4 %%DocumentFonts: (atend)
7 % FrameMaker PostScript Prolog 2.0, for use with FrameMaker 2.0
8 % Copyright (c) 1986,87,89 by Frame Technology, Inc. All rights reserved.
11 % Due to bugs in Transcript, the 'PS-Adobe-' is omitted from line 1
13 % Set up Color vs. Black-and-White
14 /FMPrintInColor systemdict /colorimage known def
15 % Uncomment this line to force b&w on color printer
16 % /FMPrintInColor false def
17 /FrameDict 190 dict def
18 systemdict /errordict known not {/errordict 10 dict def
19 errordict /rangecheck {stop} put} if
20 % The readline in 23.0 doesn't recognize cr's as nl's on AppleTalk
21 FrameDict /tmprangecheck errordict /rangecheck get put
22 errordict /rangecheck {FrameDict /bug true put} put
23 FrameDict /bug false put
25 % Some PS machines read past the CR, so keep the following 3 lines together!
26 currentfile 5 string readline
30 errordict /rangecheck FrameDict /tmprangecheck get put
40 gstring exch gindex exch put
41 /gindex gindex 1 add def
44 gstring 0 gindex getinterval true
49 /Times-Roman findfont 18 scalefont setfont
51 (FrameMaker version does not match postscript_prolog!)
74 array /FMfonts exch def
77 0 ne dup {setmanualfeed} if
82 manualfeed {true} {papersize} ifelse
83 {manualpapersize} {false} ifelse
84 {desperatepapersize} if
87 currenttransfer cvlit /orgxfer exch def
88 currentscreen cvlit /orgproc exch def
89 /organgle exch def /orgfreq exch def
99 /landscape exch 0 ne def
101 90 rotate 0 exch neg translate pop
106 /orgmatrix matrix def
126 /FMNORMALIZEGRAPHICS {
147 [/fy /fx /fh /fw /ury /urx /lly /llx] {exch def} forall
150 fw urx llx sub div fh ury lly sub div scale
151 llx neg lly neg translate
159 %%BeginFeature *ManualFeed True
160 statusdict /manualfeed true put
163 /max {2 copy lt {exch} if pop} bind def
164 /min {2 copy gt {exch} if pop} bind def
167 paperheight sub abs 16 lt exch
168 paperwidth sub abs 16 lt and
169 {/papername exch def} {pop} ifelse
171 /papersizedict FMLOCAL
173 /papersizedict 14 dict def
175 /papername /unknown def
176 /Letter 8.5 inch 11.0 inch pagedimen
177 /LetterSmall 7.68 inch 10.16 inch pagedimen
178 /Tabloid 11.0 inch 17.0 inch pagedimen
179 /Ledger 17.0 inch 11.0 inch pagedimen
180 /Legal 8.5 inch 14.0 inch pagedimen
181 /Statement 5.5 inch 8.5 inch pagedimen
182 /Executive 7.5 inch 10.0 inch pagedimen
183 /A3 11.69 inch 16.5 inch pagedimen
184 /A4 8.26 inch 11.69 inch pagedimen
185 /A4Small 7.47 inch 10.85 inch pagedimen
186 /B4 10.125 inch 14.33 inch pagedimen
187 /B5 7.16 inch 10.125 inch pagedimen
192 /Letter {lettertray} def
193 /LetterSmall {lettertray lettersmall} def
194 /Tabloid {11x17tray} def
195 /Ledger {ledgertray} def
196 /Legal {legaltray} def
197 /Statement {statementtray} def
198 /Executive {executivetray} def
201 /A4Small {a4tray a4small} def
204 /unknown {unknown} def
205 papersizedict dup papername known {papername} {/unknown} ifelse get
207 /FMdicttop countdictstack 1 add def
208 statusdict begin stopped end
209 countdictstack -1 FMdicttop {pop end} for
214 /LetterSmall {lettersmall} def
218 /Statement {statement} def
219 /Executive {executive} def
222 /A4Small {a4small} def
225 /unknown {unknown} def
226 papersizedict dup papername known {papername} {/unknown} ifelse get
230 /desperatepapersize {
231 statusdict /setpageparams known
233 paperwidth paperheight 0 1
235 {setpageparams} stopped pop
240 orgmatrix currentmatrix pop
246 /dpi 72 0 dmatrix defaultmatrix dtransform
247 dup mul exch dup mul add sqrt def
248 /freq dpi 18.75 div 8 div round dup 0 eq {pop 1} if 8 mul dpi exch div def
249 /sangle 1 0 dmatrix defaultmatrix dtransform exch atan def
251 /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef
252 /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef
253 /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef
254 /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef
255 /.notdef /.notdef /.notdef /.notdef /space /exclam /quotedbl
256 /numbersign /dollar /percent /ampersand /quotesingle /parenleft
257 /parenright /asterisk /plus /comma /hyphen /period /slash /zero /one
258 /two /three /four /five /six /seven /eight /nine /colon /semicolon
259 /less /equal /greater /question /at /A /B /C /D /E /F /G /H /I /J /K
260 /L /M /N /O /P /Q /R /S /T /U /V /W /X /Y /Z /bracketleft /backslash
261 /bracketright /asciicircum /underscore /grave /a /b /c /d /e /f /g /h
262 /i /j /k /l /m /n /o /p /q /r /s /t /u /v /w /x /y /z /braceleft /bar
263 /braceright /asciitilde /.notdef /Adieresis /Aring /Ccedilla /Eacute
264 /Ntilde /Odieresis /Udieresis /aacute /agrave /acircumflex /adieresis
265 /atilde /aring /ccedilla /eacute /egrave /ecircumflex /edieresis
266 /iacute /igrave /icircumflex /idieresis /ntilde /oacute /ograve
267 /ocircumflex /odieresis /otilde /uacute /ugrave /ucircumflex
268 /udieresis /dagger /.notdef /cent /sterling /section /bullet
269 /paragraph /germandbls /registered /copyright /trademark /acute
270 /dieresis /.notdef /AE /Oslash /.notdef /.notdef /.notdef /.notdef
271 /yen /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef
272 /ordfeminine /ordmasculine /.notdef /ae /oslash /questiondown
273 /exclamdown /logicalnot /.notdef /florin /.notdef /.notdef
274 /guillemotleft /guillemotright /ellipsis /.notdef /Agrave /Atilde
275 /Otilde /OE /oe /endash /emdash /quotedblleft /quotedblright
276 /quoteleft /quoteright /.notdef /.notdef /ydieresis /Ydieresis
277 /fraction /currency /guilsinglleft /guilsinglright /fi /fl /daggerdbl
278 /periodcentered /quotesinglbase /quotedblbase /perthousand
279 /Acircumflex /Ecircumflex /Aacute /Edieresis /Egrave /Iacute
280 /Icircumflex /Idieresis /Igrave /Oacute /Ocircumflex /.notdef /Ograve
281 /Uacute /Ucircumflex /Ugrave /dotlessi /circumflex /tilde /macron
282 /breve /dotaccent /ring /cedilla /hungarumlaut /ogonek /caron
293 Encoding StandardEncoding eq
295 /Encoding DiacriticEncoding def
314 /onbits 0 def /offbits 0 def
315 freq sangle landscape {90 add} if
318 /xindex x 1 add 2 div bpside mul cvi def
319 /yindex y 1 add 2 div bpside mul cvi def
320 bstring yindex bwidth mul xindex 8 idiv add get
321 1 7 xindex 8 mod sub bitshift and 0 ne
322 {/onbits onbits 1 add def 1}
323 {/offbits offbits 1 add def 0}
328 offbits offbits onbits add div FMsetgray
335 orgxfer cvx settransfer
336 orgfreq organgle orgproc cvx setscreen
349 % array of arrays Hue and Sat values for the separations [HUE BRIGHT]
358 [0.16 1.0] % comment / yellow
367 0 get /HUE exch store
368 1 get /BRIGHT exch store
369 HUE 0 eq BRIGHT 0 eq and
370 {1.0 SAT sub setgray}
371 {HUE SAT BRIGHT sethsbcolor}
375 /SAT exch 1.0 exch sub store
376 HUE 0 eq BRIGHT 0 eq and
377 {1.0 SAT sub setgray}
378 {HUE SAT BRIGHT sethsbcolor}
388 /FMsetgray {setgray} bind def
395 transform round exch round exch itransform
398 dtransform round exch round exch idtransform
401 0 dtransform exch cvi 2 idiv 2 mul 1 add exch idtransform pop
404 lnormalize setlinewidth
410 fillprocs exch get exec
413 gsave eofill grestore
418 /M {newpath moveto} bind def
420 /D {curveto} bind def
421 /O {closepath} bind def
428 2 1 n {pop normalize lineto} for
460 x1 y2 x2 y2 rad arcto
461 x2 y2 x2 y1 rad arcto
462 x2 y1 x1 y1 rad arcto
463 x1 y1 x1 y2 rad arcto
497 0 32 3 2 roll widthshow
503 0 32 3 2 roll widthshow
519 0 32 4 2 roll 0 exch awidthshow
525 0 32 4 2 roll 0 exch awidthshow
545 /dl dx dx mul dy dy mul add sqrt def
547 /t currentlinewidth def
551 /r [Cos Sin Sin neg Cos 0.0 0.0] def
552 /t2 t 2.5 mul 3.5 max def
557 dl t 2.7 mul sub 0.0 rlineto
560 x dx add y dy add translate
562 t 0.67 mul setlinewidth
563 t 1.61 mul neg 0.0 translate
565 t2 1.7 mul neg t2 2.0 div moveto
567 t2 1.7 mul neg t2 2.0 div neg lineto
576 normalize translate 0.0 0.0 moveto
578 0.0 0.0 1.0 5 3 roll arc
586 2 index 2 div add exch 3 index 2 div sub exch
587 normalize 2 index 2 div sub exch 3 index 2 div add exch
590 0.0 0.0 1.0 5 3 roll arc
603 /FMsaveobject FMLOCAL
607 /FMdicttop countdictstack 1 add def
608 /FMoptop count 4 sub def
609 /FMsaveobject save def
613 3 index neg 3 index neg translate
616 count -1 FMoptop {pop pop} for
617 countdictstack -1 FMdicttop {pop end} for
633 0 1 sl 1 sub {str exch val put} for
637 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0223
638 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0223
640 {0 hx} {1 hx} {2 hx} {3 hx} {4 hx} {5 hx} {6 hx} {7 hx} {8 hx} {9 hx}
641 {10 hx} {11 hx} {12 hx} {13 hx} {14 hx} {15 hx} {16 hx} {17 hx} {18 hx}
642 {19 hx} {gn hx} {0} {1} {2} {3} {4} {5} {6} {7} {8} {9} {10} {11} {12}
643 {13} {14} {15} {16} {17} {18} {19} {gn} {0 wh} {1 wh} {2 wh} {3 wh}
644 {4 wh} {5 wh} {6 wh} {7 wh} {8 wh} {9 wh} {10 wh} {11 wh} {12 wh}
645 {13 wh} {14 wh} {gn wh} {0 bl} {1 bl} {2 bl} {3 bl} {4 bl} {5 bl} {6 bl}
646 {7 bl} {8 bl} {9 bl} {10 bl} {11 bl} {12 bl} {13 bl} {14 bl} {gn bl}
647 {0 fl} {1 fl} {2 fl} {3 fl} {4 fl} {5 fl} {6 fl} {7 fl} {8 fl} {9 fl}
648 {10 fl} {11 fl} {12 fl} {13 fl} {14 fl} {gn fl}
681 ws 0 len getinterval im pos len getinterval copy pop
687 bs 0 len getinterval im pos len getinterval copy pop
694 /val cf s1 readhexstring pop 0 get def
695 pos 1 pos len add 1 sub {im exch val put} for
700 cf exch readhexstring pop pop
711 8 eq {pop} {1 eq {7 add 8 idiv} {3 add 4 idiv} ifelse} ifelse
726 translate rotate scale /h exch def /w exch def
731 /is im 0 lb getinterval def
732 ws 0 lb getinterval is copy pop
734 w h d [w 0 0 h neg 0 h]
752 translate rotate scale /h exch def /w exch def
755 /is w d wbytes string def
757 w h d [w 0 0 h neg 0 h]
758 {cf is readhexstring pop} image
766 /proc2 exch cvlit def
767 /proc1 exch cvlit def
768 /newproc proc1 length proc2 length add array def
769 newproc 0 proc1 putinterval
770 newproc proc1 length proc2 putinterval
773 /ngrayt 256 array def
775 /nbluet 256 array def
776 /ngreent 256 array def
795 /cynu 1 red indx get 255 div sub def
796 /magu 1 green indx get 255 div sub def
797 /yelu 1 blue indx get 255 div sub def
798 /k cynu magu min yelu min def
799 /u k currentundercolorremoval exec def
800 nredt indx 1 0 cynu u sub max sub redt exec put
801 ngreent indx 1 0 magu u sub max sub grnt exec put
802 nbluet indx 1 0 yelu u sub max sub blut exec put
803 ngrayt indx 1 k currentblackgeneration exec sub gryt exec put
805 {255 mul cvi nredt exch get}
806 {255 mul cvi ngreent exch get}
807 {255 mul cvi nbluet exch get}
808 {255 mul cvi ngrayt exch get}
810 {pop 0} setundercolorremoval
811 {} setblackgeneration
816 0 1 255 {/indx exch def
819 green indx get 151 mul
821 add add 256 idiv put} for
823 {255 mul cvi tran exch get 255.0 div}
824 exch Fmcc settransfer
829 translate rotate scale /h exch def /w exch def
832 /is w d wbytes string def
834 w h d [w 0 0 h neg 0 h]
835 {cf is readhexstring pop} {is} {is} true 3 colorimage
842 translate rotate scale /h exch def /w exch def
847 /is im 0 lb getinterval def
848 ws 0 lb getinterval is copy pop
850 w h d [w 0 0 h neg 0 h]
851 {ip} {is} {is} true 3 colorimage
856 8 {fakecolorsetup} COMMONBITMAP
859 8 {fakecolorsetup} COMMONBITMAPc
867 1 1 612 792 0 1 7 FMDOCUMENT
868 /fillprocs 32 array def
869 fillprocs 0 { 0.000000 grayness } put
870 fillprocs 1 { 0.100000 grayness } put
871 fillprocs 2 { 0.300000 grayness } put
872 fillprocs 3 { 0.500000 grayness } put
873 fillprocs 4 { 0.700000 grayness } put
874 fillprocs 5 { 0.900000 grayness } put
875 fillprocs 6 { 0.970000 grayness } put
876 fillprocs 7 { 1.000000 grayness } put
877 fillprocs 8 {<0f87c3e1f0783c1e> 8 1 setpattern } put
878 fillprocs 9 {<0f1e3c78f0e1c387> 8 1 setpattern } put
879 fillprocs 10 {<cccccccccccccccc> 8 1 setpattern } put
880 fillprocs 11 {<ffff0000ffff0000> 8 1 setpattern } put
881 fillprocs 12 {<8142241818244281> 8 1 setpattern } put
882 fillprocs 13 {<8040201008040201> 8 1 setpattern } put
883 fillprocs 14 {<03060c183060c081> 8 1 setpattern } put
885 fillprocs 16 { 1.000000 grayness } put
886 fillprocs 17 { 0.900000 grayness } put
887 fillprocs 18 { 0.700000 grayness } put
888 fillprocs 19 { 0.500000 grayness } put
889 fillprocs 20 { 0.300000 grayness } put
890 fillprocs 21 { 0.100000 grayness } put
891 fillprocs 22 { 0.030000 grayness } put
892 fillprocs 23 { 0.000000 grayness } put
893 fillprocs 24 {<f0783c1e0f87c3e1> 8 1 setpattern } put
894 fillprocs 25 {<f0e1c3870f1e3c78> 8 1 setpattern } put
895 fillprocs 26 {<3333333333333333> 8 1 setpattern } put
896 fillprocs 27 {<0000ffff0000ffff> 8 1 setpattern } put
897 fillprocs 28 {<7ebddbe7e7dbbd7e> 8 1 setpattern } put
898 fillprocs 29 {<7fbfdfeff7fbfdfe> 8 1 setpattern } put
899 fillprocs 30 {<fcf9f3e7cf9f3f7e> 8 1 setpattern } put
902 0 12 /Times-Roman FMDEFINEFONT
903 1 24 /Times-Roman FMDEFINEFONT
904 2 16 /Times-Bold FMDEFINEFONT
906 %%BeginPaperSize: Letter
908 612 792 0 FMBEGINPAGE
915 (Network Working Group) 72 712 T
916 (J. Moy, Editor) 470.7 712 T
917 (Request for Comments: 1245) 72 698 T
918 (Proteon, Inc.) 478.38 698 T
919 (July 1991) 493.02 684 T
925 ([Page 1]) 499.7 75.95 T
931 (OSPF protocol analysis) 192.72 587 T
933 (Status of this Memo) 72 514.33 T
935 -0.23 (This memo provides information for the Internet community) 72 487 P
936 -0.23 (. It does not specify any Internet stan-) 360.42 487 P
937 (dard. Distribution of this memo is unlimited.) 72 473 T
941 -0.11 (This is the \336rst of two reports on the OSPF protocol. These reports are required by the IAB/IESG ) 72 421 P
942 (in order for an Internet routing protocol to advance to Draft Standard Status. OSPF is a TCP/IP ) 72 407 T
943 -0.28 (routing protocol, designed to be used internal to an Autonomous System \050in other words, OSPF is ) 72 393 P
944 (an Interior Gateway Protocol\051.) 72 379 T
946 -0.09 (ersion 1 of the OSPF protocol was published in RFC 1) 79.33 353 P
947 -0.09 (131. Since then OSPF version 2 has been ) 339.85 353 P
948 -0.22 (developed. V) 72 339 P
949 -0.22 (ersion 2 has been documented in RFC 1247. The changes between version 1 and ver-) 134.4 339 P
950 -0 (sion 2 of the OSPF protocol are explained in Appendix F of RFC 1247. It is OSPF V) 72 325 P
951 -0 (ersion 2 that ) 477.72 325 P
952 (is the subject of this report.) 72 311 T
953 (This report attempts to summarize the key features of OSPF V2. It also attempts to analyze how ) 72 285 T
954 (the protocol will perform and scale in the Internet.) 72 271 T
955 (Please send comments to ospf@trantor) 72 245 T
956 (.umd.edu.) 258.27 245 T
959 1 10 /Times-Roman FMDEFINEFONT
961 612 792 0 FMBEGINPAGE
969 (OSPF protocol analysis) 249.36 712 T
970 (July 1991) 493.02 712 T
976 ([Page 2]) 499.7 73 T
983 (able of Contents) 81.19 673.33 T
986 (Introduction) 108 650 T
987 (..............................................................................................................) 167.91 650 T
991 (Acknowledgments) 144 635.33 T
992 (...............................................................................................................) 219.88 635.33 T
996 (Key features of the OSPF protocol) 108 616 T
997 (..........................................................................) 275.85 616 T
1000 (Cost of the protocol) 108 596 T
1001 (..................................................................................................) 203.89 596 T
1005 ( Operational data) 144 581.33 T
1006 (.................................................................................................................) 214.88 581.33 T
1009 (Link bandwidth) 144 567.33 T
1010 (...................................................................................................................) 209.88 567.33 T
1013 (Router memory) 144 553.33 T
1014 (....................................................................................................................) 207.39 553.33 T
1017 (Router CPU) 144 539.33 T
1018 (.......................................................................................................................) 194.89 539.33 T
1019 (10) 494.01 539.33 T
1021 (Role of Designated Router) 144 525.33 T
1022 (................................................................................................) 252.36 525.33 T
1026 (Summary) 144 511.33 T
1027 (...........................................................................................................................) 184.9 511.33 T
1032 (Suitable environments) 108 492 T
1033 (............................................................................................) 215.88 492 T
1036 (Unsuitable environments) 108 472 T
1037 (.......................................................................................) 230.87 472 T
1040 (Reference Documents) 108 452 T
1041 (............................................................................................) 215.88 452 T
1045 3 14 /Times-Bold FMDEFINEFONT
1047 612 792 0 FMBEGINPAGE
1055 (OSPF protocol analysis) 249.36 712 T
1056 (July 1991) 493.02 712 T
1062 ([Page 3]) 499.7 73 T
1068 (1.0 Intr) 72 673.33 T
1069 (oduction) 127.23 673.33 T
1071 -0.02 (This document addresses, for OSPF V2, the requirements set forth by the IAB/IESG for an Inter-) 72 646 P
1072 -0.19 (net routing protocol to advance to Draft Standard state. This requirements are brie\337y summarized ) 72 632 P
1074 (. The remaining sections of this report document how OSPF V2 satis\336es these require-) 100.53 618 T
1077 (What are the key features and algorithms of the protocol?) 85.54 584 T
1079 (How much link bandwidth, router memory and router CPU cycles does the protocol consume ) 85.54 564 T
1080 (under normal conditions?) 85.54 550 T
1082 (For these metrics, how does the usage scale as the routing environment grows? This should ) 85.54 530 T
1083 (include topologies at least an order of magnitude lar) 85.54 516 T
1084 (ger than the current environment.) 335.14 516 T
1086 (What are the limits of the protocol for these metrics? \050i.e., when will the routing protocol ) 85.54 496 T
1087 (break?\051 ) 85.54 482 T
1089 (For what environments is the protocol well suited, and for what is it not suitable? ) 85.54 462 T
1091 (1.1 Acknowledgments) 72 428.67 T
1093 -0.03 (The OSPF protocol has been developed by the OSPF W) 72 402 P
1094 -0.03 (orking Group of the Internet Engineering ) 339.64 402 P
1096 (ask Force. ) 78.49 388 T
1099 4 12 /Times-Bold FMDEFINEFONT
1101 612 792 0 FMBEGINPAGE
1109 (OSPF protocol analysis) 249.36 712 T
1110 (July 1991) 493.02 712 T
1116 ([Page 4]) 499.7 73 T
1122 (2.0 Key featur) 72 673.33 T
1123 (es of the OSPF pr) 172.97 673.33 T
1124 (otocol) 293.49 673.33 T
1126 (This section summarizes the key features of the OSPF protocol. OSPF is an) 72 646 T
1128 ( Internal gateway ) 434.78 646 T
1130 -0.2 (otocol) 83.78 632 P
1132 -0.2 (; it is designed to be used internal to a single Autonomous System. OSPF uses) 114.42 632 P
1134 -0.2 ( link-state ) 486.43 632 P
1135 -0.36 (or SPF-based) 72 618 P
1137 -0.36 ( technology \050as compared to the distance-vector or Bellman-Ford technology found ) 140.6 618 P
1138 -0.48 (in routing protocols such as RIP\051. Individual ) 72 604 P
1140 -0.48 (link state advertisements \050LSAs\051) 285.2 604 P
1142 -0.48 ( describe pieces of ) 449.99 604 P
1143 -0.13 (the OSPF routing domain \050Autonomous System\051. These LSAs are \337ooded throughout the routing ) 72 590 P
1144 (domain, forming the ) 72 576 T
1146 (link state database) 173.27 576 T
1148 (. Each router has an identical link state database; syn-) 268.56 576 T
1149 (chronization of link state databases is maintained via a ) 72 562 T
1152 (eliable \337ooding algorithm) 341.92 562 T
1154 (. From this ) 473.2 562 T
1155 (link state database, each router builds a routing table by calculating a shortest-path tree, with the ) 72 548 T
1156 (root of the tree being the calculating router itself. This calculation is commonly referred to as the ) 72 534 T
1158 (Dijkstra pr) 72 520 T
1159 (ocedur) 129.41 520 T
1163 (Link state advertisements are small. Each advertisement describes a small pieces of the OSPF ) 72 494 T
1164 (routing domain, namely either: the neighborhood of a single router) 72 480 T
1165 (, the neighborhood of a single ) 391.97 480 T
1166 (transit network, a single inter) 72 466 T
1167 (-area route \050see below\051 or a single external route.) 212 466 T
1168 (The other key features of the OSPF protocol are:) 72 440 T
1171 -0.31 (Adjacency bringup) 85.54 420 P
1173 -0.31 (. ) 183.51 420 P
1175 -0.31 (Certain pairs of OSPF r) 189.2 420 P
1176 -0.31 (outers become \322adjacent\323) 311.01 420 P
1178 -0.31 (. As an adjacency is ) 442.96 420 P
1179 (formed, the two routers synchronize their link state databases by ) 85.54 406 T
1181 (exchanging database sum-) 397.64 406 T
1182 (maries) 85.54 392 T
1184 ( in the form of OSPF Database Exchange packets. Adjacent routers then maintain syn-) 120.17 392 T
1185 (chronization of their link state databases through the ) 85.54 378 T
1188 (eliable \337ooding algorithm) 345.13 378 T
1190 (. Routers ) 476.41 378 T
1191 -0.27 (connected by serial lines always become adjacent. On multi-access networks \050e.g., ethernets or ) 85.54 364 P
1192 (X.25 PDNs\051, all routers attached to the network become adjacent to both the Designated ) 85.54 350 T
1193 (Router and the Backup Designated router) 85.54 336 T
1197 -0.02 (Designated r) 85.54 316 P
1198 -0.02 (outer) 150.26 316 P
1199 -0.02 (.) 176.46 316 P
1201 -0.02 ( A Designated Router is elected on all multi-access networks \050e.g., ether-) 179.46 316 P
1202 (nets or X.25 PDNs\051. The network\325) 85.54 302 T
1203 (s Designated Router ) 250.42 302 T
1205 (originates the network LSA) 350.69 302 T
1207 ( describ-) 492.27 302 T
1208 (ing the network\325) 85.54 288 T
1209 (s local environment. It also plays a ) 164.15 288 T
1211 (special r) 334.04 288 T
1212 (ole in the \337ooding algorithm) 376.8 288 T
1215 (since all routers on the network are synchronizing their link state databases by sending and ) 85.54 274 T
1216 (receiving LSAs to/from the Designated Router during the \337ooding process.) 85.54 260 T
1219 -0.46 (Backup Designated Router) 85.54 240 P
1221 -0.46 (. A Backup Designated Router is elected on multi-access networks ) 221.87 240 P
1222 (to speed/ease the transition of Designated Routers when the current Designated Router disap-) 85.54 226 T
1223 (pears. In that event, the Backup DR takes over) 85.54 212 T
1224 (, and does not need to go through the adjacency ) 308.22 212 T
1225 -0.13 (bringup process on the LAN \050since it already had done this in its Backup capacity\051. Also, even ) 85.54 198 P
1226 (before the disappearance of the Designated Router is noticed, the Backup DR will enable the ) 85.54 184 T
1227 (reliable \337ooding algorithm to proceed in the DR\325) 85.54 170 T
1228 (s absence.) 320.39 170 T
1232 612 792 0 FMBEGINPAGE
1240 (OSPF protocol analysis) 249.36 712 T
1241 (July 1991) 493.02 712 T
1247 ([Page 5]) 499.7 73 T
1254 (Non-br) 85.54 676 T
1255 (oadcast multi-access network support.) 122.63 676 T
1257 ( OSPF treats these networks \050e.g., X.25 ) 318.51 676 T
1258 -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
1259 -0.29 (ated\051. Additional con\336guration information is needed however for routers attached to these net-) 85.54 648 P
1260 (work to initially \336nd each other) 85.54 634 T
1264 (OSPF ar) 85.54 614 T
1267 (. OSPF allows the Autonomous Systems to be broken up into regions call areas. ) 146.28 614 T
1268 (This is useful for several reasons. First, it provides an extra level of ) 85.54 600 T
1271 (outing pr) 416.75 600 T
1272 (otection) 464.18 600 T
1274 (: rout-) 504.81 600 T
1275 -0.29 (ing within an area is protected from all information external to the area. Second, by splitting an ) 85.54 586 P
1276 -0.3 (Autonomous System into areas the ) 85.54 572 P
1278 -0.3 (cost of the Dijkstra pr) 254.27 572 P
1279 -0.3 (ocedur) 365.44 572 P
1280 -0.3 (e ) 400.53 572 P
1282 -0.3 (\050in terms of CPU cycles\051 is ) 408.55 572 P
1283 (reduced.) 85.54 558 T
1286 (Flexible import of external r) 85.54 538 T
1287 (outing information.) 230.55 538 T
1289 ( In OSPF) 330.5 538 T
1292 (each external r) 380.19 538 T
1295 ( is imported ) 478.56 538 T
1296 (into the Autonomous System in ) 85.54 524 T
1298 (a separate LSA) 240.47 524 T
1300 (. This reduces the amount of \337ooding traf) 319.08 524 T
1301 (\336c ) 518.07 524 T
1302 (\050since external routes change often, and you want to only \337ood the changes\051. It also enables ) 85.54 510 T
1304 -0.43 (partial r) 85.54 496 P
1305 -0.43 (outing table updates) 127.86 496 P
1307 -0.43 ( when only a single external route changes. OSPF external LSAs ) 230.96 496 P
1308 (also provide the following features. A ) 85.54 482 T
1310 (forwarding addr) 270.4 482 T
1313 ( can be included in the external ) 370.46 482 T
1314 (LSA, eliminating extra-hops at the edge of the Autonomous System. There are two levels of ) 85.54 468 T
1315 (external metrics that can be speci\336ed, ) 85.54 454 T
1317 (type 1) 269.06 454 T
1319 ( and ) 300.04 454 T
1321 (type 2) 323.35 454 T
1323 (. Also, external routes can be tagged ) 354.33 454 T
1324 (with a 32-bit number \050the ) 85.54 440 T
1326 (external r) 211.12 440 T
1327 (oute tag) 261.19 440 T
1329 (; commonly used as an AS number of the route\325) 302.16 440 T
1331 (origin\051, simplifying external route management in a transit Autonomous System.) 85.54 426 T
1334 (Four level r) 85.54 406 T
1335 (outing hierar) 145.27 406 T
1338 (. OSPF has a four level routing hierarchy) 229.9 406 T
1339 (, or trust model: ) 426.32 406 T
1341 (intra-) 505.94 406 T
1347 (inter) 113.96 392 T
1353 (external type 1) 170.58 392 T
1355 ( and ) 246.52 392 T
1357 (external type 2) 269.84 392 T
1359 ( routes. This enables multiple levels of ) 345.78 392 T
1360 (routing protection, and simpli\336es routing management in an Autonomous System.) 85.54 378 T
1364 (irtual links) 93.75 358 T
1366 (. By allowing the con\336guration of virtual links, OSPF ) 150.07 358 T
1369 (emoves topological ) 416.05 358 T
1371 (estrictions) 90.64 344 T
1373 ( on area layout in an Autonomous System.) 143.27 344 T
1376 -0.32 (Authentication of r) 85.54 324 P
1377 -0.32 (outing pr) 182.62 324 P
1378 -0.32 (otocol exchanges) 229.74 324 P
1380 -0.32 (. Every time an OSPF router receives a routing ) 315.03 324 P
1381 (protocol packet, it authenticates the packet before processing it further) 85.54 310 T
1385 -0.03 (Flexible r) 85.54 290 P
1386 -0.03 (outing metric.) 134.26 290 P
1388 -0.03 ( In OSPF) 206.18 290 P
1389 -0.03 (, metric are assigned to outbound router interfaces. The cost ) 249.82 290 P
1390 (of a path is then the sum of the path\325) 85.54 276 T
1391 (s component interfaces. The routing metric itself can be ) 260.42 276 T
1392 (assigned by the system administrator to indicate any combination of network characteristics ) 85.54 262 T
1393 (\050e.g., delay) 85.54 248 T
1394 (, bandwidth, dollar cost, etc.\051.) 138.04 248 T
1397 -0.09 (Equal-cost multipath.) 85.54 228 P
1399 -0.09 ( When multiple best cost routes to a destination exist, OSPF \336nds them ) 196.73 228 P
1400 (and they can be then used to load share traf) 85.54 214 T
1401 (\336c to the destination.) 292.82 214 T
1405 (OS-based r) 93.32 194 T
1406 (outing.) 150.74 194 T
1408 ( Separate sets of routes can be calculated for each IP type of service. For ) 186.4 194 T
1409 (example, low delay traf) 85.54 180 T
1410 (\336c could be routed on one path, while high bandwidth traf) 198.56 180 T
1411 (\336c is routed ) 477.16 180 T
1412 -0.39 (on another) 85.54 166 P
1413 -0.39 (. This is done by \050optionally\051 assigning, to each outgoing router interface, one metric ) 135.44 166 P
1414 (for each IP T) 85.54 152 T
1419 (ariable-length subnet support.) 93.09 132 T
1421 ( OSPF includes support for variable-length subnet masks by ) 248.02 132 T
1422 (carrying a network mask with each advertised destination.) 85.54 118 T
1426 612 792 0 FMBEGINPAGE
1434 (OSPF protocol analysis) 249.36 712 T
1435 (July 1991) 493.02 712 T
1441 ([Page 6]) 499.7 73 T
1448 -0.08 (Stub ar) 85.54 676 P
1449 -0.08 (ea support. ) 123.56 676 P
1451 -0.08 (T) 183.69 676 P
1452 -0.08 (o support routers having insuf) 190.18 676 P
1453 -0.08 (\336cient memory) 333.53 676 P
1454 -0.08 (, areas can be con\336gured as ) 405.63 676 P
1455 (stubs. External LSAs \050often making up the bulk of the Autonomous System\051 are not \337ooded ) 85.54 662 T
1456 (into/throughout stub areas. Routing to external destinations in stub areas is based solely on ) 85.54 648 T
1457 (default.) 85.54 634 T
1461 612 792 0 FMBEGINPAGE
1469 (OSPF protocol analysis) 249.36 712 T
1470 (July 1991) 493.02 712 T
1476 ([Page 7]) 499.7 73 T
1482 (3.0 Cost of the pr) 72 673.33 T
1483 (otocol) 193.4 673.33 T
1485 -0.1 (This section attempts to analyze how the OSPF protocol will perform and scale in the Internet. In ) 72 646 P
1486 (this analysis, we will concentrate on the following four areas:) 72 632 T
1489 (Link bandwidth) 85.54 612 T
1491 (. In OSPF) 168.53 612 T
1492 (, a reliable \337ooding mechanism is used to ensure that router link ) 215.22 612 T
1493 (state databases are remained synchronized. Individual components of the link state databases ) 85.54 598 T
1494 -0.17 (\050the LSAs\051 are refreshed infrequently \050every 30 minutes\051, at least in the absence of topological ) 85.54 584 P
1495 (changes. Still, as the size of the database increases, the amount of link bandwidth used by the ) 85.54 570 T
1496 (\337ooding procedure also increases.) 85.54 556 T
1499 -0.03 (Router memory) 85.54 536 P
1501 -0.03 (. The size of an OSPF link state database can get quite lar) 166.32 536 P
1502 -0.03 (ge, especially in the ) 441.86 536 P
1503 (presence of many external LSAs. This imposes requirements on the amount of router memory ) 85.54 522 T
1504 (available.) 85.54 508 T
1507 (CPU usage) 85.54 488 T
1509 (. In OSPF) 141.83 488 T
1510 (, this is dominated by the length of time it takes to run the shortest path ) 188.52 488 T
1511 (calculation \050Dijkstra procedure\051. This is a function of the number of routers in the OSPF sys-) 85.54 474 T
1515 (Role of the Designated Router) 85.54 440 T
1518 ( The Designated router receives and sends more packets on a ) 241.32 440 T
1519 -0.46 (multi-access networks than the other routers connected to the network. Also, there is some time ) 85.54 426 P
1520 (involved in cutting over to a new Designated Router after the old one fails \050especially when ) 85.54 412 T
1521 (both the Backup Designated Router and the Designated Router fail at the same time\051. For this ) 85.54 398 T
1522 -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
1524 (The remaining section will analyze these areas, estimating how much resources the OSPF proto-) 72 344 T
1525 -0.05 (col will consume, both now and in the future. T) 72 330 P
1526 -0.05 (o aid in this analysis, the next section will present ) 298.93 330 P
1527 (some data that have been collected in actual OSPF \336eld deployments.) 72 316 T
1529 (3.1 Operational data) 72 282.67 T
1531 -0.44 (The OSPF protocol has been deployed in a number of places in the Internet. For a summary of this ) 72 256 P
1532 (deployment, see [1]. Some statistics have been gathered from this operational experience, via ) 72 242 T
1533 -0.03 (local network management facilities. Some of these statistics are presented in the following table:) 72 228 P
1536 5 10 /Times-Bold FMDEFINEFONT
1538 612 792 0 FMBEGINPAGE
1546 (OSPF protocol analysis) 249.36 712 T
1547 (July 1991) 493.02 712 T
1553 ([Page 8]) 499.7 73 T
1558 72 671.98 540 671.98 2 L
1569 (ABLE 1. Pertinent operational statistics) 77.93 677.33 T
1570 (Statistic) 72 655.34 T
1571 (BARRNet) 216 655.34 T
1573 (OARnet) 432 655.34 T
1575 (Data gathering \050duration\051) 72 638.34 T
1576 (99 hours) 216 638.34 T
1577 (277 hours) 324 638.34 T
1578 (28 hours) 432 638.34 T
1579 (Dijkstra frequency) 72 622.34 T
1580 (50 minutes) 216 622.34 T
1581 (25 minutes) 324 622.34 T
1582 (13 minutes) 432 622.34 T
1583 (External incremental frequency) 72 606.34 T
1584 (1.2 minutes) 216 606.34 T
1585 (.98 minutes) 324 606.34 T
1586 (not gathered) 432 606.34 T
1587 (Database turnover) 72 590.34 T
1588 (29.7 minutes) 216 590.34 T
1589 (30.9 minutes) 324 590.34 T
1590 (28.2 minutes) 432 590.34 T
1591 (LSAs per packet) 72 574.34 T
1595 (Flooding retransmits) 72 558.34 T
1600 (The \336rst line in the above table show the length of time that statistics were gathered on the three ) 72 533.01 T
1601 (networks. A brief description of the other statistics follows:) 72 519.01 T
1604 (Dijkstra fr) 85.54 499.01 T
1605 (equency) 140.27 499.01 T
1606 (. ) 181.59 499.01 T
1608 (In OSPF) 187.59 499.01 T
1609 (, the Dijkstra calculation involves only those routers and transit ) 228.28 499.01 T
1610 -0.14 (networks belonging to the AS. The Dijkstra is run only when something in the system changes ) 85.54 485.01 P
1611 (\050like a serial line between two routers goes down\051. Note that in these operational systems, the ) 85.54 471.01 T
1612 (Dijkstra process runs only infrequently \050the most frequent being every 13 minutes\051.) 85.54 457.01 T
1615 (External incr) 85.54 437.01 T
1616 (emental fr) 153.61 437.01 T
1617 (equency) 206.35 437.01 T
1619 (. In OSPF) 247.54 437.01 T
1620 (, when an external route changes only its entry in ) 294.23 437.01 T
1621 -0.13 (the routing table is recalculated. These are called external incremental updates. Note that these ) 85.54 423.01 P
1622 (happen much more frequently than the Dijkstra procedure. \050in other words, incremental ) 85.54 409.01 T
1623 (updates are saving quite a bit of processor time\051.) 85.54 395.01 T
1626 -0.45 (Database turnover) 85.54 375.01 P
1627 -0.45 (.) 179.58 375.01 P
1629 -0.45 ( In OSPF) 182.58 375.01 P
1630 -0.45 (, link state advertisements are refreshed at a minimum of every 30 ) 225.36 375.01 P
1631 (minutes. New advertisement instances are sent out more frequently when some part of the ) 85.54 361.01 T
1632 -0.2 (topology changes. The table shows that, even taking topological changes into account, on aver-) 85.54 347.01 P
1633 (age an advertisement is updated close to only every 30 minutes. This statistic will be used in ) 85.54 333.01 T
1634 (the link bandwidth calculations below) 85.54 319.01 T
1635 (. Note that NSI actually shows advertisements updated ) 267.31 319.01 T
1636 (every 30.7 \050> 30\051 minutes. This probably means that at one time earlier in the measurement ) 85.54 305.01 T
1637 (period, NSI had a smaller link state database that it did at the end.) 85.54 291.01 T
1640 -0.39 (LSAs per packet.) 85.54 271.01 P
1642 -0.39 ( In OSPF) 173.04 271.01 P
1643 -0.39 (, multiple LSAs can be included in either Link State Update or Link ) 215.95 271.01 P
1644 -0.35 (State Acknowledgment packets.The table shows that, on average, around 3 LSAs are carried in ) 85.54 257.01 P
1645 (a single packet. This statistic is used when calculating the header overhead in the link band-) 85.54 243.01 T
1646 (width calculation below) 85.54 229.01 T
1647 (. This statistic was derived by diving the number of LSAs \337ooded by ) 200.01 229.01 T
1648 (the number of \050non-hello\051 multicasts sent.) 85.54 215.01 T
1651 (Flooding r) 85.54 195.01 T
1652 (etransmits.) 138.97 195.01 T
1654 ( This counts both retransmission of LS Update packets and Link State ) 195.92 195.01 T
1655 (Acknowledgment packets, as a percentage of the original multicast \337ooded packets. The table ) 85.54 181.01 T
1656 (shows that \337ooding is working well, and that retransmits can be ignored in the link bandwidth ) 85.54 167.01 T
1657 (calculation below) 85.54 153.01 T
1662 612 792 0 FMBEGINPAGE
1670 (OSPF protocol analysis) 249.36 712 T
1671 (July 1991) 493.02 712 T
1677 ([Page 9]) 499.7 73 T
1683 (3.2 Link bandwidth) 72 674.67 T
1685 -0.02 (In this section we attempt to calculate how much link bandwidth is consumed by the OSPF \337ood-) 72 648 P
1686 (ing process. The amount of link bandwidth consumed increases linearly with the number of ) 72 634 T
1687 (advertisements present in the OSPF database.W) 72 620 T
1688 (e assume that the majority of advertisements in ) 300.88 620 T
1689 (the database will be AS external LSAs \050operationally this is true, see [1]\051.) 72 606 T
1690 (From the statistics presented in Section 3.1, any particular advertisement is \337ooded \050on average\051 ) 72 580 T
1691 (every 30 minutes. In addition, three advertisements \336t in a single packet. \050This packet could be ) 72 566 T
1692 (either a Link State Update packet or a Link State Acknowledgment packet; in this analysis we ) 72 552 T
1693 (select the Link State Update packet, which is the lar) 72 538 T
1694 (ger\051. An AS external LSA is 36 bytes long. ) 320.93 538 T
1695 (Adding one third of a packet header \050IP header plus OSPF Update packet\051 yields 52 bytes. T) 72 524 T
1696 (rans-) 515.59 524 T
1697 (mitting this amount of data every 30 minutes gives an average rate of 23/100 bits/second.) 72 510 T
1698 -0.05 (If you want to limit your routing traf) 72 484 P
1699 -0.05 (\336c to 5% of the link\325) 247.03 484 P
1700 -0.05 (s total bandwidth, you get the following ) 345.75 484 P
1701 (maximums for database size:) 72 470 T
1703 72 439.98 540 439.98 2 L
1714 (ABLE 2. Database size as a function of link speed \0505% utilization\051) 77.93 445.33 T
1715 (Speed) 180 423.34 T
1716 (# external advertisements) 288 423.34 T
1718 (9.6 Kb) 180 406.34 T
1720 (56 Kb) 180 390.34 T
1721 (12,174) 288 390.34 T
1723 -0.46 (Higher line speeds have not been included, because other factors will then limit database size \050like ) 72 365.01 P
1724 -0.12 (router memory\051 before line speed becomes a factor) 72 351.01 P
1725 -0.12 (. Note that in the above calculation, the size of ) 315.32 351.01 P
1726 -0.06 (the data link header was not taken into account. Also, note that while the OSPF database is likely ) 72 337.01 P
1727 (to be mostly external LSAs, other LSAs have a size also. As a ballpark estimate, router links and ) 72 323.01 T
1728 -0.01 (network links are generally three times as lar) 72 309.01 P
1729 -0.01 (ge as an AS external link, with summary link adver-) 287.18 309.01 P
1730 (tisements being the same size as external link LSAs.) 72 295.01 T
1731 (OSPF consumes considerably less link bandwidth than RIP) 72 269.01 T
1732 (. This has been shown experimentally ) 355.51 269.01 T
1733 (in the NSI network. See Jef) 72 255.01 T
1734 (frey Bur) 203.69 255.01 T
1735 (gan\325) 243.77 255.01 T
1736 (s \322NASA Sciences Internet\323 report in [3].) 264.42 255.01 T
1738 (3.3 Router memory) 72 221.67 T
1740 -0.1 (Memory requirements in OSPF are dominated by the size of the link state database. As in the pre-) 72 195.01 P
1741 (vious section, it is probably safe to assume that most of the advertisements in the database are ) 72 181.01 T
1742 (external LSAs. While an external LSA is 36 bytes long, it is generally stored by an OSPF imple-) 72 167.01 T
1743 -0.34 (mentation together with some support data. So a good estimate of router memory consumed by an ) 72 153.01 P
1744 (external LSA is probably 64 bytes. So a database having 10,000 external LSAs will consume ) 72 139.01 T
1745 (640K bytes of router memory) 72 125.01 T
1746 (. OSPF de\336nitely requires more memory than RIP) 213.79 125.01 T
1751 612 792 0 FMBEGINPAGE
1759 (OSPF protocol analysis) 249.36 712 T
1760 (July 1991) 493.02 712 T
1766 ([Page 10]) 493.7 73 T
1771 -0.35 (Using the Proteon P4200 implementation as an example, the P4200 has 2Mbytes of memory) 72 676 P
1772 -0.35 (. This ) 510.38 676 P
1773 -0.02 (is shared between instruction, data and packet buf) 72 662 P
1774 -0.02 (fer memory) 310.78 662 P
1775 -0.02 (. The P4200 has enough memory to ) 366.26 662 P
1776 (store 10, 000 external LSAs, and still have enough packet buf) 72 648 T
1777 (fer memory available to run a rea-) 367.58 648 T
1778 (sonable number of interfaces.) 72 634 T
1779 (Also, note that while the OSPF database is likely to be mostly external LSAs, other LSAs have a ) 72 608 T
1780 -0.06 (size also. As a ballpark estimate, router links and network links consume generally three times as ) 72 594 P
1781 (much memory as an AS external link, with summary link advertisements being the same size as ) 72 580 T
1782 (external link LSAs.) 72 566 T
1784 (3.4 Router CPU) 72 532.67 T
1786 (Assume that, as the size of the OSPF routing domain grows, the number of interfaces per router ) 72 506 T
1787 (stays bounded. Then the Dijkstra calculation is of order \050n * log \050n\051\051, where n is the number of ) 72 492 T
1788 (routers in the routing domain. \050This is the complexity of the Dijkstra algorithm in a sparse net-) 72 478 T
1789 (work\051. Of course, it is implementation speci\336c as to how expensive the Dijkstra really is.) 72 464 T
1791 (e have no experimental numbers for the cost of the Dijkstra calculation in a real OSPF imple-) 82.36 438 T
1792 (mentation. However) 72 424 T
1793 (, Steve Deering presented results for the Dijkstra calculation in the \322MOSPF ) 169.45 424 T
1794 (meeting report\323 in [3]. Steve\325) 72 410 T
1795 (s calculation was done on a DEC 5000 \05010 mips processor\051, using ) 212.9 410 T
1796 (the Stanford internet as a model. His graphs are based on numbers of networks, not number of ) 72 396 T
1797 (routers. However) 72 382 T
1798 (, if we extrapolate that the ratio of routers to networks remains the same, the ) 154.78 382 T
1799 (time to run Dijkstra for 200 routers in Steve\325) 72 368 T
1800 (s implementation was around 15 milliseconds.) 285.87 368 T
1801 -0.46 (This seems a reasonable cost, particularly when you notice that the Dijkstra calculation is run very ) 72 342 P
1802 (infrequently in operational deployments. In the three networks presented in Section 3.1, Dijkstra ) 72 328 T
1803 -0.35 (was run on average only every 13 to 50 minutes. Since the Dijkstra is run so infrequently) 72 314 P
1804 -0.35 (, it seems ) 493.06 314 P
1805 -0.02 (likely that OSPF overall consumes less CPU than RIP \050because of RIP\325) 72 300 P
1806 -0.02 (s frequent updates, requir-) 413.95 300 P
1807 (ing routing table lookups\051.) 72 286 T
1808 (As another example, the routing algorithm in MILNET is SPF-based. MILNET\325) 72 260 T
1809 (s current size is ) 456.42 260 T
1810 -0.02 (230 nodes, and the routing calculation still consumes less than 5% of the MILNET switches\325 pro-) 72 246 P
1811 (cessor bandwidth [4]. Because the routing algorithm in the MILNET adapts to network load, it ) 72 232 T
1812 (runs the Dijkstra process quite frequently \050on the order of seconds as compared to OSPF\325) 72 218 T
1813 (s min-) 499.7 218 T
1814 (utes\051. However) 72 204 T
1815 (, it should be noted that the routing algorithm in MILNET incrementally updates ) 144.79 204 T
1816 (the SPF-tree, while OSPF rebuilds it from scratch at each Dijkstra calculation) 72 190 T
1818 (s Area capability provides a way to reduce Dijkstra overhead, if it becomes a burden. The ) 104 164 T
1819 -0 (routing domain can be split into areas. The extent of the Dijkstra calculation \050and its complexity\051 ) 72 150 P
1820 (is limited to a single area at a time.) 72 136 T
1824 612 792 0 FMBEGINPAGE
1832 (OSPF protocol analysis) 249.36 712 T
1833 (July 1991) 493.02 712 T
1839 ([Page 11]) 493.7 73 T
1845 (3.5 Role of Designated Router) 72 674.67 T
1847 (This section explores the number of routers that can be attached to a single network. As the num-) 72 648 T
1848 -0.36 (ber of routers attached to a network grows, so does the amount of OSPF routing traf) 72 634 P
1849 -0.36 (\336c seen on the ) 469.48 634 P
1850 (network. Some of this is Hello traf) 72 620 T
1851 (\336c, which is generally multicast by each router every 10 sec-) 238.01 620 T
1852 -0.07 (onds. This burden is borne by all routers attached to the network. However) 72 606 P
1853 -0.07 (, because of its special ) 429.77 606 P
1854 -0.08 (role in the \337ooding process, the Designated router ends up sending more Link State Updates than ) 72 592 P
1855 (the other routers on the network. Also, the Designated Router receives Link State Acknowledg-) 72 578 T
1856 -0.15 (ments from all attached routers, while the other routers just receive them from the DR. \050Although ) 72 564 P
1857 (it is important to note that the rate of Link State Acknowledgments will generally be limited to ) 72 550 T
1858 (one per second from each router) 72 536 T
1859 (, because acknowledgments are generally delayed.\051) 226.38 536 T
1860 -0.22 (So, if the amount of protocol traf) 72 510 P
1861 -0.22 (\336c on the LAN becomes a limiting factor) 228.71 510 P
1862 -0.22 (, the limit is likely to be ) 424.24 510 P
1863 (detected in the Designated Router \336rst. However) 72 496 T
1864 (, such a limit is not expected to be reached in ) 305.68 496 T
1865 (practice. The amount of routing protocol traf) 72 482 T
1866 (\336c generated by OSPF has been shown to be small ) 286.62 482 T
1867 -0.11 (\050see Section 3.2\051. Also, if need be OSPF\325) 72 468 P
1868 -0.11 (s hello timers can be con\336gured to reduce the amount of ) 268.43 468 P
1869 (protocol traf) 72 454 T
1870 (\336c on the network. Note that more than 50 routers have been simulated attached to a ) 131.4 454 T
1871 (single LAN \050see [1]\051. Also, in interoperability testing 13 routers have been attached to a single ) 72 440 T
1872 (ethernet with no problems encountered.) 72 426 T
1873 -0.02 (Another factor in the number of routers attached to a single network is the cutover time when the ) 72 400 P
1874 -0.17 (Designated Router fails. OSPF has a Backup Designated Router so that the cutover does not have ) 72 386 P
1875 -0.31 (to wait for the new DR to synchronize \050the adjacency bring-up process mentioned earlier\051 with all ) 72 372 P
1876 -0.43 (the other routers on the LAN; as a Backup DR it had already synchronized. However) 72 358 P
1877 -0.43 (, in those rare ) 473.46 358 P
1878 -0.33 (cases when both DR and Backup DR crash at the same time, the new DR will have to synchronize ) 72 344 P
1879 (\050via the adjacency bring-up process\051 with all other routers before becoming functional. Field ) 72 330 T
1880 -0.44 (experience show that this synchronization process takes place in a timely fashion \050see the OARnet ) 72 316 P
1881 (report in [1]\051. However) 72 302 T
1882 (, this may be an issue in systems that have many routers attached to a sin-) 183.42 302 T
1883 (gle network.) 72 288 T
1884 -0.15 (In the unlikely event that the number of routers attached to a LAN becomes a problem, either due ) 72 262 P
1885 (to the amount of routing protocol traf) 72 248 T
1886 (\336c or the cutover time, the LAN can be split into separate ) 251 248 T
1887 (pieces \050similar to splitting up the AS into separate areas\051.) 72 234 T
1889 (3.6 Summary) 72 200.67 T
1891 (In summary) 72 174 T
1892 (, it seems like the most likely limitation to the size of an OSPF system is available ) 128.85 174 T
1893 -0.4 (router memory) 72 160 P
1894 -0.4 (. W) 142.43 160 P
1895 -0.4 (e have given as 10,000 as the number of external LSAs that can be supported by ) 158.39 160 P
1896 (the memory available in one con\336guration of a particular implementation \050the Proteon P4200\051. ) 72 146 T
1897 -0.09 (Other implementations may vary; nowadays routers are being built with more and more memory) 72 132 P
1898 -0.09 (. ) 534.09 132 P
1902 612 792 0 FMBEGINPAGE
1910 (OSPF protocol analysis) 249.36 712 T
1911 (July 1991) 493.02 712 T
1917 ([Page 12]) 493.7 73 T
1922 (Note that 10,000 routes is considerably lar) 72 676 T
1923 (ger than the lar) 275.31 676 T
1924 (gest \336eld implementation \050BARRNet; ) 347.37 676 T
1925 (which at 1816 external LSAs is still very lar) 72 662 T
1926 (ge\051.) 283.65 662 T
1927 (Note that there may be ways to reduce database size in a routing domain. First, the domain can ) 72 636 T
1928 -0.19 (make use of default routing, reducing the number of external routes that need to be imported. Sec-) 72 622 P
1930 (, an EGP can be used that will transport its own information through the AS instead of rely-) 98.54 608 T
1931 -0.21 (ing on the IGP \050OSPF in this case\051 to do transfer the information for it \050the EGP\051. Thirdly) 72 594 P
1932 -0.21 (, routers ) 498.11 594 P
1933 (having insuf) 72 580 T
1934 (\336cient memory may be able to be assigned to stub areas \050whose databases are drasti-) 131.41 580 T
1935 (cally smaller\051. Lastly) 72 566 T
1936 (, if the Internet went away from a \337at address space the amount of external ) 172.82 566 T
1937 (information imported into an OSPF domain could be reduced drastically) 72 552 T
1939 (While not as likely) 72 526 T
1940 (, there could be other issues that would limit the size of an OSPF routing ) 162.17 526 T
1941 (domain. If there are slow lines \050like 9600 baud\051, the size of the database will be limited \050see Sec-) 72 512 T
1942 (tion 3.2\051. Dijkstra may get to be expensive when there are hundreds of routers in the OSPF ) 72 498 T
1943 (domain; although at this point the domain can be split into areas. Finally) 72 484 T
1944 (, when there are many ) 418.69 484 T
1945 (routers attached to a single network, there may be undue burden imposed upon the Designated ) 72 470 T
1946 (Router; although at that point a LAN can be split into separate LANs.) 72 456 T
1950 612 792 0 FMBEGINPAGE
1958 (OSPF protocol analysis) 249.36 712 T
1959 (July 1991) 493.02 712 T
1965 ([Page 13]) 493.7 73 T
1971 (4.0 Suitable envir) 72 673.33 T
1972 (onments) 195.21 673.33 T
1974 -0.14 (Suitable environments for the OSPF protocol range from lar) 72 646 P
1975 -0.14 (ge to small. OSPF is particular suited ) 359.11 646 P
1976 (for transit Autonomous Systems for the following reasons. OSPF can accommodate a lar) 72 632 T
1977 (ge num-) 497.84 632 T
1978 (ber of external routes. In OSPF the import of external information is very \337exible, having provi-) 72 618 T
1979 -0.39 (sions for a forwarding address, two levels of external metrics, and the ability to tag external routes ) 72 604 P
1980 -0.29 (with their AS number for easy management. Also OSPF\325) 72 590 P
1981 -0.29 (s ability to do partial updates when exter-) 343.17 590 P
1982 (nal information changes is very useful on these networks.) 72 576 T
1983 (OSPF is also suited for smaller) 72 550 T
1984 (, either stand alone or stub Autonomous Systems, because of its ) 220.44 550 T
1985 (wide array of features: fast conver) 72 536 T
1986 (gence, equal-cost-multipath, T) 235.96 536 T
1987 (OS routing, areas, etc.) 382.3 536 T
1989 (5.0 Unsuitable envir) 72 469.33 T
1990 (onments) 212.98 469.33 T
1992 -0.22 (OSPF has a very limited ability to express policy) 72 442 P
1993 -0.22 (. Basically) 304.62 442 P
1994 -0.22 (, its only policy mechanisms are in the ) 354.25 442 P
1995 (establishment of a four level routing hierarchy: intra-area, inter) 72 428 T
1996 (-area, type 1 and type 2 external ) 374.52 428 T
1997 (routes. A system wanting more sophisticated policies would have to be split up into separate ) 72 414 T
1998 (ASes, running a policy-based EGP between them.) 72 400 T
2002 612 792 0 FMBEGINPAGE
2010 (OSPF protocol analysis) 249.36 712 T
2011 (July 1991) 493.02 712 T
2017 ([Page 14]) 493.7 73 T
2023 (6.0 Refer) 72 673.33 T
2024 (ence Documents) 137.87 673.33 T
2026 (The following documents have been referenced by this report:) 72 646 T
2029 (, J., \322Experience with the OSPF protocol\323, RFC 1246, July 1991.) 129.88 626 T
2032 (, J., \322OSPF V) 129.88 608 T
2033 (ersion 2\323, RFC 1247, July 1991.) 193.85 608 T
2035 (Corporation for National Research Initiatives, \322Proceedings of the Eighteenth Internet ) 108 590 T
2036 (Engineering T) 108 576 T
2037 (ask Force\323, University of British Columbia, July 30-August 3, 1990.) 176.11 576 T
2041 612 792 0 FMBEGINPAGE
2049 (OSPF protocol analysis) 249.36 712 T
2050 (July 1991) 493.02 712 T
2056 ([Page 15]) 493.7 73 T
2062 (Security Considerations) 72 673.33 T
2064 (Security issues are not discussed in this memo.) 72 646 T
2066 (Author) 72 617.33 T
2067 (\325) 122.04 617.33 T
2068 (s Addr) 126.77 617.33 T
2069 (ess) 173.13 617.33 T
2072 (Proteon Inc.) 72 576 T
2074 (echnology Drive) 87.48 562 T
2076 (estborough, MA 01581) 82.36 548 T
2077 (Phone: \050508\051 898-2800) 72 522 T
2078 (Email: jmoy@proteon.com) 72 508 T
2082 %%BoundingBox: 0 0 612 792
2084 %%DocumentFonts: Times-Roman