344 T
(\245) 72 324 T
4 F
-0.32 (Authentication of r) 85.54 324 P
-0.32 (outing pr) 182.62 324 P
-0.32 (otocol exchanges) 229.74 324 P
0 F
-0.32 (. Every time an OSPF router receives a routing ) 315.03 324 P
(protocol packet, it authenticates the packet before processing it further) 85.54 310 T
(.) 422.61 310 T
(\245) 72 290 T
4 F
-0.03 (Flexible r) 85.54 290 P
-0.03 (outing metric.) 134.26 290 P
0 F
-0.03 ( In OSPF) 206.18 290 P
-0.03 (, metric are assigned to outbound router interfaces. The cost ) 249.82 290 P
(of a path is then the sum of the path\325) 85.54 276 T
(s component interfaces. The routing metric itself can be ) 260.42 276 T
(assigned by the system administrator to indicate any combination of network characteristics ) 85.54 262 T
(\050e.g., delay) 85.54 248 T
(, bandwidth, dollar cost, etc.\051.) 138.04 248 T
(\245) 72 228 T
4 F
-0.09 (Equal-cost multipath.) 85.54 228 P
0 F
-0.09 ( When multiple best cost routes to a destination exist, OSPF \336nds them ) 196.73 228 P
(and they can be then used to load share traf) 85.54 214 T
(\336c to the destination.) 292.82 214 T
(\245) 72 194 T
4 F
(T) 85.54 194 T
(OS-based r) 93.32 194 T
(outing.) 150.74 194 T
0 F
( Separate sets of routes can be calculated for each IP type of service. For ) 186.4 194 T
(example, low delay traf) 85.54 180 T
(\336c could be routed on one path, while high bandwidth traf) 198.56 180 T
(\336c is routed ) 477.16 180 T
-0.39 (on another) 85.54 166 P
-0.39 (. This is done by \050optionally\051 assigning, to each outgoing router interface, one metric ) 135.44 166 P
(for each IP T) 85.54 152 T
(OS.) 148.26 152 T
(\245) 72 132 T
4 F
(V) 85.54 132 T
(ariable-length subnet support.) 93.09 132 T
0 F
( OSPF includes support for variable-length subnet masks by ) 248.02 132 T
(carrying a network mask with each advertised destination.) 85.54 118 T
FMENDPAGE
%%EndPage: "5" 6
%%Page: "6" 6
612 792 0 FMBEGINPAGE
72 702 540 720 R
7 X
0 K
V
0 F
0 X
(RFC 1245) 72 712 T
(OSPF protocol analysis) 249.36 712 T
(July 1991) 493.02 712 T
72 69.05 540 81 R
7 X
V
0 X
([Moy]) 72 73 T
([Page 6]) 499.7 73 T
72 108 540 684 R
7 X
V
0 X
(\245) 72 676 T
4 F
-0.08 (Stub ar) 85.54 676 P
-0.08 (ea support. ) 123.56 676 P
0 F
-0.08 (T) 183.69 676 P
-0.08 (o support routers having insuf) 190.18 676 P
-0.08 (\336cient memory) 333.53 676 P
-0.08 (, areas can be con\336gured as ) 405.63 676 P
(stubs. External LSAs \050often making up the bulk of the Autonomous System\051 are not \337ooded ) 85.54 662 T
(into/throughout stub areas. Routing to external destinations in stub areas is based solely on ) 85.54 648 T
(default.) 85.54 634 T
FMENDPAGE
%%EndPage: "6" 7
%%Page: "7" 7
612 792 0 FMBEGINPAGE
72 702 540 720 R
7 X
0 K
V
0 F
0 X
(RFC 1245) 72 712 T
(OSPF protocol analysis) 249.36 712 T
(July 1991) 493.02 712 T
72 69.05 540 81 R
7 X
V
0 X
([Moy]) 72 73 T
([Page 7]) 499.7 73 T
72 108 540 684 R
7 X
V
2 F
0 X
(3.0  Cost of the pr) 72 673.33 T
(otocol) 193.4 673.33 T
0 F
-0.1 (This section attempts to analyze how the OSPF protocol will perform and scale in the Internet. In ) 72 646 P
(this analysis, we will concentrate on the following four areas:) 72 632 T
(\245) 72 612 T
4 F
(Link bandwidth) 85.54 612 T
0 F
(. In OSPF) 168.53 612 T
(, a reliable \337ooding mechanism is used to ensure that router link ) 215.22 612 T
(state databases are remained synchronized. Individual components of the link state databases ) 85.54 598 T
-0.17 (\050the LSAs\051 are refreshed infrequently \050every 30 minutes\051, at least in the absence of topological ) 85.54 584 P
(changes. Still, as the size of the database increases, the amount of link bandwidth used by the ) 85.54 570 T
(\337ooding procedure also increases.) 85.54 556 T
(\245) 72 536 T
4 F
-0.03 (Router memory) 85.54 536 P
0 F
-0.03 (. The size of an OSPF link state database can get quite lar) 166.32 536 P
-0.03 (ge, especially in the ) 441.86 536 P
(presence of many external LSAs. This imposes requirements on the amount of router memory ) 85.54 522 T
(available.) 85.54 508 T
(\245) 72 488 T
4 F
(CPU usage) 85.54 488 T
0 F
(. In OSPF) 141.83 488 T
(, this is dominated by the length of time it takes to run the shortest path ) 188.52 488 T
(calculation \050Dijkstra procedure\051. This is a function of the number of routers in the OSPF sys-) 85.54 474 T
(tem.) 85.54 460 T
(\245) 72 440 T
4 F
(Role of the Designated Router) 85.54 440 T
(.) 238.32 440 T
0 F
( The Designated router receives and sends more packets on a ) 241.32 440 T
-0.46 (multi-access networks than the other routers connected to the network. Also, there is some time ) 85.54 426 P
(involved in cutting over to a new Designated Router after the old one fails \050especially when ) 85.54 412 T
(both the Backup Designated Router and the Designated Router fail at the same time\051. For this ) 85.54 398 T
-0.27 (reason, it is possible that you may want to limit the number of routers connected to a single net-) 85.54 384 P
(work.) 85.54 370 T
(The remaining section will analyze these areas, estimating how much resources the OSPF proto-) 72 344 T
-0.05 (col will consume, both now and in the future. T) 72 330 P
-0.05 (o aid in this analysis, the next section will present ) 298.93 330 P
(some data that have been collected in actual OSPF \336eld deployments.) 72 316 T
3 F
(3.1   Operational data) 72 282.67 T
0 F
-0.44 (The OSPF protocol has been deployed in a number of places in the Internet. For a summary of this ) 72 256 P
(deployment, see [1]. Some statistics have been gathered from this operational experience, via ) 72 242 T
-0.03 (local network management facilities. Some of these statistics are presented in the following table:) 72 228 P
FMENDPAGE
%%EndPage: "7" 8
5 10 /Times-Bold FMDEFINEFONT
%%Page: "8" 8
612 792 0 FMBEGINPAGE
72 702 540 720 R
7 X
0 K
V
0 F
0 X
(RFC 1245) 72 712 T
(OSPF protocol analysis) 249.36 712 T
(July 1991) 493.02 712 T
72 69.05 540 81 R
7 X
V
0 X
([Moy]) 72 73 T
([Page 8]) 499.7 73 T
72 108 540 684 R
7 X
V
72 666.01 540 674 C
72 671.98 540 671.98 2 L
0.5 H
0 Z
0 X
0 K
N
0 0 612 792 C
5 F
0 X
0 K
(T) 72 677.33 T
(ABLE 1. Pertinent operational statistics) 77.93 677.33 T
(Statistic) 72 655.34 T
(BARRNet) 216 655.34 T
(NSI) 324 655.34 T
(OARnet) 432 655.34 T
1 F
(Data gathering \050duration\051) 72 638.34 T
(99 hours) 216 638.34 T
(277 hours) 324 638.34 T
(28 hours) 432 638.34 T
(Dijkstra frequency) 72 622.34 T
(50 minutes) 216 622.34 T
(25 minutes) 324 622.34 T
(13 minutes) 432 622.34 T
(External incremental frequency) 72 606.34 T
(1.2 minutes) 216 606.34 T
(.98 minutes) 324 606.34 T
(not gathered) 432 606.34 T
(Database turnover) 72 590.34 T
(29.7 minutes) 216 590.34 T
(30.9 minutes) 324 590.34 T
(28.2 minutes) 432 590.34 T
(LSAs per packet) 72 574.34 T
(3.38) 216 574.34 T
(3.16) 324 574.34 T
(2.99) 432 574.34 T
(Flooding retransmits) 72 558.34 T
(1.3%) 216 558.34 T
(1.4%) 324 558.34 T
(.7%) 432 558.34 T
0 F
(The \336rst line in the above table show the length of time that statistics were gathered on the three ) 72 533.01 T
(networks. A brief description of the other statistics follows:) 72 519.01 T
(\245) 72 499.01 T
4 F
(Dijkstra fr) 85.54 499.01 T
(equency) 140.27 499.01 T
(. ) 181.59 499.01 T
0 F
(In OSPF) 187.59 499.01 T
(, the Dijkstra calculation involves only those routers and transit ) 228.28 499.01 T
-0.14 (networks belonging to the AS. The Dijkstra is run only when something in the system changes ) 85.54 485.01 P
(\050like a serial line between two routers goes down\051. Note that in these operational systems, the ) 85.54 471.01 T
(Dijkstra process runs only infrequently \050the most frequent being every 13 minutes\051.) 85.54 457.01 T
(\245) 72 437.01 T
4 F
(External incr) 85.54 437.01 T
(emental fr) 153.61 437.01 T
(equency) 206.35 437.01 T
0 F
(. In OSPF) 247.54 437.01 T
(, when an external route changes only its entry in ) 294.23 437.01 T
-0.13 (the routing table is recalculated. These are called external incremental updates. Note that these ) 85.54 423.01 P
(happen much more frequently than the Dijkstra procedure. \050in other words, incremental ) 85.54 409.01 T
(updates are saving quite a bit of processor time\051.) 85.54 395.01 T
(\245) 72 375.01 T
4 F
-0.45 (Database turnover) 85.54 375.01 P
-0.45 (.) 179.58 375.01 P
0 F
-0.45 ( In OSPF) 182.58 375.01 P
-0.45 (, link state advertisements are refreshed at a minimum of every 30 ) 225.36 375.01 P
(minutes. New advertisement instances are sent out more frequently when some part of the ) 85.54 361.01 T
-0.2 (topology changes. The table shows that, even taking topological changes into account, on aver-) 85.54 347.01 P
(age an advertisement is updated close to only every 30 minutes. This statistic will be used in ) 85.54 333.01 T
(the link bandwidth calculations below) 85.54 319.01 T
(. Note that NSI actually shows advertisements updated ) 267.31 319.01 T
(every 30.7 \050> 30\051 minutes. This probably means that at one time earlier in the measurement ) 85.54 305.01 T
(period, NSI had a smaller link state database that it did at the end.) 85.54 291.01 T
(\245) 72 271.01 T
4 F
-0.39 (LSAs per packet.) 85.54 271.01 P
0 F
-0.39 ( In OSPF) 173.04 271.01 P
-0.39 (, multiple LSAs can be included in either Link State Update or Link ) 215.95 271.01 P
-0.35 (State Acknowledgment packets.The table shows that, on average, around 3 LSAs are carried in ) 85.54 257.01 P
(a single packet. This statistic is used when calculating the header overhead in the link band-) 85.54 243.01 T
(width calculation below) 85.54 229.01 T
(. This statistic was derived by diving the number of LSAs \337ooded by ) 200.01 229.01 T
(the number of \050non-hello\051 multicasts sent.) 85.54 215.01 T
(\245) 72 195.01 T
4 F
(Flooding r) 85.54 195.01 T
(etransmits.) 138.97 195.01 T
0 F
( This counts both retransmission of LS Update packets and Link State ) 195.92 195.01 T
(Acknowledgment packets, as a percentage of the original multicast \337ooded packets. The table ) 85.54 181.01 T
(shows that \337ooding is working well, and that retransmits can be ignored in the link bandwidth ) 85.54 167.01 T
(calculation below) 85.54 153.01 T
(.) 169.69 153.01 T
FMENDPAGE
%%EndPage: "8" 9
%%Page: "9" 9
612 792 0 FMBEGINPAGE
72 702 540 720 R
7 X
0 K
V
0 F
0 X
(RFC 1245) 72 712 T
(OSPF protocol analysis) 249.36 712 T
(July 1991) 493.02 712 T
72 69.05 540 81 R