One document matched: draft-irtf-tmrg-metrics-11.ps

%!PS-Adobe-3.0
%%Creator: groff version 1.18.1
%%CreationDate: Fri Oct  5 21:00:57 2007
%%DocumentNeededResources: font Courier-Bold
%%+ font Times-Bold
%%+ font Times-Roman
%%DocumentSuppliedResources: procset grops 1.18 1
%%Pages: 19
%%PageOrder: Ascend
%%Orientation: Portrait
%%EndComments
%%BeginProlog
%%BeginResource: procset grops 1.18 1
/setpacking where{
pop
currentpacking
true setpacking
}if
/grops 120 dict dup begin
/SC 32 def
/A/show load def
/B{0 SC 3 -1 roll widthshow}bind def
/C{0 exch ashow}bind def
/D{0 exch 0 SC 5 2 roll awidthshow}bind def
/E{0 rmoveto show}bind def
/F{0 rmoveto 0 SC 3 -1 roll widthshow}bind def
/G{0 rmoveto 0 exch ashow}bind def
/H{0 rmoveto 0 exch 0 SC 5 2 roll awidthshow}bind def
/I{0 exch rmoveto show}bind def
/J{0 exch rmoveto 0 SC 3 -1 roll widthshow}bind def
/K{0 exch rmoveto 0 exch ashow}bind def
/L{0 exch rmoveto 0 exch 0 SC 5 2 roll awidthshow}bind def
/M{rmoveto show}bind def
/N{rmoveto 0 SC 3 -1 roll widthshow}bind def
/O{rmoveto 0 exch ashow}bind def
/P{rmoveto 0 exch 0 SC 5 2 roll awidthshow}bind def
/Q{moveto show}bind def
/R{moveto 0 SC 3 -1 roll widthshow}bind def
/S{moveto 0 exch ashow}bind def
/T{moveto 0 exch 0 SC 5 2 roll awidthshow}bind def
/SF{
findfont exch
[exch dup 0 exch 0 exch neg 0 0]makefont
dup setfont
[exch/setfont cvx]cvx bind def
}bind def
/MF{
findfont
[5 2 roll
0 3 1 roll
neg 0 0]makefont
dup setfont
[exch/setfont cvx]cvx bind def
}bind def
/level0 0 def
/RES 0 def
/PL 0 def
/LS 0 def
/MANUAL{
statusdict begin/manualfeed true store end
}bind def
/PLG{
gsave newpath clippath pathbbox grestore
exch pop add exch pop
}bind def
/BP{10 -52 translate
/level0 save def
1 setlinecap
1 setlinejoin
72 RES div dup scale
LS{
90 rotate
}{
0 PL translate
}ifelse
1 -1 scale
}bind def
/EP{
level0 restore
showpage
}bind def
/DA{
newpath arcn stroke
}bind def
/SN{
transform
.25 sub exch .25 sub exch
round .25 add exch round .25 add exch
itransform
}bind def
/DL{
SN
moveto
SN
lineto stroke
}bind def
/DC{
newpath 0 360 arc closepath
}bind def
/TM matrix def
/DE{
TM currentmatrix pop
translate scale newpath 0 0 .5 0 360 arc closepath
TM setmatrix
}bind def
/RC/rcurveto load def
/RL/rlineto load def
/ST/stroke load def
/MT/moveto load def
/CL/closepath load def
/Fr{
setrgbcolor fill
}bind def
/Fk{
setcmykcolor fill
}bind def
/Fg{
setgray fill
}bind def
/FL/fill load def
/LW/setlinewidth load def
/Cr/setrgbcolor load def
/Ck/setcmykcolor load def
/Cg/setgray load def
/RE{
findfont
dup maxlength 1 index/FontName known not{1 add}if dict begin
{
1 index/FID ne{def}{pop pop}ifelse
}forall
/Encoding exch def
dup/FontName exch def
currentdict end definefont pop
}bind def
/DEFS 0 def
/EBEGIN{
moveto
DEFS begin
}bind def
/EEND/end load def
/CNT 0 def
/level1 0 def
/PBEGIN{
/level1 save def
translate
div 3 1 roll div exch scale
neg exch neg exch translate
0 setgray
0 setlinecap
1 setlinewidth
0 setlinejoin
10 setmiterlimit
[]0 setdash
/setstrokeadjust where{
pop
false setstrokeadjust
}if
/setoverprint where{
pop
false setoverprint
}if
newpath
/CNT countdictstack def
userdict begin
/showpage{}def
}bind def
/PEND{
clear
countdictstack CNT sub{end}repeat
level1 restore
}bind def
end def
/setpacking where{
pop
setpacking
}if
%%EndResource
%%IncludeResource: font Courier-Bold
%%IncludeResource: font Times-Bold
%%IncludeResource: font Times-Roman
grops begin/DEFS 1 dict def DEFS begin/u{.001 mul}bind def end/RES 72
def/PL 841.89 def/LS false def/ENC0[/asciicircum/asciitilde/Scaron
/Zcaron/scaron/zcaron/Ydieresis/trademark/quotesingle/Euro/.notdef
/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef
/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef
/.notdef/.notdef/.notdef/space/exclam/quotedbl/numbersign/dollar/percent
/ampersand/quoteright/parenleft/parenright/asterisk/plus/comma/hyphen
/period/slash/zero/one/two/three/four/five/six/seven/eight/nine/colon
/semicolon/less/equal/greater/question/at/A/B/C/D/E/F/G/H/I/J/K/L/M/N/O
/P/Q/R/S/T/U/V/W/X/Y/Z/bracketleft/backslash/bracketright/circumflex
/underscore/quoteleft/a/b/c/d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/w/x/y
/z/braceleft/bar/braceright/tilde/.notdef/quotesinglbase/guillemotleft
/guillemotright/bullet/florin/fraction/perthousand/dagger/daggerdbl
/endash/emdash/ff/fi/fl/ffi/ffl/dotlessi/dotlessj/grave/hungarumlaut
/dotaccent/breve/caron/ring/ogonek/quotedblleft/quotedblright/oe/lslash
/quotedblbase/OE/Lslash/.notdef/exclamdown/cent/sterling/currency/yen
/brokenbar/section/dieresis/copyright/ordfeminine/guilsinglleft
/logicalnot/minus/registered/macron/degree/plusminus/twosuperior
/threesuperior/acute/mu/paragraph/periodcentered/cedilla/onesuperior
/ordmasculine/guilsinglright/onequarter/onehalf/threequarters
/questiondown/Agrave/Aacute/Acircumflex/Atilde/Adieresis/Aring/AE
/Ccedilla/Egrave/Eacute/Ecircumflex/Edieresis/Igrave/Iacute/Icircumflex
/Idieresis/Eth/Ntilde/Ograve/Oacute/Ocircumflex/Otilde/Odieresis
/multiply/Oslash/Ugrave/Uacute/Ucircumflex/Udieresis/Yacute/Thorn
/germandbls/agrave/aacute/acircumflex/atilde/adieresis/aring/ae/ccedilla
/egrave/eacute/ecircumflex/edieresis/igrave/iacute/icircumflex/idieresis
/eth/ntilde/ograve/oacute/ocircumflex/otilde/odieresis/divide/oslash
/ugrave/uacute/ucircumflex/udieresis/yacute/thorn/ydieresis]def
/Times-Roman@0 ENC0/Times-Roman RE/Times-Bold@0 ENC0/Times-Bold RE
/Courier-Bold@0 ENC0/Courier-Bold RE
%%EndProlog
%%Page: 1 1
%%BeginPageSetup
BP
%%EndPageSetup
/F0 10/Courier-Bold@0 SF(Internet Engineering Task Force)72 85 Q
(Sally Floyd)179.998 E 305.998(INTERNET-DRAFT Editor)72 98 R
(Intended status: Informational)72 111 Q 6(5O)167.998 G(ctober 2007)-6 E
(Expires: April 2008)72 124 Q/F1 14/Times-Bold@0 SF(Metrics f)102.304
149 Q(or the Ev)-.35 E(aluation of Congestion Contr)-.14 E
(ol Mechanisms)-.252 E(draft-irtf-tmr)203.328 162 Q(g-metrics-11.ps)-.14
E/F2 11/Times-Bold@0 SF(Status of this Memo)72 194 Q/F3 11/Times-Roman@0
SF(By submitting this Internet-Draft, each author represents that an)
100.8 210.6 Q 2.75(ya)-.165 G(pplicable patent or other)-2.75 E
(IPR claims of which he or she is a)100.8 223.6 Q -.11(wa)-.165 G(re ha)
.11 E .33 -.165(ve b)-.22 H(een or will be disclosed, and an).165 E 2.75
(yo)-.165 G 2.75(fw)-2.75 G(hich he)-2.75 E(or she becomes a)100.8 236.6
Q -.11(wa)-.165 G
(re will be disclosed, in accordance with Section 6 of BCP 79.).11 E
(Internet-Drafts are w)100.8 253.2 Q
(orking documents of the Internet Engineering T)-.11 E(ask F)-.88 E
(orce \(IETF\), its)-.165 E(areas, and its w)100.8 266.2 Q
(orking groups.)-.11 E(Note that other groups may also distrib)5.5 E
(ute w)-.22 E(orking)-.11 E(documents as Internet-Drafts.)100.8 279.2 Q
(Internet-Drafts are draft documents v)100.8 295.8 Q
(alid for a maximum of six months and may be)-.275 E
(updated, replaced, or obsoleted by other documents at an)100.8 308.8 Q
2.75(yt)-.165 G 2.75(ime. It)-2.75 F(is inappropriate to use)2.75 E
(Internet-Drafts as reference material or to cite them other than as "w)
100.8 321.8 Q(ork in progress.")-.11 E
(The list of current Internet-Drafts can be accessed at http://www)100.8
338.4 Q(.ietf.or)-.715 E(g/ietf/1id-)-.198 E(abstracts.txt.)100.8 351.4
Q(The list of Internet-Draft Shado)100.8 368 Q 2.75(wD)-.275 G
(irectories can be accessed at)-2.75 E(http://www)100.8 381 Q(.ietf.or)
-.715 E(g/shado)-.198 E -.715(w.)-.275 G(html.).715 E
(This Internet-Draft will e)100.8 397.6 Q(xpire on April 2008.)-.165 E
F2(Copyright Notice)72 423.6 Q F3(Cop)100.8 440.2 Q
(yright \(C\) The IETF T)-.11 E(rust \(2007\).)-.385 E F2(Abstract)72
466.2 Q F3(This document discusses the metrics to be considered in an e)
100.8 482.8 Q -.275(va)-.275 G(luation of ne).275 E 2.75(wo)-.275 G 2.75
(rm)-2.75 G(odi\214ed)-2.75 E
(congestion control mechanisms for the Internet.)100.8 495.8 Q
(These include metrics for the e)5.5 E -.275(va)-.275 G(luation).275 E
(of ne)100.8 508.8 Q 2.75(wt)-.275 G
(ransport protocols, of proposed modi\214cations to TCP)-2.75 E 2.75(,o)
-1.221 G 2.75(fa)-2.75 G(pplication-le)-2.75 E -.165(ve)-.275 G(l).165 E
(congestion control, and of Acti)100.8 521.8 Q .33 -.165(ve Q)-.275 H
(ueue Management \(A).165 E(QM\) mechanisms in the router)-.605 E(.)
-.605 E
(This document is the \214rst in a series of documents aimed at impro)
100.8 534.8 Q(ving the models that we)-.165 E(use in the e)100.8 547.8 Q
-.275(va)-.275 G(luation of transport protocols.).275 E
(This document is a product of the T)100.8 573.8 Q
(ransport Modeling Research Group \(TMRG\), and has)-.385 E(recei)100.8
586.8 Q -.165(ve)-.275 G 2.75(dd).165 G(etailed feedback from man)-2.75
E 2.75(ym)-.165 G(embers of the Research Group \(RG\).)-2.75 E(As the)
5.5 E(document tries to mak)100.8 599.8 Q 2.75(ec)-.11 G(lear)-2.75 E
2.75(,t)-.44 G(here is not necessarily a consensus within the research)
-2.75 E(community \(or the IETF community)100.8 612.8 Q 2.75(,t)-.715 G
(he v)-2.75 E(endor community)-.165 E 2.75(,t)-.715 G
(he operations community)-2.75 E 2.75(,o)-.715 G(r)-2.75 E(an)100.8
625.8 Q 2.75(yo)-.165 G(ther community\) about the metrics that congest\
ion control mechanisms should be)-2.75 E
(designed to optimize, in terms of tradeof)100.8 638.8 Q
(fs between throughput and delay)-.275 E 2.75(,f)-.715 G
(airness between)-2.86 E(competing \215o)100.8 651.8 Q(ws, and the lik)
-.275 E 2.75(e. Ho)-.11 F(we)-.275 E -.165(ve)-.275 G .88 -.44(r, w).165
H 2.75(eb).44 G(elie)-2.75 E .33 -.165(ve t)-.275 H
(hat there is a clear consensus that).165 E
(congestion control mechanisms should be e)100.8 664.8 Q -.275(va)-.275
G(luated in terms of tradeof).275 E(fs between a range)-.275 E
(of metrics, rather than in terms of optimizing for a single metric.)
100.8 677.8 Q(Flo)72 769 Q 366.891(yd [P)-.11 F(age 1])-.165 E 0 Cg EP
%%Page: 2 2
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E/F1 13/Times-Bold@0 SF -1.196
(Ta)253.526 85 S(ble of Contents)1.196 E/F2 10/Times-Roman@0 SF
(1. Introduction)100.8 110 Q F0 11(.......................)11.01 G F2(5)
11.5 E(2. Metrics)100.8 122 Q F0 11(.........................)2.4 G F2
(6)11.5 E(2.1. Throughput, Delay)115.8 134 Q 2.5(,a)-.65 G
(nd Loss Rates)-2.5 E F0 11(...............)6.95 G F2(6)11.5 E
(2.1.1. Throughput)130.8 146 Q F0 11(....................)9.48 G F2(7)
11.5 E(2.1.2. Delay)130.8 158 Q F0 11(......................)5.32 G F2
(7)11.5 E(2.1.3. P)130.8 170 Q(ack)-.15 E(et Loss Rates)-.1 E F0 11
(..................)11.68 G F2(8)11.5 E(2.2. Response T)115.8 182 Q
(imes and Minimizing Oscillations)-.35 E F0 11(............)6.77 G F2(8)
11.5 E(2.2.1. Response to Changes)130.8 194 Q F0 11(.................)
12.4 G F2(9)11.5 E(2.2.2. Minimizing Oscillations)130.8 206 Q F0 11
(.................).44 G F2(9)11.5 E(2.3. F)115.8 218 Q(airness and Con)
-.15 E -.15(ve)-.4 G -.18(rg).15 G(ence).18 E F0 11(..................)
2.05 G F2(10)6.5 E(2.3.1. Metrics for f)130.8 230 Q
(airness between \215o)-.1 E(ws)-.25 E F0 11(..............)1.25 G F2
(10)6.5 E(2.3.2. Metrics for f)130.8 242 Q(airness between \215o)-.1 E
(ws with dif)-.25 E(ferent resource requirements)-.25 E F0 11(...)4.46 G
F2(11)6.5 E(2.3.3. Comments on f)130.8 254 Q(airness)-.1 E F0 11
(.................)8.61 G F2(12)6.5 E(2.4. Rob)115.8 266 Q
(ustness for Challenging En)-.2 E(vironments)-.4 E F0 11(.............)
6.33 G F2(13)6.5 E(2.5. Rob)115.8 278 Q(ustness to F)-.2 E
(ailures and to Misbeha)-.15 E(ving Users)-.2 E F0 11(...........)5.73 G
F2(13)6.5 E(2.6. Deplo)115.8 290 Q(yability)-.1 E F0 11
(.....................)10.55 G F2(13)6.5 E
(2.7. Metrics for Speci\214c T)115.8 302 Q(ypes of T)-.8 E(ransport)-.35
E F0 11(..............)4.27 G F2(14)6.5 E(2.8. User)115.8 314 Q
(-Based Metrics)-.2 E F0 11(...................)13.45 G F2(14)6.5 E
(3. Metrics in the IP Performance Metrics \(IPPM\) W)100.8 326 Q
(orking Group)-.8 E F0 11(.........)1.57 G F2(14)6.5 E
(4. Comments on Methodology)100.8 338 Q F0 11(...................)3.22 G
F2(14)6.5 E(5. Security Considerations)100.8 350 Q F0 11
(....................)5.87 G F2(15)6.5 E(6. IAN)100.8 362 Q 2.5(AC)-.35
G(onsiderations)-2.5 E F0 11(.....................).81 G F2(15)6.5 E
(7. Ackno)100.8 374 Q(wledgements)-.25 E F0 11(.....................)
8.77 G F2(15)6.5 E(Informati)100.8 386 Q .3 -.15(ve R)-.25 H(eferences)
.15 E F0 11(.....................)4.22 G F2(15)6.5 E(Authors' Addresses)
100.8 398 Q F0 11(......................)5.05 G F2(18)6.5 E(Full Cop)
100.8 410 Q(yright Statement)-.1 E F0 11(....................)10.12 G F2
(18)6.5 E(Intellectual Property)100.8 422 Q F0 11
(......................)2.27 G F2(19)6.5 E F0(Flo)72 769 Q 366.891
(yd [P)-.11 F(age 2])-.165 E 0 Cg EP
%%Page: 3 3
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E .396 -.198(TO B)100.8 85 T
2.75(ED).198 G(ELETED BY THE RFC EDIT)-2.75 E(OR UPON PUBLICA)-.198 E
(TION:)-1.221 E(Changes from draft-irtf-tmr)100.8 111 Q
(g-metrics-10.txt:)-.198 E 2.75(*A)100.8 137 S
(dded tunnels \(e.g., IPsec\) and non-router queues)-2.75 E
(\(e.g. Ethernet switches\) to the discussion on deplo)106.3 150 Q
(yability)-.11 E(.)-.715 E(Feedback from Aaron F)106.3 163 Q
(alk in the IRSG poll.)-.165 E 2.75(*M)100.8 189 S
(inor editing in response to feedback from Lachlan Andre)-2.75 E -.715
(w.)-.275 G 2.75(*A)100.8 215 S(dded a ne)-2.75 E 2.75(wc)-.275 G
(itation to [DM06].)-2.75 E(Changes from draft-irtf-tmr)100.8 241 Q
(g-metrics-09.txt:)-.198 E 2.75(*C)100.8 267 S
(opied a paragraph from the Abstract to the Introduction, and)-2.75 E
(updated references.)106.3 280 Q(From feedback from T)5.5 E(on)-.88 E
2.75(yL)-.165 G 2.75(ii)-2.75 G 2.75(nt)-2.75 G(he IRSG)-2.75 E(re)106.3
293 Q(vie)-.275 E -.715(w.)-.275 G(Changes from draft-irtf-tmr)100.8 319
Q(g-metrics-08.txt:)-.198 E 2.75(*G)100.8 345 S
(eneral feedback from Michael W)-2.75 E(elzl, adding e)-.88 E
(xplanations)-.165 E(to some of the f)106.3 358 Q(airness discussions.)
-.11 E(Changes from draft-irtf-tmr)100.8 384 Q(g-metrics-07.txt:)-.198 E
2.75(*G)100.8 410 S(eneral feedback from Mark Allman.)-2.75 E 2.75(*F)
100.8 436 S(eedback and contrib)-2.75 E(utions from Lachlan Andre)-.22 E
2.75(wa)-.275 G(bout f)-2.75 E(airness.)-.11 E
(Changes from draft-irtf-tmr)100.8 462 Q(g-metrics-06.txt:)-.198 E 2.75
(*A)100.8 488 S(dded a paragraph about tradeof)-2.75 E(fs between)-.275
E -.11(fa)106.3 501 S(irness and throughput.).11 E 2.75(*M)100.8 527 S
(inor editing changes.)-2.75 E(Changes from draft-irtf-tmr)100.8 553 Q
(g-metrics-05.txt:)-.198 E 2.75(*A)100.8 579 S
(dded a sentence about drop synchronization.)-2.75 E(Feedback from Da)
106.3 592 Q(vid Ros.)-.22 E 2.75(*A)100.8 618 S(dded a citation to f)
-2.75 E(airness draft by Briscoe.)-.11 E(Changes from draft-irtf-tmr)
100.8 644 Q(g-metrics-04.txt:)-.198 E 2.75(*A)100.8 670 S(dded te)-2.75
E(xt to the abstract.)-.165 E(Changes from draft-irtf-tmr)100.8 696 Q
(g-metrics-03.txt:)-.198 E 2.75(*A)100.8 722 S
(dded a paragraph about sudden changes due to mobility)-2.75 E(Flo)72
769 Q 366.891(yd [P)-.11 F(age 3])-.165 E 0 Cg EP
%%Page: 4 4
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E
(and the heterogeneity of wireless access types.)106.3 85 Q
(Suggestion from Andras V)106.3 98 Q(eres.)-1.221 E 2.75(*A)100.8 124 S
(dd co)-2.75 E -.275(va)-.165 G
(riance as one of the metrics for oscillations.).275 E
(Suggestion from Sa)106.3 137 Q -.165(ve)-.22 G
(rio Mascolo, original te).165 E(xt)-.165 E(contrib)106.3 150 Q
(ution from Injong Rhee.)-.22 E(Changes from draft-irtf-tmr)100.8 176 Q
(g-metrics-02.txt:)-.198 E 2.75(*A)100.8 202 S(dded a fe)-2.75 E 2.75
(ws)-.275 G(entences to the Abstract about the)-2.75 E
(status of the document.)106.3 215 Q(Changes from draft-irtf-tmr)100.8
241 Q(g-metrics-01.txt:)-.198 E 2.75(*A)100.8 267 S
(dded a discussion about the metrics in IPPM.)-2.75 E
(Changes from draft-irtf-tmr)100.8 293 Q(g-metrics-01c.txt:)-.198 E 2.75
(*A)100.8 319 S(dded to the discussion of netw)-2.75 E(ork-based, \215o)
-.11 E(w-based,)-.275 E(and user)106.3 332 Q
(-based metrics, based on email from Dado Colussi,)-.22 E
(Sean Moore, Damon W)106.3 345 Q(ischik, Dah Ming Chiu, and others.)-.44
E 2.75(*C)100.8 371 S(hanged "pack)-2.75 E(et drop rate" to "pack)-.11 E
(et loss rate".)-.11 E(Suggestion from Nelson F)106.3 384 Q(onseca.)
-.165 E 2.75(*A)100.8 410 S
(dded a discussion of the Colussi et al. paper on a ne)-2.75 E(w)-.275 E
(de\214nition of f)106.3 423 Q(airness.)-.11 E 2.75(*A)100.8 449 S
(dded a discussion of the Chiu and T)-2.75 E(an paper on rede\214ning)
-.88 E -.11(fa)106.3 462 S(irness for inelastic traf).11 E(\214c.)-.275
E(Changes from draft-irtf-tmr)100.8 488 Q(g-metrics-01b)-.198 E(.txt:)
-.44 E 2.75(*A)100.8 514 S(dded a discussion of goodput vs. throughput.)
-2.75 E(Suggestion from Nelson F)106.3 527 Q(onseca.)-.165 E
(Changes from draft-irtf-tmr)100.8 553 Q(g-metrics-01a.txt:)-.198 E 2.75
(*A)100.8 579 S(dded to the discussion of pack)-2.75 E
(et drop rate metrics.)-.11 E(Suggestions from Janardhan Iyeng)106.3 592
Q(ar)-.055 E 2.75(,S)-.44 G(ean Moore,)-2.75 E
(Armando Caro, and Nelson F)106.3 605 Q(onseca.)-.165 E 2.75(*A)100.8
631 S(dded a sentence about throughput used as a metric for)-2.75 E
(transfer times for v)106.3 644 Q(ery short \215o)-.165 E(ws.)-.275 E
(Response to email from Sean Moore.)106.3 657 Q
(Changes from draft-irtf-tmr)100.8 683 Q(g-metrics-00.txt:)-.198 E 2.75
(*A)100.8 709 S(dded a list of rele)-2.75 E -.275(va)-.275 G
(nt congestion control mechanisms to).275 E(the abstract.)106.3 722 Q
(Suggestion from Sean Moore.)5.5 E(Flo)72 769 Q 366.891(yd [P)-.11 F
(age 4])-.165 E 0 Cg EP
%%Page: 5 5
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E 2.75(*A)100.8 85 S
(dded to the Introduction. Suggestion from Dado Colussi.)-2.75 E 2.75
(*A)100.8 111 S
(dded a sentence about jitter to the discussion of minimizing)-2.75 E
2.75(oscillations. Suggestion)106.3 124 R(from W)2.75 E(esle)-.88 E 2.75
(yE)-.165 G(ddy)-2.75 E(.)-.715 E 2.75(*A)100.8 150 S
(dded a note about con)-2.75 E -.165(ve)-.44 G -.198(rg).165 G
(ence between e).198 E(xisting \215o)-.165 E(ws after)-.275 E 2.75(ac)
106.3 163 S(hange in bandwidth.)-2.75 E(Suggestion from W)5.5 E(esle)
-.88 E 2.75(yE)-.165 G(ddy)-2.75 E(.)-.715 E 2.75(*A)100.8 189 S
(dded to the section on deplo)-2.75 E(yability)-.11 E 5.5(.S)-.715 G
(uggestion from)-5.5 E -.88(We)106.3 202 S(sle).88 E 2.75(yE)-.165 G
(ddy)-2.75 E(.)-.715 E(Changes from draft-\215o)100.8 228 Q
(yd-transport-metrics-00.txt:)-.11 E 2.75(*A)100.8 254 S
(dded metrics for:)-2.75 E 2.75(-r)106.3 267 S(ob)-2.75 E
(ustness in challenging en)-.22 E(vironments,)-.44 E 2.75(-d)106.3 280 S
(eplo)-2.75 E(yability)-.11 E(,)-.715 E 2.75(-r)106.3 293 S(ob)-2.75 E
(ustness to f)-.22 E(ailures and to misbeha)-.11 E(ving users)-.22 E
2.75(*A)100.8 319 S(dded a discussion of f)-2.75 E(airness and pack)-.11
E(et size.)-.11 E/F1 11/Times-Bold@0 SF(1.)72 358 Q/F2 14/Times-Bold@0
SF(Intr)5.5 E(oduction)-.252 E F0(As a step to)100.8 374.6 Q -.11(wa)
-.275 G(rds impro).11 E(ving our methodologies for e)-.165 E -.275(va)
-.275 G(luating congestion control).275 E(mechanisms, in this document \
we discuss some of the metrics to be considered.)100.8 387.6 Q 1.76 -.88
(We a)5.5 H(lso).88 E
(consider the relationship between metrics, e.g., the well-kno)100.8
400.6 Q(wn tradeof)-.275 E 2.75(fb)-.275 G(etween)-2.75 E
(throughput and delay)100.8 413.6 Q 5.5(.T)-.715 G(his document doesn')
-5.5 E 2.75(ta)-.198 G(ttempt to specify e)-2.75 E -.165(ve)-.275 G
(ry metric that a study).165 E(might consider; for e)100.8 426.6 Q
(xample, there are domain-speci\214c metrics that researchers might)
-.165 E(consider that are o)100.8 439.6 Q -.165(ve)-.165 G 2.75(ra).165
G(nd abo)-2.75 E .33 -.165(ve t)-.165 H
(he metrics laid out in this document.).165 E 1.76 -.88(We c)100.8 465.6
T(onsider metrics for aggre).88 E -.055(ga)-.165 G(te traf).055 E
(\214c \(taking into account the ef)-.275 E(fect of \215o)-.275 E(ws on)
-.275 E(competing traf)100.8 478.6 Q(\214c in the netw)-.275 E
(ork\) as well as the heterogeneous goals of dif)-.11 E(ferent)-.275 E
(applications or transport protocols \(e.g., of high throughput for b)
100.8 491.6 Q(ulk data transfer)-.22 E 2.75(,a)-.44 G(nd of)-2.75 E(lo)
100.8 504.6 Q 2.75(wd)-.275 G(elay for interacti)-2.75 E .33 -.165(ve v)
-.275 H(oice or video\).)-.055 E(Dif)5.5 E
(ferent transport protocols or A)-.275 E(QM)-.605 E(mechanisms might ha)
100.8 517.6 Q .33 -.165(ve g)-.22 H(oals of optimizing dif).165 E
(ferent sets of metrics, with one transport)-.275 E
(protocol optimized for per)100.8 530.6 Q(-\215o)-.22 E 2.75(wt)-.275 G
(hroughput and another optimized for rob)-2.75 E(ustness o)-.22 E -.165
(ve)-.165 G(r).165 E(wireless links, and with dif)100.8 543.6 Q
(ferent de)-.275 E(grees of attention to f)-.165 E
(airness with competing traf)-.11 E(\214c.)-.275 E 1.76 -.88(We h)100.8
556.6 T(ope this document will be used as a step in e).88 E -.275(va)
-.275 G(luating proposed congestion control).275 E
(mechanisms for a wide range of metrics, for e)100.8 569.6 Q
(xample, noting that Mechanism X is good at)-.165 E
(optimizing Metric A, b)100.8 582.6 Q
(ut pays the price with poor performance for Metric B.)-.22 E(The goal)
5.5 E -.11(wo)100.8 595.6 S(uld be to ha).11 E .33 -.165(ve a b)-.22 H
(road vie).165 E 2.75(wo)-.275 G 2.75(fb)-2.75 G
(oth the strengths and weaknesses of ne)-2.75 E(wly-proposed)-.275 E
(congestion control mechanisms.)100.8 608.6 Q(Subsequent documents are \
planned to present sets of simulation and testbed scenarios for)100.8
634.6 Q(the e)100.8 647.6 Q -.275(va)-.275 G(luation of transport proto\
cols and of congestion control mechanisms, based on the).275 E
(best current practice of the research community)100.8 660.6 Q 5.5(.T)
-.715 G(hese are not intended to be complete or)-5.5 E
(\214nal benchmark test suites, b)100.8 673.6 Q
(ut simply to be one step of man)-.22 E 2.75(yt)-.165 G 2.75(ob)-2.75 G
2.75(eu)-2.75 G(sed by researchers in)-2.75 E -.275(eva)100.8 686.6 S
(luating congestion control mechanisms.).275 E
(Subsequent documents are also planned on the)5.5 E
(methodologies in using these sets of scenarios.)100.8 699.6 Q
(This document is a product of the T)100.8 725.6 Q
(ransport Modeling Research Group \(TMRG\), and has)-.385 E(Flo)72 769 Q
317.391(yd Section)-.11 F 2.75(1. [P)2.75 F(age 5])-.165 E 0 Cg EP
%%Page: 6 6
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E(recei)100.8 85 Q -.165(ve)
-.275 G 2.75(dd).165 G(etailed feedback from man)-2.75 E 2.75(ym)-.165 G
(embers of the Research Group \(RG\).)-2.75 E(As the)5.5 E
(document tries to mak)100.8 98 Q 2.75(ec)-.11 G(lear)-2.75 E 2.75(,t)
-.44 G(here is not necessarily a consensus within the research)-2.75 E
(community \(or the IETF community)100.8 111 Q 2.75(,t)-.715 G(he v)
-2.75 E(endor community)-.165 E 2.75(,t)-.715 G(he operations community)
-2.75 E 2.75(,o)-.715 G(r)-2.75 E(an)100.8 124 Q 2.75(yo)-.165 G(ther c\
ommunity\) about the metrics that congestion control mechanisms should \
be)-2.75 E(designed to optimize, in terms of tradeof)100.8 137 Q
(fs between throughput and delay)-.275 E 2.75(,f)-.715 G
(airness between)-2.86 E(competing \215o)100.8 150 Q(ws, and the lik)
-.275 E 2.75(e. Ho)-.11 F(we)-.275 E -.165(ve)-.275 G .88 -.44(r, w).165
H 2.75(eb).44 G(elie)-2.75 E .33 -.165(ve t)-.275 H
(hat there is a clear consensus that).165 E
(congestion control mechanisms should be e)100.8 163 Q -.275(va)-.275 G
(luated in terms of tradeof).275 E(fs between a range)-.275 E
(of metrics, rather than in terms of optimizing for a single metric.)
100.8 176 Q/F1 11/Times-Bold@0 SF(2.)72 202 Q/F2 14/Times-Bold@0 SF
(Metrics)5.5 E F0(The metrics that we discuss are the follo)100.8 218.6
Q(wing:)-.275 E 12.65<8354>100.8 235.2 S(hroughput;)-12.65 E 12.65<8344>
100.8 251.8 S(elay;)-12.65 E 12.65<8350>100.8 268.4 S(ack)-12.815 E
(et loss rates;)-.11 E 12.65<8352>100.8 285 S
(esponse to sudden changes or to transient e)-12.65 E -.165(ve)-.275 G
(nts;).165 E 12.65<834d>100.8 301.6 S
(inimizing oscillations in throughput or in delay;)-12.65 E 12.65<8346>
100.8 318.2 S(airness and con)-12.815 E -.165(ve)-.44 G -.198(rg).165 G
(ence times;).198 E 12.65<8352>100.8 334.8 S(ob)-12.65 E
(ustness for challenging en)-.22 E(vironments;)-.44 E 12.65<8352>100.8
351.4 S(ob)-12.65 E(ustness to f)-.22 E(ailures and to misbeha)-.11 E
(ving users;)-.22 E 12.65<8344>100.8 368 S(eplo)-12.65 E(yability;)-.11
E 12.65<834d>100.8 384.6 S(etrics for speci\214c types of transport;)
-12.65 E 12.65<8355>100.8 401.2 S(ser)-12.65 E(-based metrics.)-.22 E
1.76 -.88(We c)100.8 417.8 T(onsider each of these belo).88 E 4.18 -.715
(w. M)-.275 H(an).715 E 2.75(yo)-.165 G 2.75(ft)-2.75 G(he metrics ha)
-2.75 E .33 -.165(ve n)-.22 H(etw).165 E(ork-based, \215o)-.11 E
(w-based,)-.275 E(and user)100.8 430.8 Q(-based interpretations.)-.22 E
-.165(Fo)5.5 G 2.75(re).165 G(xample, netw)-2.915 E
(ork-based metrics can consider)-.11 E(aggre)100.8 443.8 Q -.055(ga)
-.165 G(te bandwidth and aggre).055 E -.055(ga)-.165 G
(te drop rates, \215o).055 E(w-based metrics can consider end-to-end)
-.275 E
(transfer times for \214le transfers or end-to-end delay and pack)100.8
456.8 Q(et drop rates for interacti)-.11 E -.165(ve)-.275 G(traf)100.8
469.8 Q(\214c, and user)-.275 E(-based metrics can consider user w)-.22
E(ait time or user satisf)-.11 E(action with the)-.11 E(multimedia e)
100.8 482.8 Q 2.75(xperience. Our)-.165 F
(main goal in this document is to e)2.75 E
(xplain the set of metrics that)-.165 E(can be rele)100.8 495.8 Q -.275
(va)-.275 G(nt, and not to le).275 E
(gislate on the most appropriate methodology for using each)-.165 E
(general metric.)100.8 508.8 Q -.165(Fo)100.8 525.4 S 2.75(rs).165 G
(ome of the metrics, such as f)-2.75 E
(airness, there is not a clear agreement in the netw)-.11 E(ork)-.11 E
(community about the desired goals.)100.8 538.4 Q
(In these cases, the document attempts to present the)5.5 E
(range of approaches.)100.8 551.4 Q F1(2.1.)72 577.4 Q/F3 13
/Times-Bold@0 SF(Thr)5.5 E(oughput, Delay)-.234 E 3.25(,a)-.715 G
(nd Loss Rates)-3.25 E F0(Because of the clear tradeof)100.8 594 Q
(fs between throughput, delay)-.275 E 2.75(,a)-.715 G
(nd loss rates, it can be useful to)-2.75 E
(consider these three metrics together)100.8 607 Q 5.5(.T)-.605 G
(he tradeof)-5.5 E(fs are most clear in terms of queue)-.275 E
(management at the router; is the queue con\214gured to maximize aggre)
100.8 620 Q -.055(ga)-.165 G(te throughput, to).055 E
(minimize delay and loss rates, or some compromise between the tw)100.8
633 Q 2.75(o? An)-.11 F(alternati)2.75 E -.165(ve)-.275 G -.11(wo)100.8
646 S(uld be to consider a separate metric such as po).11 E(wer)-.275 E
2.75(,d)-.44 G(e\214ned in this conte)-2.75 E(xt as throughput)-.165 E
-.165(ove)100.8 659 S 2.75(rd).165 G(elay)-2.75 E 2.75(,t)-.715 G
(hat combines throughput and delay)-2.75 E 5.5(.H)-.715 G -.275(ow)-5.5
G -2.365 -.275(ev e).275 H .88 -.44(r, w).275 H 2.75(ed).44 G 2.75(on)
-2.75 G(ot propose in this)-2.75 E(document a clear tar)100.8 672 Q
(get in terms of the tradeof)-.198 E
(fs between throughput and delay; we are)-.275 E
(simply proposing that the e)100.8 685 Q -.275(va)-.275 G
(luation of transport protocols include an e).275 E(xploration of the)
-.165 E(competing metrics.)100.8 698 Q(Using \215o)100.8 724 Q
(w-based metrics instead of router)-.275 E
(-based metrics, the relationship between per)-.22 E(-\215o)-.22 E(w)
-.275 E(Flo)72 769 Q 309.141(yd Section)-.11 F 2.75(2.1. [P)2.75 F
(age 6])-.165 E 0 Cg EP
%%Page: 7 7
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E(throughput, delay)100.8 85 Q
2.75(,a)-.715 G(nd loss rates can often be gi)-2.75 E -.165(ve)-.275 G
2.75(nb).165 G 2.75(yt)-2.75 G(he transport protocol itself.)-2.75 E
-.165(Fo)5.5 G(r).165 E -.165(ex)100.8 98 S(ample, in TCP).165 E 2.75
(,t)-1.221 G(he sending rate s in pack)-2.75 E(ets per second is gi)-.11
E -.165(ve)-.275 G 2.75(na).165 G(s)-2.75 E 2.75(s=1)109.05 124 S
(.2/\(R)-2.75 E(TT*sqrt\(p\)\),)-.66 E(for the round-trip time R)100.8
150 Q(TT and loss rate p [FF99].)-.66 E/F1 11/Times-Bold@0 SF 2.75
(2.1.1. Thr)72 176 R(oughput)-.198 E F0
(Throughput can be measured as a router)100.8 192.6 Q
(-based metric of aggre)-.22 E -.055(ga)-.165 G
(te link utilization, as a).055 E<8d6f>100.8 205.6 Q
(w-based metric of per)-.275 E(-connection transfer times, and as user)
-.22 E(-based metrics of utility)-.22 E(functions or user w)100.8 218.6
Q(ait times.)-.11 E
(It is a clear goal of most congestion control mechanisms to)5.5 E(maxi\
mize throughput, subject to application demand and to the constraints o\
f the other)100.8 231.6 Q(metrics.)100.8 244.6 Q(Throughput is sometime\
s distinguished from goodput, where throughput is the link)100.8 270.6 Q
(utilization or \215o)100.8 283.6 Q 2.75(wr)-.275 G
(ate in bytes per second, and goodput, also measured in bytes per)-2.75
E(second, is the subset of throughput consisting of useful traf)100.8
296.6 Q 2.75(\214c. That)-.275 F(is, `goodput' e)2.75 E(xcludes)-.165 E
(duplicate pack)100.8 309.6 Q(ets, pack)-.11 E
(ets that will be dropped do)-.11 E(wnstream, pack)-.275 E
(et fragments or A)-.11 E(TM)-1.221 E
(cells that are dropped at the recei)100.8 322.6 Q -.165(ve)-.275 G 2.75
(rb).165 G(ecause the)-2.75 E 2.75(yc)-.165 G(an')-2.75 E 2.75(tb)-.198
G 2.75(er)-2.75 G(e-assembled into complete)-2.75 E(pack)100.8 335.6 Q
(ets, and the lik)-.11 E 2.75(e. In)-.11 F
(general, this document doesn')2.75 E 2.75(td)-.198 G
(istinguish between throughput)-2.75 E
(and goodput, and uses the general term "throughput".)100.8 348.6 Q 1.76
-.88(We n)100.8 374.6 T
(ote that maximizing throughput is of concern in a wide range of en).88
E(vironments, from)-.44 E(highly-congested netw)100.8 387.6 Q
(orks to under)-.11 E(-utilized ones, and from long-li)-.22 E -.165(ve)
-.275 G 2.75<648d>.165 G -.275(ow)-2.75 G 2.75(st).275 G 2.75(ov)-2.75 G
(ery short)-2.915 E 2.75(ones. As)100.8 400.6 R(an e)2.75 E
(xample, throughput has been used as one of the metrics for e)-.165 E
-.275(va)-.275 G(luating Quick-).275 E(Start, a proposal to allo)100.8
413.6 Q 2.75<778d>-.275 G -.275(ow)-2.75 G 2.75(st).275 G 2.75(os)-2.75
G(tart-up f)-2.75 E(aster than slo)-.11 E
(w-start, where throughput has been)-.275 E -.275(eva)100.8 426.6 S(lua\
ted in terms of the transfer times for connections with a range of tran\
sfer sizes).275 E([RFC4782] [SAF06].)100.8 439.6 Q(In some conte)100.8
465.6 Q(xts, it might be suf)-.165 E(\214cient to consider the aggre)
-.275 E -.055(ga)-.165 G(te throughput or the mean).055 E(per)100.8
478.6 Q(-\215o)-.22 E 2.75(wt)-.275 G
(hroughput [DM06], while in other conte)-2.75 E
(xts it might be necessary to consider the)-.165 E(distrib)100.8 491.6 Q
(ution of per)-.22 E(-\215o)-.22 E 2.75(wt)-.275 G 2.75(hroughput. Some)
-2.75 F(researchers e)2.75 E -.275(va)-.275 G
(luate transport protocols in terms).275 E(of maximizing the aggre)100.8
504.6 Q -.055(ga)-.165 G(te user utility).055 E 2.75(,w)-.715 G
(here a user')-2.75 E 2.75(su)-.605 G
(tility is generally de\214ned as a)-2.75 E(function of the user')100.8
517.6 Q 2.75(st)-.605 G(hroughput [KMT98].)-2.75 E(Indi)100.8 543.6 Q
(vidual applications can ha)-.275 E .33 -.165(ve a)-.22 H
(pplication-speci\214c needs in terms of throughput.).165 E -.165(Fo)5.5
G(r).165 E -.165(ex)100.8 556.6 S(ample, real-time video traf).165 E
(\214c can ha)-.275 E .33 -.165(ve h)-.22 H(ighly v).165 E
(ariable bandwidth demands;)-.275 E -1.419(Vo)5.5 G(IP traf)1.419 E
<8c63>-.275 E(is sensiti)100.8 569.6 Q .33 -.165(ve t)-.275 H 2.75(ot)
.165 G(he amount of bandwidth recei)-2.75 E -.165(ve)-.275 G 2.75(di)
.165 G(mmediately after idle periods.)-2.75 E(Thus, user)5.5 E
(metrics for throughput can be more comple)100.8 582.6 Q 2.75(xt)-.165 G
(han simply the per)-2.75 E(-connection transfer time.)-.22 E F1 2.75
(2.1.2. Delay)72 608.6 R F0(Lik)100.8 625.2 Q 2.75(et)-.11 G
(hroughput, delay can be measured as a router)-2.75 E
(-based metric of queueing delay o)-.22 E -.165(ve)-.165 G(r).165 E
(time, or as a \215o)100.8 638.2 Q(w-based metric in terms of per)-.275
E(-pack)-.22 E(et transfer times.)-.11 E(Per)5.5 E(-pack)-.22 E
(et delay can)-.11 E(also include delay at the sender w)100.8 651.2 Q
(aiting for the transport protocol to send the pack)-.11 E 2.75(et. F)
-.11 F(or)-.165 E(reliable transfer)100.8 664.2 Q 2.75(,t)-.44 G(he per)
-2.75 E(-pack)-.22 E
(et transfer time seen by the application includes the possible)-.11 E
(delay of retransmitting a lost pack)100.8 677.2 Q(et.)-.11 E
(Users of b)100.8 703.2 Q
(ulk data transfer applications might care about per)-.22 E(-pack)-.22 E
(et transfer times only in)-.11 E(so f)100.8 716.2 Q(ar as the)-.11 E
2.75(ya)-.165 G -.275(ff)-2.75 G(ect the per).275 E
(-connection transfer time.)-.22 E
(On the other end of the spectrum, for)5.5 E(Flo)72 769 Q 300.891
(yd Section)-.11 F 2.75(2.1.2. [P)2.75 F(age 7])-.165 E 0 Cg EP
%%Page: 8 8
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E
(users of streaming media, per)100.8 85 Q(-pack)-.22 E
(et delay can be a signi\214cant concern.)-.11 E(Note that in some)5.5 E
(cases the a)100.8 98 Q -.165(ve)-.22 G(rage delay might not capture th\
e metric of interest to the users; for e).165 E(xample,)-.165 E
(some users might care about the w)100.8 111 Q(orst-case delay)-.11 E
2.75(,o)-.715 G 2.75(ra)-2.75 G(bout the tail of the delay distrib)-2.75
E(ution.)-.22 E
(Note that queueing delay at a router is shared by all \215o)100.8 137 Q
(ws at a FIFO \(First-In First-Out\))-.275 E 2.75(queue. Thus,)100.8 150
R(the router)2.75 E(-based queueing delay induced by b)-.22 E
(ulk data transfers can be)-.22 E(important e)100.8 163 Q -.165(ve)-.275
G 2.75(ni).165 G 2.75(ft)-2.75 G(he b)-2.75 E
(ulk data transfers themselv)-.22 E(es are not concerned with per)-.165
E(-pack)-.22 E(et)-.11 E(transfer times.)100.8 176 Q/F1 11/Times-Bold@0
SF 2.75(2.1.3. P)72 202 R(ack)-.11 E(et Loss Rates)-.11 E F0 -.165(Pa)
100.8 218.6 S(ck).165 E(et loss rates can be measured as a netw)-.11 E
(ork-based or as a \215o)-.11 E(w-based metric.)-.275 E(When e)100.8
244.6 Q -.275(va)-.275 G(luating the ef).275 E(fect of pack)-.275 E
(et losses or ECN marks \(Explicit Congestion)-.11 E(Noti\214cation\) [\
RFC3168] on the performance of a congestion control mechanism for an)
100.8 257.6 Q(indi)100.8 270.6 Q(vidual \215o)-.275 E 1.43 -.715(w, r)
-.275 H(esearchers often use both the pack).715 E
(et loss/mark rate for that connection,)-.11 E(and the congestion e)
100.8 283.6 Q -.165(ve)-.275 G(nt rate \(also called the loss e).165 E
-.165(ve)-.275 G(nt rate\), where a congestion e).165 E -.165(ve)-.275 G
(nt or).165 E(loss e)100.8 296.6 Q -.165(ve)-.275 G
(nt consists of one or more lost or mark).165 E(ed pack)-.11 E
(ets in one round-trip time)-.11 E([RFC3448].)100.8 309.6 Q
(Some users might care about the pack)100.8 335.6 Q
(et loss rate only in so f)-.11 E(ar as it af)-.11 E(fects per)-.275 E
(-connection)-.22 E
(transfer times, while other users might care about pack)100.8 348.6 Q
(et loss rates directly)-.11 E 5.5(.R)-.715 G(FC 3611,)-5.5 E -.66(RT)
100.8 361.6 S 2.75(PC).66 G
(ontrol Protocol Extended Reports, describes a V)-2.75 E
(oIP performance-reporting standard)-1.419 E(called R)100.8 374.6 Q
(TCP XR, which includes a set of b)-.66 E(urst metrics.)-.22 E
(In RFC 3611, a b)5.5 E(urst is de\214ned as)-.22 E
(the maximal sequence starting and ending with a lost pack)100.8 387.6 Q
(et, and not including a sequence)-.11 E(of Gmin or more pack)100.8
400.6 Q(ets that are not lost [RFC3611].)-.11 E(The b)5.5 E
(urst metrics in RFC 3611)-.22 E(consist of the b)100.8 413.6 Q
(urst density \(the fraction of pack)-.22 E(ets in b)-.11 E(ursts\), g)
-.22 E(ap density \(the fraction of)-.055 E(pack)100.8 426.6 Q
(ets in the g)-.11 E(aps between b)-.055 E(ursts\), b)-.22 E
(urst duration \(the mean duration of b)-.22 E(ursts in)-.22 E
(seconds\), and g)100.8 439.6 Q(ap duration \(the mean duration of g)
-.055 E(aps in seconds\).)-.055 E(RFC 3357 deri)5.5 E -.165(ve)-.275 G
(s).165 E(metrics for "loss distance" and "loss period", along with sta\
tistics that capture loss patterns)100.8 452.6 Q -.165(ex)100.8 465.6 S
(perienced by pack).165 E(et streams on the Internet [RFC3357].)-.11 E
(In some cases it is useful to distinguish between pack)100.8 491.6 Q
(ets dropped at routers due to)-.11 E(congestion, and pack)100.8 504.6 Q
(ets lost in the netw)-.11 E(ork due to corruption.)-.11 E(One netw)
100.8 530.6 Q(ork-related reason to a)-.11 E -.22(vo)-.22 G
(id high steady-state pack).22 E(et loss rates is to a)-.11 E -.22(vo)
-.22 G(id).22 E(congestion collapse in en)100.8 543.6 Q
(vironments containing paths with multiple congested links.)-.44 E(In)
5.5 E(such en)100.8 556.6 Q(vironments, high pack)-.44 E
(et loss rates could result in congested links w)-.11 E(asting scarce)
-.11 E(bandwidth by carrying pack)100.8 569.6 Q
(ets that will only be dropped do)-.11 E(wnstream, before being)-.275 E
(deli)100.8 582.6 Q -.165(ve)-.275 G(red to the recei).165 E -.165(ve)
-.275 G 2.75(r[).165 G 2.75(RFC2914]. W)-2.75 F 2.75(ea)-.88 G
(lso note that in some cases the retransmit rate)-2.75 E
(can be high, and the goodput correspondingly lo)100.8 595.6 Q 1.43
-.715(w, e)-.275 H -.165(ve).44 G 2.75(nw).165 G(ith a lo)-2.75 E 2.75
(wp)-.275 G(ack)-2.75 E(et drop rate)-.11 E([AEO03].)100.8 608.6 Q F1
(2.2.)72 634.6 Q/F2 13/Times-Bold@0 SF(Response T)5.5 E
(imes and Minimizing Oscillations)-.234 E F0
(In this section we consider response times and oscillations together)
100.8 651.2 Q 2.75(,a)-.44 G 2.75(st)-2.75 G(here are well-)-2.75 E(kno)
100.8 664.2 Q(wn tradeof)-.275 E(fs between impro)-.275 E
(ving response times and minimizing oscillations.)-.165 E(In)5.5 E(addi\
tion, the scenarios that illustrate the dangers of poor response times \
are often quite)100.8 677.2 Q(dif)100.8 690.2 Q(ferent from the scenari\
os that illustrate the dangers of unnecessary oscillations.)-.275 E(Flo)
72 769 Q 309.141(yd Section)-.11 F 2.75(2.2. [P)2.75 F(age 8])-.165 E 0
Cg EP
%%Page: 9 9
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E/F1 11/Times-Bold@0 SF 2.75
(2.2.1. Response)72 85 R(to Changes)2.75 E F0(One of the k)100.8 101.6 Q
.33 -.165(ey c)-.11 H
(oncerns in the design of congestion control mechanisms has been the)
.165 E(response times to sudden congestion in the netw)100.8 114.6 Q
2.75(ork. On)-.11 F(the one hand, congestion control)2.75 E(mechanisms \
should respond reasonably promptly to sudden congestion from routing or)
100.8 127.6 Q(bandwidth changes, or from a b)100.8 140.6 Q
(urst of competing traf)-.22 E 2.75(\214c. At)-.275 F
(the same time, congestion)2.75 E
(control mechanisms should not respond too se)100.8 153.6 Q -.165(ve)
-.275 G(rely to transient changes, e.g., to a sudden).165 E
(increase in delay that will dissipate in less than the connection')
100.8 166.6 Q 2.75(sr)-.605 G(ound-trip time.)-2.75 E
(Congestion control mechanisms also ha)100.8 192.6 Q .33 -.165(ve t)-.22
H 2.75(oc).165 G(ontend with sudden changes in the)-2.75 E
(bandwidth-delay product due to mobility)100.8 205.6 Q 5.5(.S)-.715 G
(uch bandwith-delay product changes are)-5.5 E -.165(ex)100.8 218.6 S
(pected to become more frequent and to ha).165 E .33 -.165(ve g)-.22 H
(reater impact than path changes today).165 E 5.5(.A)-.715 G(s)-5.5 E
2.75(ar)100.8 231.6 S(esult of both mobility and of the heterogeneity o\
f wireless access types \(802.11b,a,g,)-2.75 E(WIMAX, WCDMA, HS-WCDMA, \
E-GPRS, Bluetooth, etc.\), both the bandwidth and the)100.8 244.6 Q
(round-trip delay can change suddenly)100.8 257.6 Q 2.75(,s)-.715 G
(ometimes by se)-2.75 E -.165(ve)-.275 G(ral orders of magnitude.).165 E
(Ev)100.8 283.6 Q(aluating the response to sudden or transient changes \
can be of particular concern for)-.275 E(slo)100.8 296.6 Q(wly-respondi\
ng congestion control mechanisms such as equation-based congestion)-.275
E(control [RFC3448], and for AIMD \(Additi)100.8 309.6 Q .33 -.165(ve I)
-.275 H(ncrease Multiplicati).165 E .33 -.165(ve D)-.275 H
(ecrease\) or related).165 E(mechanisms using parameters that mak)100.8
322.6 Q 2.75(et)-.11 G(hem more slo)-2.75 E
(wly-responding than TCP [BB01])-.275 E([BBFS01].)100.8 335.6 Q
(In addition to the responsi)100.8 361.6 Q -.165(ve)-.275 G
(ness and smoothness of aggre).165 E -.055(ga)-.165 G(te traf).055 E
(\214c, one can consider the)-.275 E(tradeof)100.8 374.6 Q
(fs between responsi)-.275 E -.165(ve)-.275 G
(ness, smoothness, and aggressi).165 E -.165(ve)-.275 G
(ness for an indi).165 E(vidual)-.275 E(connection [FHP00] [YKL01].)
100.8 387.6 Q(In this case smoothness can be de\214ned by the lar)5.5 E
(gest)-.198 E(reduction in the sending rate in one round-trip time, in \
a deterministic en)100.8 400.6 Q(vironment with a)-.44 E(pack)100.8
413.6 Q(et drop e)-.11 E(xactly e)-.165 E -.165(ve)-.275 G(ry 1/p pack)
.165 E 2.75(ets. The)-.11 F(responsi)2.75 E -.165(ve)-.275 G
(ness is de\214ned as the number of).165 E(round-trip times of sustaine\
d congestion required for the sender to halv)100.8 426.6 Q 2.75(et)-.165
G(he sending rate,)-2.75 E(and the aggressi)100.8 439.6 Q -.165(ve)-.275
G(ness is de\214ned as the maximum increase in the sending rate in one)
.165 E(round-trip time, in pack)100.8 452.6 Q
(ets per second, in the absence of congestion.)-.11 E(This aggressi)5.5
E -.165(ve)-.275 G(ness).165 E
(can be a function of the mode of the transport protocol; for TCP)100.8
465.6 Q 2.75(,t)-1.221 G(he aggressi)-2.75 E -.165(ve)-.275 G(ness of)
.165 E(slo)100.8 478.6 Q(w-start is quite dif)-.275 E
(ferent from the aggressi)-.275 E -.165(ve)-.275 G(ness of congestion a)
.165 E -.22(vo)-.22 G(idance mode.).22 E F1 2.75(2.2.2. Minimizing)72
517.6 R(Oscillations)2.75 E F0(One goal is that of stability)100.8 534.2
Q 2.75(,i)-.715 G 2.75(nt)-2.75 G
(erms of minimizing oscillations of queueing delay or of)-2.75 E 2.75
(throughput. In)100.8 547.2 R(practice, stability is frequently associa\
ted with rate \215uctuations or v)2.75 E(ariance.)-.275 E(Rate v)100.8
560.2 Q(ariations can result in \215uctuations in router queue size and\
 therefore of queue)-.275 E -.165(ove)100.8 573.2 S(r\215o).165 E 2.75
(ws. These)-.275 F(queue o)2.75 E -.165(ve)-.165 G(r\215o).165 E
(ws can cause loss synchronizations across co-e)-.275 E(xisting \215o)
-.165 E(ws)-.275 E(and periodic under)100.8 586.2 Q
(-utilization of link capacity)-.22 E 2.75(,b)-.715 G
(oth of which are considered to be general)-2.75 E(signs of netw)100.8
599.2 Q(ork instability)-.11 E 2.75(.T)-.715 G
(hus, measuring the rate v)-2.75 E(ariations of \215o)-.275 E
(ws is often used to)-.275 E
(measure the stability of transport protocols.)100.8 612.2 Q 1.76 -.88
(To m)5.5 H(easure rate v).88 E(ariations, [JWL04], [RX05],)-.275 E
(and [FHPW00] use the coef)100.8 625.2 Q(\214cient of v)-.275 E
(ariation \(CoV\) of per)-.275 E(-\215o)-.22 E 2.75(wt)-.275 G
(ransmission rates and)-2.75 E([WCL05] suggests the use of standard de)
100.8 638.2 Q(viations of per)-.275 E(-\215o)-.22 E 2.75(wr)-.275 G 2.75
(ates. Since)-2.75 F(rate v)2.75 E(ariations)-.275 E
(are a function of time scales, it mak)100.8 651.2 Q
(es sense to measure these rate v)-.11 E(ariations o)-.275 E -.165(ve)
-.165 G 2.75(rv).165 G(arious)-3.025 E(time scales.)100.8 664.2 Q
(Measuring per)100.8 690.2 Q(-\215o)-.22 E 2.75(wr)-.275 G(ate v)-2.75 E
(ariations, ho)-.275 E(we)-.275 E -.165(ve)-.275 G .88 -.44(r, i).165 H
2.75(so).44 G(nly one aspect of transport protocol)-2.75 E(stability)
100.8 703.2 Q 5.5(.A)-.715 G(realistic e)-2.75 E(xperiment setting al)
-.165 E -.11(wa)-.11 G(ys in).11 E -.22(vo)-.44 G(lv).22 E
(es multiple \215o)-.165 E(ws of the transport)-.275 E
(protocol being observ)100.8 716.2 Q
(ed, along with a signi\214cant amount of cross traf)-.165 E
(\214c, with rates v)-.275 E(arying)-.275 E(Flo)72 769 Q 300.891
(yd Section)-.11 F 2.75(2.2.2. [P)2.75 F(age 9])-.165 E 0 Cg EP
%%Page: 10 10
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E -.165(ove)100.8 85 S 2.75(rt)
.165 G(ime, on both the forw)-2.75 E(ard and re)-.11 E -.165(ve)-.275 G
(rse paths. As a congestion control protocol must).165 E
(adapt its rate to the v)100.8 98 Q(arying rates of competing traf)-.275
E(\214c, just measuring the per)-.275 E(-\215o)-.22 E 2.75(ws)-.275 G
(tatistics)-2.75 E(of a subset of the traf)100.8 111 Q(\214c could be m\
isleading because it measures the rate \215uctuations due)-.275 E
(in part to the adaptation to competing traf)100.8 124 Q
(\214c on the path.)-.275 E(Thus, per)5.5 E(-\215o)-.22 E 2.75(ws)-.275
G(tatistics are most)-2.75 E(meaningful if the)100.8 137 Q 2.75(ya)-.165
G(re accompanied by the statistics measured at the netw)-2.75 E(ork le)
-.11 E -.165(ve)-.275 G 2.75(l. As).165 F(a)2.75 E
(complementary metric to the per)100.8 150 Q(-\215o)-.22 E 2.75(ws)-.275
G(tatistics, [HKLRX06] uses measurements of the)-2.75 E(rate v)100.8 163
Q(ariations of the aggre)-.275 E -.055(ga)-.165 G(te \215o).055 E
(ws observ)-.275 E(ed in bottleneck routers o)-.165 E -.165(ve)-.165 G
2.75(rv).165 G(arious time)-3.025 E(scales.)100.8 176 Q(Minimizing osci\
llations in queueing delay or throughput has related per)100.8 202 Q
(-\215o)-.22 E 2.75(wm)-.275 G(etrics of)-2.75 E
(minimizing jitter in round-trip times and loss rates.)100.8 215 Q(An o\
rthogonal goal for some congestion control mechanisms, e.g., for equati\
on-based)100.8 241 Q(congestion control, is to minimize the oscillation\
s in the sending rate for an indi)100.8 254 Q(vidual)-.275 E
(connection, gi)100.8 267 Q -.165(ve)-.275 G 2.75(na).165 G 2.75(ne)
-2.75 G -.44(nv)-2.75 G(ironment with a \214x).44 E
(ed, steady-state pack)-.165 E(et drop rate.)-.11 E(\(As is well)5.5 E
(kno)100.8 280 Q(wn, TCP congestion control is characterized by a prono\
unced oscillation in the sending)-.275 E(rate, with the sender halving \
the sending rate in response to congestion.\))100.8 293 Q
(One metric for)5.5 E(the le)100.8 306 Q -.165(ve)-.275 G 2.75(lo).165 G
2.75(fo)-2.75 G(scillations is the smoothness metric gi)-2.75 E -.165
(ve)-.275 G 2.75(ni).165 G 2.75(nS)-2.75 G(ection 2.2.1 abo)-2.75 E
-.165(ve)-.165 G(.).165 E
(As discussed in [FK07], the synchronization of loss e)100.8 332 Q -.165
(ve)-.275 G(nts can also af).165 E(fect con)-.275 E -.165(ve)-.44 G
-.198(rg).165 G(ence).198 E(times, throughput, and delay)100.8 345 Q(.)
-.715 E/F1 11/Times-Bold@0 SF(2.3.)72 384 Q/F2 13/Times-Bold@0 SF -.325
(Fa)5.5 G(ir).325 E(ness and Con)-.195 E -.13(ve)-.52 G -.13(rg).13 G
(ence).13 E F0(Another set of metrics is that of f)100.8 400.6 Q
(airness and con)-.11 E -.165(ve)-.44 G -.198(rg).165 G(ence times.).198
E -.165(Fa)5.5 G(irness can be).165 E(considered between \215o)100.8
413.6 Q(ws of the same protocol, and between \215o)-.275 E(ws using dif)
-.275 E(ferent)-.275 E
(protocols \(e.g., TCP-friendliness, referring to f)100.8 426.6 Q
(airness between TCP and a ne)-.11 E 2.75(wt)-.275 G(ransport)-2.75 E
2.75(protocol\). F)100.8 439.6 R(airness can also be considered between\
 sessions, between users, or between)-.165 E(other entities.)100.8 452.6
Q(There are a number of dif)100.8 478.6 Q(ferent f)-.275 E
(airness measures.)-.11 E(These include max-min f)5.5 E(airness)-.11 E
([HG86], proportional f)100.8 491.6 Q(airness [KMT98] [K01], the f)-.11
E(airness inde)-.11 E 2.75(xp)-.165 G(roposed in [JCH84],)-2.75 E
(and the product measure, a v)100.8 504.6 Q(ariant of netw)-.275 E
(ork po)-.11 E(wer [BJ81].)-.275 E F1 2.75(2.3.1. Metrics)72 543.6 R
-.275(fo)2.75 G 2.75(rf).275 G(air)-2.75 E(ness between \215o)-.165 E
(ws)-.11 E F0(This section discusses f)100.8 569.6 Q
(airness metrics that consider the f)-.11 E(airness between \215o)-.11 E
(ws, b)-.275 E(ut that)-.22 E(don')100.8 582.6 Q 2.75(tt)-.198 G(ak)
-2.75 E 2.75(ei)-.11 G(nto account dif)-2.75 E
(ferent characteristics of \215o)-.275 E
(ws \(e.g., the number of links in the)-.275 E
(path, or the round-trip time\).)100.8 595.6 Q -.165(Fo)5.5 G 2.75(rt)
.165 G(he discussion of f)-2.75 E(airness metrics, let x_i be the)-.11 E
(throughput for the i-th connection.)100.8 608.6 Q(Jain')100.8 647.6 Q
2.75(sf)-.605 G(airness inde)-2.86 E(x: The f)-.165 E(airness inde)-.11
E 2.75(xi)-.165 G 2.75(n[)-2.75 G(JCH84] is)-2.75 E
(\(\( sum_i x_i \)^2\) / \(n * sum_i \( \(x_i\)^2 \)\),)109.05 673.6 Q
(where there are n users.)100.8 699.6 Q(This f)5.5 E(airness inde)-.11 E
2.75(xr)-.165 G(anges from 0 to 1, and is maximum when all)-2.75 E
(users recei)100.8 712.6 Q .33 -.165(ve t)-.275 H(he same allocation.)
.165 E(This inde)5.5 E 2.75(xi)-.165 G 2.75(sk)-2.75 G
(/n when k users equally share the)-2.75 E
(resource, and the other n-k users recei)100.8 725.6 Q .33 -.165(ve z)
-.275 H(ero allocation.).165 E(Flo)72 769 Q 295.391(yd Section)-.11 F
2.75(2.3.1. [P)2.75 F(age 10])-.165 E 0 Cg EP
%%Page: 11 11
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E
(The product measure: The product measure)100.8 85 Q(product_i x_i ,)
109.05 111 Q(the product of the throughput of the indi)100.8 137 Q
(vidual connections, is also used as a measure of)-.275 E -.11(fa)100.8
150 S 2.75(irness. \(In).11 F(some conte)2.75 E(xts x_i is tak)-.165 E
(en as the po)-.11 E(wer of the i-th connection, and the product)-.275 E
(measure is referred to as netw)100.8 163 Q(ork po)-.11 E(wer)-.275 E
2.75(.\) The)-.605 F(product measure is particularly sensiti)2.75 E .33
-.165(ve t)-.275 H(o).165 E(se)100.8 176 Q(gre)-.165 E -.055(ga)-.165 G
(tion; the product measure is zero if an).055 E 2.75(yc)-.165 G
(onnection recei)-2.75 E -.165(ve)-.275 G 2.75(sz).165 G
(ero throughput.)-2.75 E(In)5.5 E([MS91] it is sho)100.8 189 Q
(wn that for a netw)-.275 E(ork with man)-.11 E 2.75(yc)-.165 G
(onnections and one shared g)-2.75 E(ate)-.055 E -.11(wa)-.275 G 1.43
-.715(y, t).11 H(he).715 E
(product measure is maximized when all connections recei)100.8 202 Q .33
-.165(ve t)-.275 H(he same throughput.).165 E(Epsilon-f)100.8 228 Q
(airness: A rate allocation is de\214ned as epsilon-f)-.11 E(air if)-.11
E(\(min_i x_i\) / \(max_i x_i\) >= 1 - epsilon.)109.05 254 Q(Epsilon-f)
100.8 280 Q(airness measures the w)-.11 E(orst-case ratio between an)
-.11 E 2.75(yt)-.165 G .22 -.11(wo t)-2.75 H(hroughput rates [ZKL04].)
.11 E(Epsilon-f)100.8 293 Q(airness is related to max-min f)-.11 E
(airness, de\214ned later in this document.)-.11 E/F1 11/Times-Bold@0 SF
2.75(2.3.2. Metrics)72 332 R -.275(fo)2.75 G 2.75(rf).275 G(air)-2.75 E
(ness between \215o)-.165 E(ws with differ)-.11 E(ent r)-.198 E(esour)
-.198 E(ce r)-.198 E(equir)-.198 E(ements)-.198 E F0
(This section discusses metrics for f)100.8 358 Q(airness between \215o)
-.11 E(ws with dif)-.275 E(ferent resource)-.275 E
(requirements, that is, with dif)100.8 371 Q
(ferent utility functions, round-trip times, or numbers of links)-.275 E
(on the path.)100.8 384 Q(Man)5.5 E 2.75(yo)-.165 G 2.75(ft)-2.75 G
(hese metrics can be described as solutions to utility maximization)
-2.75 E(problems [K01]; the total utility quanti\214es both the f)100.8
397 Q(airness and the throughput.)-.11 E(Max-min f)100.8 423 Q
(airness: In order to satisfy the max-min f)-.11 E
(airness criteria, the smallest throughput)-.11 E(rate must be as lar)
100.8 436 Q(ge as possible. Gi)-.198 E -.165(ve)-.275 G 2.75(nt).165 G
(his condition, the ne)-2.75 E(xt-smallest throughput rate)-.165 E
(must be as lar)100.8 449 Q(ge as possible, and so on.)-.198 E
(Thus, the max-min f)5.5 E(airness gi)-.11 E -.165(ve)-.275 G 2.75(sa)
.165 G(bsolute priority)-2.75 E(to the smallest \215o)100.8 462 Q 2.75
(ws. \(Max-min)-.275 F -.11(fa)2.75 G(irness can be e).11 E
(xplained by the progressi)-.165 E .33 -.165(ve \214)-.275 H(lling).165
E(algorithm, where all \215o)100.8 475 Q 2.75(wr)-.275 G
(ates start at zero, and the rates all gro)-2.75 E 2.75(wa)-.275 G 2.75
(tt)-2.75 G(he same pace.)-2.75 E(Each)5.5 E<8d6f>100.8 488 Q 2.75(wr)
-.275 G(ate stops gro)-2.75 E
(wing only when one or more links on the path reaches link capacity)
-.275 E(.\))-.715 E(Proportional f)100.8 514 Q(airness: In contrast, a \
feasible allocation x is de\214ned as proportionally f)-.11 E(air if)
-.11 E(for an)100.8 527 Q 2.75(yo)-.165 G
(ther feasible allocation x*, the aggre)-2.75 E -.055(ga)-.165 G
(te of proportional changes is zero or).055 E(ne)100.8 540 Q -.055(ga)
-.165 G(ti).055 E -.165(ve)-.275 G(:).165 E
(sum_i \( \(x*_i - x_i\)/x_i \) <= 0.)109.05 566 Q("This criterion f)
100.8 592 Q -.22(avo)-.11 G(urs smaller \215o).22 E(ws, b)-.275 E
(ut less emphatically than max-min f)-.22 E(airness" [K01].)-.11 E
(\(Using the language of utility functions, proportional f)100.8 605 Q
(airness can be achie)-.11 E -.165(ve)-.275 G 2.75(db).165 G 2.75(yu)
-2.75 G(sing)-2.75 E(log)100.8 618 Q
(arithmic utility functions, and maximizing the sum of the per)-.055 E
(-\215o)-.22 E 2.75(wu)-.275 G(tility functions; see)-2.75 E
([KMT98] for a fuller e)100.8 631 Q(xplanation.\))-.165 E
(Minimum potential delay f)100.8 657 Q
(airness: Minimum potential delay f)-.11 E(airness has been sho)-.11 E
(wn to)-.275 E(model TCP [KS03], and is a compromise between max-min f)
100.8 670 Q(airness and proportional)-.11 E -.11(fa)100.8 683 S 2.75
(irness. An).11 F(allocation x is de\214ned as ha)2.75 E
(ving minimum potential delay f)-.22 E(airness if)-.11 E
(sum_i \(1/x_i\))109.05 709 Q(Flo)72 769 Q 295.391(yd Section)-.11 F
2.75(2.3.2. [P)2.75 F(age 11])-.165 E 0 Cg EP
%%Page: 12 12
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E(is smaller than for an)100.8
85 Q 2.75(yo)-.165 G(ther feasible allocation.)-2.75 E(That is, it w)5.5
E(ould minimize the a)-.11 E -.165(ve)-.22 G(rage).165 E(do)100.8 98 Q
(wnload time if each \215o)-.275 E 2.75(ww)-.275 G
(as an equal-sized \214le.)-2.86 E
(In [CRM05], Colussi et al. propose a ne)100.8 124 Q 2.75(wd)-.275 G
(e\214nition of TCP f)-2.75 E(airness, that "a set of TCP f)-.11 E(air)
-.11 E<8d6f>100.8 137 Q
(ws do not cause more congestion than a set of TCP \215o)-.275 E(ws w)
-.275 E(ould cause", where)-.11 E
(congestion is de\214ned in terms of queueing delay)100.8 150 Q 2.75(,q)
-.715 G(ueueing delay v)-2.75 E(ariation, the congestion)-.275 E -2.365
-.275(ev e)100.8 163 T(nt rate [e.g., loss e).275 E -.165(ve)-.275 G
(nt rate], and the pack).165 E(et loss rate.)-.11 E(Chiu and T)100.8 189
Q(an in [CT06] ar)-.88 E(gue for rede\214ning the notion of f)-.198 E
(airness when studying traf)-.11 E<8c63>-.275 E
(controls for inelastic traf)100.8 202 Q
(\214c, proposing that inelastic \215o)-.275 E(ws adopt other traf)-.275
E(\214c controls such)-.275 E(as admission control.)100.8 215 Q
(Briscoe in [B07] ar)100.8 241 Q(gues that \215o)-.198 E(w-rate f)-.275
E(airness should not be a desired goal, and that)-.11 E
(instead "a realistic f)100.8 254 Q
(airness mechanism must share out the `cost' of each users actions on)
-.11 E(others".)100.8 267 Q/F1 11/Times-Bold@0 SF 2.75(2.3.3. Comments)
72 306 R(on fair)2.75 E(ness)-.165 E F0 -.385(Tr)100.8 332 S(adeof).385
E(fs between f)-.275 E(airness and throughput: The f)-.11 E
(airness measures in the section abo)-.11 E -.165(ve)-.165 G
(generally measure both f)100.8 345 Q(airness and throughput, gi)-.11 E
(ving dif)-.275 E(ferent weights to each.)-.275 E(Potential)5.5 E
(tradeof)100.8 358 Q(fs between f)-.275 E
(airness and throughput are also discussed by T)-.11 E
(ang et al. in [TWL06], for)-.88 E 2.75(af)100.8 371 S(rame)-2.75 E -.11
(wo)-.275 G(rk where max-min f).11 E(airness is de\214ned as the most f)
-.11 E(air)-.11 E 5.5(.I)-.605 G 2.75(np)-5.5 G(articular)-2.75 E 2.75
(,[)-.44 G(TWL06])-2.75 E(sho)100.8 384 Q
(ws that in some topologies throughput is proportional to f)-.275 E
(airness, while in other)-.11 E(topologies throughput is in)100.8 397 Q
-.165(ve)-.44 G(rsely proportional to f).165 E(airness.)-.11 E -.165(Fa)
100.8 423 S(irness and the number of congested links: Some of these f)
.165 E(airness metrics are discussed in)-.11 E(more detail in [F91].)
100.8 436 Q 1.76 -.88(We n)5.5 H
(ote that there is not a clear consensus for the f).88 E
(airness goals, in)-.11 E(particular for f)100.8 449 Q
(airness between \215o)-.11 E(ws that tra)-.275 E -.165(ve)-.22 G
(rse dif).165 E(ferent numbers of congested links)-.275 E 2.75
([F91]. Utility)100.8 462 R(maximization pro)2.75 E(vides one frame)
-.165 E -.11(wo)-.275 G(rk for describing this tradeof).11 E 2.75(fi)
-.275 G(n)-2.75 E -.11(fa)100.8 475 S(irness.).11 E -.165(Fa)100.8 501 S
(irness and round-trip times: One goal cited in a number of ne).165 E
2.75(wt)-.275 G(ransport protocols has)-2.75 E(been that of f)100.8 514
Q(airness between \215o)-.11 E(ws with dif)-.275 E
(ferent round-trip times [KHR02] [XHR04]. W)-.275 E(e)-.88 E
(note that there is not a consensus in the netw)100.8 527 Q
(orking community about the desirability of)-.11 E(this goal, or about \
the implications and interactions between this goal and other metrics)
100.8 540 Q([FJ92] \(Section 3.3\).)100.8 553 Q(One common ar)5.5 E
(gument ag)-.198 E(ainst the goal of f)-.055 E(airness between \215o)
-.11 E(ws)-.275 E(with dif)100.8 566 Q
(ferent round-trip times has been that \215o)-.275 E
(ws with long round-trip times consume)-.275 E
(more resources; this aspect is co)100.8 579 Q -.165(ve)-.165 G
(red by the pre).165 E(vious paragraph.)-.275 E(Researchers ha)5.5 E .33
-.165(ve a)-.22 H(lso).165 E(noted the dif)100.8 592 Q
(ference between the R)-.275 E(TT)-.66 E(-unf)-1.012 E
(airness of standard TCP)-.11 E 2.75(,a)-1.221 G(nd the greater R)-2.75
E(TT)-.66 E(-)-1.012 E(unf)100.8 605 Q
(airness of some proposed modi\214cations to TCP [LLS05].)-.11 E -.165
(Fa)100.8 631 S(irness and pack).165 E(et size: One f)-.11 E
(airness issue is that of the relati)-.11 E .33 -.165(ve f)-.275 H
(airness for \215o).055 E(ws with)-.275 E(dif)100.8 644 Q(ferent pack)
-.275 E(et sizes.)-.11 E(Man)5.5 E 2.75<798c>-.165 G
(le transfer applications will use the maximum pack)-2.75 E(et size)-.11
E 2.75(possible; in)100.8 657 R(contrast, lo)2.75 E(w-bandwidth V)-.275
E(oIP \215o)-1.419 E(ws are lik)-.275 E(ely to send small pack)-.11 E
(ets, sending a)-.11 E(ne)100.8 670 Q 2.75(wp)-.275 G(ack)-2.75 E(et e)
-.11 E -.165(ve)-.275 G(ry 10 to 40 ms., to limit delay).165 E 5.5(.S)
-.715 G(hould a small-pack)-5.5 E(et V)-.11 E(oIP connection)-1.419 E
(recei)100.8 683 Q .33 -.165(ve t)-.275 H
(he same sending rate in *bytes* per second as a lar).165 E(ge-pack)
-.198 E(et TCP connection in)-.11 E(the same en)100.8 696 Q
(vironment, or should it recei)-.44 E .33 -.165(ve t)-.275 H
(he same sending rate in *pack).165 E(ets* per second?)-.11 E(This f)
100.8 709 Q(airness issue has been discussed in more detail in [RFC3714\
], with [RFC4828] also)-.11 E(describing the w)100.8 722 Q
(ays that pack)-.11 E(et size can ef)-.11 E(fect the pack)-.275 E
(et drop rate e)-.11 E(xperienced by a \215o)-.165 E -.715(w.)-.275 G
(Flo)72 769 Q 295.391(yd Section)-.11 F 2.75(2.3.3. [P)2.75 F(age 12])
-.165 E 0 Cg EP
%%Page: 13 13
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E(Con)100.8 85 Q -.165(ve)-.44
G -.198(rg).165 G(ence times: Con).198 E -.165(ve)-.44 G -.198(rg).165 G
(ence times concern the time for con).198 E -.165(ve)-.44 G -.198(rg)
.165 G(ence to f).198 E(airness)-.11 E(between an e)100.8 98 Q
(xisting \215o)-.165 E 2.75(wa)-.275 G(nd a ne)-2.75 E
(wly-starting one, and are a special concern for)-.275 E(en)100.8 111 Q
(vironments with high-bandwidth long-delay \215o)-.44 E 2.75(ws. Con)
-.275 F -.165(ve)-.44 G -.198(rg).165 G(ence times also concern the).198
E(time for con)100.8 124 Q -.165(ve)-.44 G -.198(rg).165 G(ence to f)
.198 E(airness after a sudden change such as a change in the netw)-.11 E
(ork)-.11 E(path, the competing cross-traf)100.8 137 Q
(\214c, or the characteristics of a wireless link.)-.275 E(As with f)5.5
E(airness,)-.11 E(con)100.8 150 Q -.165(ve)-.44 G -.198(rg).165 G
(ence times can matter both between \215o).198 E
(ws of the same protocol, and between \215o)-.275 E(ws)-.275 E
(using dif)100.8 163 Q(ferent protocols [SLFK03].)-.275 E
(One metric used for con)8.25 E -.165(ve)-.44 G -.198(rg).165 G
(ence times is the delta-).198 E -.11(fa)100.8 176 S(ir con).11 E -.165
(ve)-.44 G -.198(rg).165 G(ence time, de\214ned as the time tak).198 E
(en for tw)-.11 E 2.75<6f8d>-.11 G -.275(ow)-2.75 G 2.75(sw).275 G
(ith the same round-trip)-2.75 E(time to go from shares of 100/101-th a\
nd 1/101-th of the link bandwidth, to ha)100.8 189 Q(ving close to)-.22
E -.11(fa)100.8 202 S(ir sharing with shares of \(1+delta\)/2 and \(1-d\
elta\)/2 of the link bandwidth [BBFS01].).11 E(A)5.5 E
(similar metric for con)100.8 215 Q -.165(ve)-.44 G -.198(rg).165 G
(ence times measures the con).198 E -.165(ve)-.44 G -.198(rg).165 G
(ence time as the number of).198 E(round-trip times for tw)100.8 228 Q
2.75<6f8d>-.11 G -.275(ow)-2.75 G 2.75(st).275 G 2.75(or)-2.75 G
(each epsilon-f)-2.75 E(airness, when starting from a maximally-)-.11 E
(unf)100.8 241 Q(air state [ZKL04].)-.11 E/F1 11/Times-Bold@0 SF(2.4.)72
280 Q/F2 13/Times-Bold@0 SF(Rob)5.5 E(ustness f)-.26 E
(or Challenging En)-.325 E(vir)-.52 E(onments)-.234 E F0
(While congestion control mechanisms are generally e)100.8 296.6 Q -.275
(va)-.275 G(luated \214rst o).275 E -.165(ve)-.165 G 2.75(re).165 G -.44
(nv)-2.75 G(ironments with).44 E(static routing in a netw)100.8 309.6 Q
(ork of tw)-.11 E(o-w)-.11 E(ay point-to-point links, some en)-.11 E
(vironments bring up)-.44 E
(more challenging problems \(e.g., corrupted pack)100.8 322.6 Q
(ets, reordering, v)-.11 E(ariable bandwidth,)-.275 E
(mobility\) as well as ne)100.8 335.6 Q 2.75(wm)-.275 G
(etrics to be considered \(e.g., ener)-2.75 E(gy consumption\).)-.198 E
(Rob)100.8 361.6 Q(ustness for challenging en)-.22 E(vironments: Rob)
-.44 E(ustness needs to be e)-.22 E(xplored for paths with)-.165 E
(reordering, corruption, v)100.8 374.6 Q
(ariable bandwidth, asymmetric routing, router con\214guration)-.275 E
(changes, mobility)100.8 387.6 Q 2.75(,a)-.715 G(nd the lik)-2.75 E 2.75
(e[)-.11 G 2.75(GF04]. In)-2.75 F
(general, the Internet architecture has v)2.75 E(alued)-.275 E(rob)100.8
400.6 Q(ustness o)-.22 E -.165(ve)-.165 G 2.75(re).165 G -.275<668c>
-2.75 G(cienc).275 E 1.43 -.715(y, e)-.165 H
(.g., when there are tradeof).715 E(fs between rob)-.275 E
(ustness and the)-.22 E(throughput, delay)100.8 413.6 Q 2.75(,a)-.715 G
(nd f)-2.75 E(airness metrics described abo)-.11 E -.165(ve)-.165 G(.)
.165 E(Ener)100.8 439.6 Q(gy consumption: In mobile en)-.198 E
(vironments the ener)-.44 E(gy consumption for the mobile end-)-.198 E
(node can be a k)100.8 452.6 Q .33 -.165(ey m)-.11 H(etric that is af)
.165 E(fected by the transport protocol [TM02].)-.275 E
(The goodput ratio: F)100.8 478.6 Q(or wireless netw)-.165 E
(orks, the goodput ratio can be a useful metric, where)-.11 E
(the goodput ratio can be de\214ned as the useful data deli)100.8 491.6
Q -.165(ve)-.275 G(red to users as a fraction of the).165 E
(total amount of data transmitted on the netw)100.8 504.6 Q 2.75(ork. A)
-.11 F(high goodput ratio indicates an ef)2.75 E(\214cient)-.275 E
(use of the radio spectrum and lo)100.8 517.6 Q
(wer interference with other users.)-.275 E F1(2.5.)72 556.6 Q F2(Rob)
5.5 E(ustness to F)-.26 E(ailur)-.325 E(es and to Misbeha)-.234 E
(ving Users)-.325 E F0
(One goal is for congestion control mechanisms to be rob)100.8 573.2 Q
(ust to misbeha)-.22 E(ving users, such as)-.22 E(recei)100.8 586.2 Q
-.165(ve)-.275 G(rs that `lie' to data senders about the congestion e)
.165 E(xperienced along the path or)-.165 E(otherwise attempt to bypass\
 the congestion control mechanisms of the sender [SCW)100.8 599.2 Q
(A99].)-1.32 E
(Another goal is for congestion control mechanisms to be as rob)100.8
612.2 Q(ust as possible to f)-.22 E(ailures,)-.11 E(such as f)100.8
625.2 Q(ailures of routers in using e)-.11 E
(xplicit feedback to end-nodes or f)-.165 E(ailures of end-nodes)-.11 E
(to follo)100.8 638.2 Q 2.75(wt)-.275 G(he prescribed protocols.)-2.75 E
F1(2.6.)72 664.2 Q F2(Deployability)5.5 E F0
(One metric for congestion control mechanisms is their deplo)100.8 680.8
Q(yability in the current Internet.)-.11 E(Metrics related to deplo)
100.8 693.8 Q(yability include the ease of f)-.11 E
(ailure diagnosis, and the o)-.11 E -.165(ve)-.165 G(rhead in).165 E
(terms of pack)100.8 706.8 Q(et header size or added comple)-.11 E
(xity at end-nodes or routers.)-.165 E(Flo)72 769 Q 303.641(yd Section)
-.11 F 2.75(2.6. [P)2.75 F(age 13])-.165 E 0 Cg EP
%%Page: 14 14
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E(One k)100.8 85 Q .33 -.165
(ey a)-.11 H(spect of deplo).165 E(yability concerns the range of deplo)
-.11 E(yment needed for a ne)-.11 E(w)-.275 E
(congestion control mechanism.)100.8 98 Q(Consider the follo)5.5 E
(wing possible deplo)-.275 E(yment requirements:)-.11 E 2.75(*O)100.8
124 S(nly at the sender \(e.g., Ne)-2.75 E(wReno in TCP [RFC3782]\);)
-.275 E 2.75(*O)100.8 137 S(nly at the recei)-2.75 E -.165(ve)-.275 G
2.75(r\().165 G(e.g., delayed ackno)-2.75 E(wledgements in TCP\);)-.275
E 2.75(*B)100.8 150 S(oth the sender and recei)-2.75 E -.165(ve)-.275 G
2.75(r\().165 G(e.g., SA)-2.75 E(CK TCP [RFC2018]\);)-.44 E 2.75(*A)
100.8 163 S 2.75(tas)-2.75 G(ingle router \(e.g., RED [FJ93]\);)-2.75 E
2.75(*A)100.8 176 S(ll of the routers along the end-to-end path;)-2.75 E
2.75(*B)100.8 189 S
(oth end nodes and all routers along the path \(e.g., XCP [KHR02]\).)
-2.75 E(Some congestion control mechanisms \(e.g., XCP [KHR02], Quick-S\
tart [RFC4782]\) may)100.8 205.6 Q(also ha)100.8 218.6 Q .33 -.165(ve d)
-.22 H(eplo).165 E(yment issues with IPsec, IP in IP)-.11 E 2.75(,M)
-1.221 G(PLS, or other tunnels, or with non-)-2.75 E
(router queues such as those in Ethernet switches.)100.8 231.6 Q
(Another deplo)100.8 248.2 Q(yability issue concerns the comple)-.11 E
(xity of the code.)-.165 E(Ho)5.5 E 2.75(wc)-.275 G(omple)-2.75 E 2.75
(xi)-.165 G 2.75(st)-2.75 G(he code)-2.75 E
(required to implement the mechanism in softw)100.8 261.2 Q 2.75
(are? Is)-.11 F(\215oating point math required?)2.75 E(Ho)5.5 E(w)-.275
E(much ne)100.8 274.2 Q 2.75(ws)-.275 G(tate must be k)-2.75 E
(ept to implement the ne)-.11 E 2.75(wm)-.275 G
(echanism, and who holds that state,)-2.75 E
(the routers or the end-nodes?)100.8 287.2 Q 1.76 -.88(We n)5.5 H
(ote that we don').88 E 2.75(ts)-.198 G(uggest these questions as w)
-2.75 E(ays to)-.11 E(reduce the deplo)100.8 300.2 Q(yability metric to\
 a single number; we suggest them as issues that could be)-.11 E
(considered in e)100.8 313.2 Q -.275(va)-.275 G(luating the deplo).275 E
(yability of a proposed congestion control mechanism.)-.11 E/F1 11
/Times-Bold@0 SF(2.7.)72 339.2 Q/F2 13/Times-Bold@0 SF(Metrics f)5.5 E
(or Speci\214c T)-.325 E(ypes of T)-.962 E(ransport)-.962 E F0
(In some cases modi\214ed metrics are needed for e)100.8 355.8 Q -.275
(va)-.275 G(luating transport protocols intended for).275 E
(QoS-enabled en)100.8 368.8 Q(vironments or for belo)-.44 E(w-best-ef)
-.275 E(fort traf)-.275 E(\214c [VKD02] [KK03].)-.275 E F1(2.8.)72 394.8
Q F2(User)5.5 E(-Based Metrics)-.481 E F0
(An alternate approach that has been proposed for the e)100.8 411.4 Q
-.275(va)-.275 G(luation of congestion control).275 E(mechanisms w)100.8
424.4 Q(ould be to e)-.11 E -.275(va)-.275 G
(luate in terms of user metrics such as user satisf).275 E
(action, or in)-.11 E
(terms of application-speci\214c utility functions.)100.8 437.4 Q
(Such an approach w)5.5 E(ould require the)-.11 E(de\214nition of a ran\
ge of user metrics or of application-speci\214c utility functions for t\
he range)100.8 450.4 Q(of applications under consideration \(e.g., FTP)
100.8 463.4 Q 2.75(,H)-1.221 G(TTP)-2.75 E 2.75(,V)-1.221 G(oIP\).)
-4.169 E F1(3.)72 489.4 Q/F3 14/Times-Bold@0 SF(Metrics in the IP P)5.5
E(erf)-.28 E(ormance Metrics \(IPPM\) W)-.35 E(orking Gr)-1.05 E(oup)
-.252 E F0(The IPPM W)100.8 506 Q(orking Group [IPPM] w)-.88 E
(as established to de\214ne performance metrics to be)-.11 E
(used by netw)100.8 519 Q
(ork operators, end users, or independent testing groups.)-.11 E
(The metrics include)5.5 E(metrics for connecti)100.8 532 Q
(vity [RFC2678], delay and loss [RFC2679] [RFC2680] [RFC2681],)-.275 E
(delay v)100.8 545 Q(ariation [RFC3393], loss patterns [RFC3357], pack)
-.275 E(et reordering [RFC4737], b)-.11 E(ulk)-.22 E
(transfer capacity [RFC3148], and link capacity [CI07].)100.8 558 Q
(The IPPM documents gi)5.5 E -.165(ve)-.275 G(concrete, well-de\214ned \
metrics, along with a methodology for measuring the metric.)100.8 571 Q
(The)5.5 E(metrics discussed in this document ha)100.8 584 Q .33 -.165
(ve a d)-.22 H(if).165 E(ferent purpose from the IPPM metrics; in this)
-.275 E(document we are discussing metrics as used in analysis, simulat\
ions, and e)100.8 597 Q(xperiments for)-.165 E(the e)100.8 610 Q -.275
(va)-.275 G(luation of congestion control mechanisms.).275 E(Further)5.5
E 2.75(,w)-.44 G 2.75(ea)-2.75 G(re discussing these metrics)-2.75 E(in\
 a general sense, rather than looking for speci\214c concrete de\214nit\
ions for each metric.)100.8 623 Q(Ho)100.8 636 Q(we)-.275 E -.165(ve)
-.275 G .88 -.44(r, t).165 H(here are man).44 E 2.75(yc)-.165 G
(ases where the metric de\214nitions from IPPM could be useful, for)
-2.75 E(speci\214c issues of ho)100.8 649 Q 2.75(wt)-.275 G 2.75(om)
-2.75 G(easure these metrics in simulations or in testbeds, and for)
-2.75 E(pro)100.8 662 Q
(viding common de\214nitions for talking about these metrics.)-.165 E F1
(4.)72 688 Q F3(Comments on Methodology)5.5 E F0(The types of scenarios\
 that are used to test speci\214c metrics, and the range of parameters)
100.8 704.6 Q(that it is useful to consider)100.8 717.6 Q 2.75(,w)-.44 G
(ill be discussed in separate documents, e.g., along with)-2.75 E(Flo)72
769 Q 311.891(yd Section)-.11 F 2.75(4. [P)2.75 F(age 14])-.165 E 0 Cg
EP
%%Page: 15 15
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E
(speci\214c scenarios for use in e)100.8 85 Q -.275(va)-.275 G
(luating congestion control mechanisms.).275 E 1.76 -.88(We n)100.8
101.6 T(ote that it can be important to e).88 E -.275(va)-.275 G
(luate metrics o).275 E -.165(ve)-.165 G 2.75(raw).165 G
(ide range of en)-2.75 E(vironments,)-.44 E
(with a range of link bandwidths, congestion le)100.8 114.6 Q -.165(ve)
-.275 G(ls, and le).165 E -.165(ve)-.275 G(ls of statistical multiple)
.165 E 2.75(xing. It)-.165 F(is also important to e)100.8 127.6 Q -.275
(va)-.275 G
(luate congestion control mechanisms in a range of scenarios,).275 E(in\
cluding typical ranges of connection sizes and round-trip times [FK02].\
 It is also useful)100.8 140.6 Q(to compare metrics for ne)100.8 153.6 Q
2.75(wo)-.275 G 2.75(rm)-2.75 G
(odi\214ed transport protocols with those of the current)-2.75 E
(standards for TCP)100.8 166.6 Q(.)-1.221 E(As an e)100.8 183.2 Q
(xample from the literature on e)-.165 E -.275(va)-.275 G
(luating transport protocols, Li et al. in).275 E("Experimental Ev)100.8
196.2 Q(aluation of TCP Protocols for High-Speed Netw)-.275 E
(orks" [LLS05] focus on)-.11 E
(the performance of TCP in high-speed netw)100.8 209.2 Q
(orks, and consider metrics for aggre)-.11 E -.055(ga)-.165 G(te).055 E
(throughput, loss rates, f)100.8 222.2 Q(airness \(including f)-.11 E
(airness between \215o)-.11 E(ws with dif)-.275 E(ferent round-trip)
-.275 E(times\), response times \(including con)100.8 235.2 Q -.165(ve)
-.44 G -.198(rg).165 G(ence times\), and incremental deplo).198 E 2.75
(yment. More)-.11 F(general references on methodology include [J91]. P)
100.8 248.2 Q(apers that discuss the range of metrics)-.165 E(for e)
100.8 261.2 Q -.275(va)-.275 G
(luating congestion control include [MTZ04].).275 E/F1 11/Times-Bold@0
SF(5.)72 287.2 Q/F2 14/Times-Bold@0 SF(Security Considerations)5.5 E F0
(There are no security considerations in this document.)100.8 303.8 Q F1
(6.)72 329.8 Q F2(IAN)5.5 E 3.5(AC)-.28 G(onsiderations)-3.5 E F0
(There are no IAN)100.8 346.4 Q 2.75(Ac)-.385 G
(onsiderations in this document.)-2.75 E F1(7.)72 372.4 Q F2(Ackno)5.5 E
(wledgements)-.14 E F0(Thanks to Lachlan Andre)100.8 389 Q 1.43 -.715
(w, M)-.275 H(ark Allman, Armando Caro, Dah Ming Chiu, Eric Coe,).715 E
(Dado Colussi, W)100.8 402 Q(esle)-.88 E 2.75(yE)-.165 G(ddy)-2.75 E
2.75(,N)-.715 G(elson F)-2.75 E(onseca, Janardhan Iyeng)-.165 E(ar)-.055
E 2.75(,D)-.44 G(oug Leith, T)-2.75 E(on)-.88 E 2.75(yL)-.165 G(i,)-2.75
E(Sa)100.8 415 Q -.165(ve)-.22 G
(rio Mascolo, Sean Moore, Injong Rhee, Da).165 E(vid Ros, Juer)-.22 E
(gen Schoenw)-.198 E(aelder)-.11 E 2.75(,A)-.44 G(ndras)-2.75 E -1.221
(Ve)100.8 428 S(res, Michael W)1.221 E(elzl, and Damon W)-.88 E
(ischik, and members of the T)-.44 E(ransport Modeling)-.385 E
(Research Group for feedback and contrib)100.8 441 Q(utions.)-.22 E F2
(Inf)72 467 Q(ormati)-.35 E .28 -.14(ve R)-.14 H(efer).14 E(ences)-.252
E F0([AEO03] M. Allman, W)100.8 483.6 Q 2.75(.E)-1.012 G(ddy)-2.75 E
2.75(,a)-.715 G(nd S. Ostermann, Estimating Loss Rates W)-2.75 E
(ith TCP)-.44 E 2.75(,A)-1.221 G(CM)-3.19 E(Performance Ev)129.6 496.6 Q
(aluation Re)-.275 E(vie)-.275 E 1.43 -.715(w, 3)-.275 H
(1\(3\), December 2003.).715 E([BB01] D. Bansal and H. Balakrishnan, Bi\
nomial Congestion Control Algorithms, IEEE)100.8 513.2 Q
(Infocom, April 2001.)129.6 526.2 Q
([BBFS01] D. Bansal, H. Balakrishnan, S. Flo)100.8 542.8 Q
(yd, and S. Shenk)-.11 E(er)-.11 E 2.75(,D)-.44 G(ynamic Beha)-2.75 E
(vior of)-.22 E(Slo)129.6 555.8 Q(wly-Responsi)-.275 E .33 -.165(ve C)
-.275 H(ongestion Control Algorithms, SIGCOMM 2001.).165 E
([BJ81] K. Bharath-K)100.8 572.4 Q(umar and J. Jef)-.165 E(fre)-.275 E
1.43 -.715(y, A N)-.165 H .55 -.275(ew A).715 H
(pproach to Performance-Oriented Flo).275 E(w)-.275 E(Control, IEEE T)
129.6 585.4 Q(ransactions on Communications, V)-.385 E
(ol.29 N.4, April 1981.)-1.419 E([B07] B. Briscoe, Flo)100.8 602 Q 2.75
(wR)-.275 G(ate F)-2.75 E
(airness: Dismantling a Religion, internet-draft draft-briscoe-)-.165 E
(tsv)129.6 615 Q(area-f)-.275 E(air)-.11 E(-02.txt, w)-.22 E
(ork-in-progress, July 2007.)-.11 E([CI07] P)100.8 631.6 Q 2.75(.C)
-1.221 G(himento and J. Ishac, De\214ning Netw)-2.75 E(ork Capacity)-.11
E 2.75(,i)-.715 G(nternet-draft draft-ietf-ippm-)-2.75 E
(bw-capacity-05, w)129.6 644.6 Q(ork-in-progress, May 2007.)-.11 E
([CRM05] D. Colussi, A Ne)100.8 661.2 Q 2.75(wA)-.275 G
(pproach to TCP-F)-2.75 E(airness, Report C-2005-49, Uni)-.165 E -.165
(ve)-.275 G(rsity of).165 E(Helsinki, Finland, 2005.)129.6 674.2 Q
([CT06] D. Chiu and A. T)100.8 690.8 Q(am, Rede\214ning F)-.88 E
(airness in the Study of TCP-friendly T)-.165 E(raf)-.385 E<8c63>-.275 E
(Controls, T)129.6 703.8 Q(echnical Report, 2006.)-.77 E(Flo)72 769 Q
361.391(yd [P)-.11 F(age 15])-.165 E 0 Cg EP
%%Page: 16 16
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E
([DM06] N. Dukkipati and N. McK)100.8 85 Q(eo)-.275 E(wn, Wh)-.275 E
2.75(yF)-.055 G(lo)-2.75 E(w-Completion T)-.275 E
(ime is the Right Metric)-.385 E(for Congestion Control, A)129.6 98 Q
(CM SIGCOMM, January 2006.)-.44 E([F91] S. Flo)100.8 114.6 Q
(yd, Connections with Multiple Congested Gate)-.11 E -.11(wa)-.275 G
(ys in P).11 E(ack)-.165 E(et-Switched)-.11 E(Netw)129.6 127.6 Q(orks P)
-.11 E(art 1: One-w)-.165 E(ay T)-.11 E(raf)-.385 E
(\214c, Computer Communication Re)-.275 E(vie)-.275 E 1.43 -.715(w, V)
-.275 H(ol.21 No.5,)-.704 E(October 1991, p. 30-47.)129.6 140.6 Q
([FF99] Flo)100.8 157.2 Q(yd, S., and F)-.11 E
(all, K., "Promoting the Use of End-to-End Congestion Control in)-.165 E
(the Internet", IEEE/A)129.6 170.2 Q(CM T)-.44 E(ransactions on Netw)
-.385 E(orking, August 1999.)-.11 E([FHP00] S. Flo)100.8 186.8 Q
(yd, M. Handle)-.11 E 1.43 -.715(y, a)-.165 H(nd J. P).715 E(adh)-.165 E
(ye, A Comparison of Equation-Based and)-.055 E
(AIMD Congestion Control, May 2000.)129.6 199.8 Q(URL "http://www)8.25 E
(.icir)-.715 E(.or)-.605 E(g/tfrc/".)-.198 E([FHPW00] S. Flo)100.8 216.4
Q(yd, M. Handle)-.11 E 1.43 -.715(y, J)-.165 H 2.75(.P).715 G(adh)-2.915
E(ye, and J. W)-.055 E(idmer)-.44 E 2.75(,E)-.44 G
(quation-Based Congestion)-2.75 E
(Control for Unicast Applications, SIGCOMM 2000, August 2000.)129.6
229.4 Q([FJ92] S. Flo)100.8 246 Q(yd and V)-.11 E 2.75(.J)-1.419 G
(acobson, On T)-2.75 E(raf)-.385 E(\214c Phase Ef)-.275 E(fects in P)
-.275 E(ack)-.165 E(et-Switched Gate)-.11 E -.11(wa)-.275 G(ys,).11 E
(Internetw)129.6 259 Q(orking: Research and Experience, V)-.11 E
(.3 N.3, September 1992, p.115-156.)-1.419 E([FJ93] S. Flo)100.8 275.6 Q
(yd and V)-.11 E 2.75(.J)-1.419 G(acobson, Random Early Detection g)
-2.75 E(ate)-.055 E -.11(wa)-.275 G(ys for Congestion).11 E -1.342 -.814
(Av o)129.6 288.6 T(idance, IEEE/A).814 E(CM T)-.44 E
(ransactions on Netw)-.385 E(orking, V)-.11 E(.1 N.4, August 1993,)
-1.419 E([FK02] S. Flo)100.8 305.2 Q(yd and E. K)-.11 E(ohler)-.385 E
2.75(,I)-.44 G(nternet Research Needs Better Models, Hotnets-I. October)
-2.75 E(2002.)129.6 318.2 Q([FK07] S. Flo)100.8 334.8 Q(yd and E. K)-.11
E(ohler)-.385 E 2.75(,T)-.44 G(ools for the Ev)-3.63 E
(aluation of Simulation and T)-.275 E(estbed)-.77 E
(Scenarios, internet-draft draft-irtf-tmr)129.6 347.8 Q
(g-tools-04.txt, July 2007.)-.198 E([GF04] A. Gurto)100.8 364.4 Q 2.75
(va)-.165 G(nd S. Flo)-2.75 E(yd, Modeling W)-.11 E(ireless Links for T)
-.44 E(ransport Protocols, A)-.385 E(CM)-.44 E
(CCR, 34\(2\):85-96, April 2004.)129.6 377.4 Q([HKLRX06] S. Ha, Y)100.8
394 Q 2.75(.K)-1.419 G(im, L. Le, I. Rhee, and L. Xu, A Step T)-2.75 E
-2.31 -.275(ow a)-.88 H(rd Realistic Ev).275 E(aluation)-.275 E(of High\
-speed TCP Protocols, technical report, North Carolina State Uni)129.6
407 Q -.165(ve)-.275 G(rsity).165 E(,)-.715 E(January 2006.)129.6 420 Q
([HG86] E. Hahne and R. Gallager)100.8 436.6 Q 2.75(,R)-.44 G
(ound Robin Scheduling for F)-2.75 E(air Flo)-.165 E 2.75(wC)-.275 G
(ontrol in Data)-2.75 E(Communications Netw)129.6 449.6 Q
(orks, IEEE International Conference on Communications,)-.11 E
(June 1986.)129.6 462.6 Q([IPPM] IP Performance Metrics \(IPPM\) W)100.8
479.2 Q(orking Group, URL)-.88 E("http://www)129.6 492.2 Q(.ietf.or)
-.715 E(g/html.charters/ippm-charter)-.198 E(.html".)-.605 E
([J91] R. Jain, The Art of Computer Systems Performance Analysis: T)
100.8 508.8 Q(echniques for)-.77 E
(Experimental Design, Measurement, Simulation, and Modeling, John W)
129.6 521.8 Q(ile)-.44 E 2.75(y&S)-.165 G(ons,)-2.75 E(1991.)129.6 534.8
Q([JCH84] R. Jain, D.M. Chiu, and W)100.8 551.4 Q 2.75(.H)-1.012 G -.165
(aw)-2.75 G(e, A Quantitati).165 E .33 -.165(ve M)-.275 H(easure of F)
.165 E(airness and)-.165 E(Discrimination for Resource Allocation in Sh\
ared Systems, DEC TR-301, Littleton,)129.6 564.4 Q
(MA: Digital Equipment Corporation, 1984.)129.6 577.4 Q
([JWL04] C. Jin, D. W)100.8 594 Q(ei, and S. Lo)-.88 E 1.43 -.715(w, F)
-.275 H(AST TCP: Moti)-.099 E -.275(va)-.275 G
(tion, Architecture, Algorithms,).275 E
(Performance, IEEE INFOCOM, March 2004.)129.6 607 Q([K01] F)100.8 623.6
Q 2.75(.K)-.88 G(elly)-3.025 E 2.75(,M)-.715 G
(athematical Modelling of the Internet, "Mathematics Unlimited - 2001)
-2.75 E(and Be)129.6 636.6 Q(yond" \(Editors B. Engquist and W)-.165 E
5.5(.S)-1.012 G(chmid\), Springer)-5.5 E(-V)-.22 E(erlag, Berlin, pp.)
-1.221 E(685-702, 2001.)129.6 649.6 Q([KHR02] D. Katabi, M. Handle)100.8
666.2 Q 1.43 -.715(y, a)-.165 H
(nd C. Rohrs, Congestion Control for High Bandwidth-).715 E
(Delay Product Netw)129.6 679.2 Q(orks, A)-.11 E(CM Sigcomm, 2002.)-.44
E([KK03] A. K)100.8 695.8 Q(uzmano)-.165 E(vic and E. W)-.165 E 2.75(.K)
-1.012 G(nightly)-2.75 E 2.75(,T)-.715 G(CP-LP: A Distrib)-2.75 E
(uted Algorithm for Lo)-.22 E(w)-.275 E(Priority Data T)129.6 708.8 Q
(ransfer)-.385 E 2.75(,I)-.44 G(EEE INFOCOM 2003, April 2003.)-2.75 E
(Flo)72 769 Q 361.391(yd [P)-.11 F(age 16])-.165 E 0 Cg EP
%%Page: 17 17
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E([KMT98] F)100.8 85 Q 2.75(.K)
-.88 G(elly)-3.025 E 2.75(,A)-.715 G 2.75(.M)-2.75 G(aulloo and D. T)
-2.75 E(an, Rate Control in Communication Netw)-.88 E(orks:)-.11 E
(Shado)129.6 98 Q 2.75(wP)-.275 G(rices, Proportional F)-2.75 E
(airness and Stability)-.165 E 5.5(.J)-.715 G(ournal of the Operational)
-5.5 E(Research Society 49, pp. 237-252, 1998.)129.6 111 Q([KS03] S. K)
100.8 127.6 Q
(unniyur and R. Srikant, End-to-end Congestion Control Schemes: Utility)
-.165 E(Functions, Random Losses and ECN Marks, IEEE/A)129.6 140.6 Q
(CM T)-.44 E(ransactions on)-.385 E(Netw)129.6 153.6 Q
(orking, 11\(5\):689-702, October 2003.)-.11 E([LLS05] Y)100.8 170.2 Q
(-T)-1.221 E 2.75(.L)-.814 G
(i, D. Leith, and R. Shorten, Experimental Ev)-2.75 E
(aluation of TCP Protocols for)-.275 E(High-Speed Netw)129.6 183.2 Q
(orks, Hamilton Institute, 2005.)-.11 E(URL)5.5 E("http://www)129.6
196.2 Q(.hamilton.ie/net/e)-.715 E -.275(va)-.275 G
(l/results-HI2005.pdf".).275 E([MS91] D. Mitra and J. Seery)100.8 212.8
Q 2.75(,D)-.715 G(ynamic Adapti)-2.75 E .33 -.165(ve W)-.275 H(indo)
-.275 E(ws for High Speed Data Netw)-.275 E(orks)-.11 E(with Multiple P)
129.6 225.8 Q(aths and Propag)-.165 E
(ation Delays, INFOCOM '91, pp 39-48.)-.055 E([MTZ04] L. Mamatas, V)
100.8 242.4 Q 2.75(.T)-1.419 G
(saoussidis, and C. Zhang, Approaches to Congestion Control in)-2.75 E
-.165(Pa)129.6 255.4 S(ck).165 E(et Netw)-.11 E(orks, 2004.)-.11 E
([RFC2018] TCP Selecti)100.8 272 Q .33 -.165(ve A)-.275 H(ckno).165 E
(wledgement Options.)-.275 E(M. Mathis, J. Mahda)5.5 E(vi, D. Flo)-.22 E
(yd,)-.11 E(and A. Romano)129.6 285 Q 4.18 -.715(w. R)-.275 H
(FC 2018, Proposed Standard, April 1996.).715 E
([RFC2680] Almes, G., Kalidindi, S., and M. Zekauskas, "A One-w)100.8
301.6 Q(ay P)-.11 E(ack)-.165 E(et Loss Metric)-.11 E
(for IPPM", RFC 2680, September 1999.)129.6 314.6 Q([RFC2678] Mahda)
100.8 331.2 Q(vi, J. and V)-.22 E 2.75(.P)-1.419 G
(axson, "IPPM Metrics for Measuring Connecti)-2.915 E(vity", RFC)-.275 E
(2678, September 1999.)129.6 344.2 Q([RFC2679] Almes, G.,)100.8 360.8 Q
(Kalidindi, S.)5.5 E(and M. Zekauskas, "A One-w)5.5 E
(ay Delay Metric for)-.11 E(IPPM", RFC 2679, September 1999.)129.6 373.8
Q([RFC2681] Almes, G., Kalidindi, S., Zekauskas, M., "A Round-T)100.8
390.4 Q(rip Delay Metric for)-.385 E
(IPPM". RFC 2681, STD 1, September 1999.)129.6 403.4 Q([RFC2914] S. Flo)
100.8 420 Q
(yd, Congestion Control Principles, RFC 2914, September 2000.)-.11 E
([RFC3148] Mathis, M. and M. Allman, "A Frame)100.8 436.6 Q -.11(wo)
-.275 G(rk for De\214ning Empirical Bulk).11 E -.385(Tr)129.6 449.6 S
(ansfer Capacity Metrics", RFC 3148, July 2001.).385 E
([RFC3168] K.K. Ramakrishnan, S. Flo)100.8 466.2 Q
(yd, and D. Black, The Addition of Explicit)-.11 E
(Congestion Noti\214cation \(ECN\) to IP)129.6 479.2 Q 2.75(,R)-1.221 G
(FC 3168, September 2001.)-2.75 E([RFC3357] R. K)100.8 495.8 Q
(oodli and R. Ra)-.385 E(vikanth, One-w)-.22 E(ay Loss P)-.11 E
(attern Sample Metrics, RFC)-.165 E(3357, Informational, August 2002.)
129.6 508.8 Q([RFC3393] Demichelis, C. and P)100.8 525.4 Q 2.75(.C)
-1.221 G(himento, "IP P)-2.75 E(ack)-.165 E(et Delay V)-.11 E
(ariation Metric for IP)-1.221 E
(Performance Metrics \(IPPM\)", RFC 3393, No)129.6 538.4 Q -.165(ve)
-.165 G(mber 2002.).165 E([RFC3448] M. Handle)100.8 555 Q 1.43 -.715
(y, S)-.165 H 2.75(.F).715 G(lo)-2.75 E(yd, J. P)-.11 E(adh)-.165 E
(ye, and J. W)-.055 E(idmer)-.44 E 2.75(,T)-.44 G
(CP Friendly Rate Control)-2.75 E(\(TFRC\): Protocol Speci\214cation, R\
FC 3448, Proposed Standard, January 2003.)129.6 568 Q([RFC3611] T)100.8
584.6 Q 2.75(.F)-.814 G(riedman, R. Caceres, and A. Clark, R)-2.75 E
(TP Control Protocol Extended)-.66 E(Reports \(R)129.6 597.6 Q
(TCP XR\), RFC 3611, No)-.66 E -.165(ve)-.165 G(mber 2003.).165 E
([RFC3714] S. Flo)100.8 614.2 Q(yd and J. K)-.11 E
(empf, IAB Concerns Re)-.275 E -.055(ga)-.165 G
(rding Congestion Control for V).055 E(oice)-1.419 E -.385(Tr)129.6
627.2 S(af).385 E(\214c in the Internet, RFC 3714, March 2004.)-.275 E
([RFC3782] S. Flo)100.8 643.8 Q(yd, T)-.11 E 2.75(.H)-.814 G
(enderson, and A. Gurto)-2.75 E 1.43 -.715(v, T)-.165 H(he Ne).715 E
(wReno Modi\214cation to TCP')-.275 E(s)-.605 E -.165(Fa)129.6 656.8 S
(st Reco).165 E -.165(ve)-.165 G
(ry Algorithm, RFC 3782, Proposed Standard, April 2004.).165 E
([RFC4737] A. Morton, L. Cia)100.8 673.4 Q -.275(va)-.22 G
(ttone, G. Ramachandran, S. Shaluno).275 E 1.43 -.715(v, J)-.165 H 2.75
(.P).715 G(erser)-2.75 E 2.75(,P)-.44 G(ack)-2.915 E(et)-.11 E
(Reordering Metrics, RFC 4737, No)129.6 686.4 Q -.165(ve)-.165 G
(mber 2006.).165 E([RFC4782] S. Flo)100.8 703 Q
(yd, M. Allman, A. Jain, and P)-.11 E 2.75(.S)-1.221 G
(arolahti, Quick-Start for TCP and IP)-2.75 E(,)-1.221 E
(RFC 4782, Experimental, January 2007.)129.6 716 Q(Flo)72 769 Q 361.391
(yd [P)-.11 F(age 17])-.165 E 0 Cg EP
%%Page: 18 18
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E([RFC4828] S. Flo)100.8 85 Q
(yd and E. K)-.11 E(ohler)-.385 E 2.75(,T)-.44 G
(CP Friendly Rate Control \(TFRC\): the Small-P)-2.75 E(ack)-.165 E(et)
-.11 E(\(SP\) V)129.6 98 Q(ariant, RFC 4828, April 2007.)-1.221 E
([RX05] I. Rhee and L. Xu, CUBIC: A Ne)100.8 114.6 Q 2.75(wT)-.275 G
(CP-Friendly High-Speed TCP V)-2.75 E(ariant,)-1.221 E(PFLDnet 2005.)
129.6 127.6 Q([SAF06] P)100.8 144.2 Q 2.75(.S)-1.221 G
(arolahti, M. Allman, and S. Flo)-2.75 E
(yd, Determining an Appropriate Sending Rate)-.11 E(Ov)129.6 157.2 Q
(er an Underutilized Netw)-.165 E(ork P)-.11 E(ath, Computer Netw)-.165
E(orks, September 2006.)-.11 E([SLFK03] R.N. Shorten, D.J. Leith, J. F)
100.8 173.8 Q -.11(oy)-.165 G 2.75(,a)-.605 G(nd R. Kilduf)-2.75 E
(f, Analysis and Design of)-.275 E
(Congestion Control in Synchronised Communication Netw)129.6 186.8 Q
(orks. Proc. 12th Y)-.11 E(ale)-1.1 E -.88(Wo)129.6 199.8 S
(rkshop on Adapti).88 E .33 -.165(ve & L)-.275 H
(earning Systems, May 2003.).165 E([SCW)100.8 216.4 Q(A99] S. Sa)-1.32 E
-.275(va)-.22 G(ge, N. Cardwell, D. W).275 E(etherall, and T)-.88 E 2.75
(.A)-.814 G(nderson, TCP Congestion)-2.75 E(Control with a Misbeha)129.6
229.4 Q(ving Recei)-.22 E -.165(ve)-.275 G .88 -.44(r, AC).165 H 2.75
(MC).44 G(omputer Communications Re)-2.75 E(vie)-.275 E -.715(w,)-.275 G
(October 1999.)129.6 242.4 Q([TM02] V)100.8 259 Q 2.75(.T)-1.419 G(saou\
ssidis and I. Matta, Open Issues of TCP for Mobile Computing, Journal o\
f)-2.75 E -.44(Wi)129.6 272 S
(reless Communications and Mobile Computing: Special Issue on Reliable)
.44 E -.385(Tr)129.6 285 S
(ansport Protocols for Mobile Computing, February 2002.).385 E
([TWL06] A. T)100.8 301.6 Q(ang, J. W)-.88 E(ang and S. H. Lo)-.88 E
4.18 -.715(w. C)-.275 H(ounter).715 E(-Intuiti)-.22 E .33 -.165(ve T)
-.275 H(hroughput Beha).165 E(viors in)-.22 E(Netw)129.6 314.6 Q
(orks Under End-to-End Control, IEEE/A)-.11 E(CM T)-.44 E
(ransactions on Netw)-.385 E(orking,)-.11 E
(14\(2\):355-368, April 2006.)129.6 327.6 Q([WCL05] D. X. W)100.8 344.2
Q(ei, P)-.88 E 2.75(.C)-1.221 G(ao and S. H. Lo)-2.75 E 1.43 -.715(w, T)
-.275 H(ime for a TCP Benchmark Suite?, T).33 E(echnical)-.77 E
(Report, Caltech CS, Stanford EAS, Caltech, 2005.)129.6 357.2 Q
([VKD02] A. V)100.8 373.8 Q(enkataramani, R. K)-1.221 E
(okku, and M. Dahlin, TCP Nice: A Mechanism for)-.385 E(Background T)
129.6 386.8 Q
(ransfers, Fifth USENIX Symposium on Operating System Design and)-.385 E
(Implementation \(OSDI\), 2002.)129.6 399.8 Q([XHR04] L. Xu, K. Harfous\
h, and I. Rhee, Binary Increase Congestion Control for F)100.8 416.4 Q
(ast,)-.165 E(Long Distance Netw)129.6 429.4 Q(orks, Infocom 2004.)-.11
E([YKL01] Y)100.8 446 Q 2.75(.Y)-1.419 G(ang, M. Kim, and S. Lam, T)
-3.85 E(ransient Beha)-.385 E(viors of TCP-friendly Congestion)-.22 E
(Control Protocols, Infocom 2001.)129.6 459 Q([ZKL04] Y)100.8 475.6 Q
2.75(.Z)-1.419 G(hang, S.-R. Kang, and D. Loguino)-2.75 E 1.43 -.715
(v, D)-.165 H(elayed Stability and Performance of).715 E(Distrib)129.6
488.6 Q(uted Congestion Control, A)-.22 E(CM SIGCOMM, August 2004.)-.44
E/F1 14/Times-Bold@0 SF -.7(Au)72 514.6 S(thors' Addr).7 E(esses)-.252 E
F0(Sally Flo)100.8 531.2 Q(yd <\215o)-.11 E(yd at icir)-.11 E(.or)-.605
E(g>)-.198 E(ICSI Center for Internet Research)100.8 544.2 Q
(1947 Center Street, Suite 600)100.8 557.2 Q(Berk)100.8 570.2 Q(ele)-.11
E 1.43 -.715(y, C)-.165 H 2.75(A9).715 G(4704)-2.75 E(USA)100.8 583.2 Q
F1(Full Copyright Statement)72 622.2 Q F0(Cop)100.8 638.8 Q
(yright \(C\) The IETF T)-.11 E(rust \(2007\).)-.385 E(This document is\
 subject to the rights, licenses and restrictions contained in BCP 78, \
and)100.8 655.4 Q -.165(ex)100.8 668.4 S
(cept as set forth therein, the authors retain all their rights.).165 E
(This document and the information contained herein are pro)100.8 685 Q
(vided on an "AS IS" basis and)-.165 E(THE CONTRIB)100.8 698 Q(UT)-.11 E
(OR, THE ORGANIZA)-.198 E(TION HE/SHE REPRESENTS OR IS)-1.221 E
(SPONSORED BY \(IF ANY\), THE INTERNET SOCIETY)100.8 711 Q 2.75(,T)
-1.419 G(HE IETF TR)-2.75 E(UST AND THE)-.44 E(INTERNET ENGINEERING T)
100.8 724 Q(ASK FORCE DISCLAIM ALL W)-1.023 E(ARRANTIES,)-1.32 E(Flo)72
769 Q 361.391(yd [P)-.11 F(age 18])-.165 E 0 Cg EP
%%Page: 19 19
%%BeginPageSetup
BP
%%EndPageSetup
/F0 11/Times-Roman@0 SF(INTERNET)72 49 Q 78.297(-DRAFT TMRG,)-1.012 F
110.592(METRICS October)2.75 F(2007)2.75 E
(EXPRESS OR IMPLIED, INCLUDING B)100.8 85 Q(UT NO)-.11 E 2.75(TL)-.44 G
(IMITED T)-2.75 E 2.75(OA)-.198 G(NY W)-2.75 E(ARRANTY)-1.32 E(THA)100.8
98 Q 2.75(TT)-1.221 G(HE USE OF THE INFORMA)-2.75 E(TION HEREIN WILL NO)
-1.221 E 2.75(TI)-.44 G(NFRINGE ANY)-2.75 E(RIGHTS OR ANY IMPLIED W)
100.8 111 Q(ARRANTIES OF MERCHANT)-1.32 E(ABILITY OR FITNESS)-1.023 E
(FOR A P)100.8 124 Q(AR)-1.012 E(TICULAR PURPOSE.)-.66 E/F1 14
/Times-Bold@0 SF(Intellectual Pr)72 150 Q(operty)-.252 E F0
(The IETF tak)100.8 166.6 Q(es no position re)-.11 E -.055(ga)-.165 G
(rding the v).055 E(alidity or scope of an)-.275 E 2.75(yI)-.165 G
(ntellectual Property)-2.75 E(Rights or other rights that might be clai\
med to pertain to the implementation or use of the)100.8 179.6 Q
(technology described in this document or the e)100.8 192.6 Q
(xtent to which an)-.165 E 2.75(yl)-.165 G(icense under such rights)
-2.75 E(might or might not be a)100.8 205.6 Q -.275(va)-.22 G
(ilable; nor does it represent that it has made an).275 E 2.75(yi)-.165
G(ndependent)-2.75 E(ef)100.8 218.6 Q(fort to identify an)-.275 E 2.75
(ys)-.165 G(uch rights.)-2.75 E
(Information on the procedures with respect to rights in)5.5 E
(RFC documents can be found in BCP 78 and BCP 79.)100.8 231.6 Q
(Copies of IPR disclosures made to the IETF Secretariat and an)100.8
248.2 Q 2.75(ya)-.165 G(ssurances of licenses to)-2.75 E(be made a)100.8
261.2 Q -.275(va)-.22 G(ilable, or the result of an attempt made to obt\
ain a general license or).275 E(permission for the use of such propriet\
ary rights by implementers or users of this)100.8 274.2 Q(speci\214cati\
on can be obtained from the IETF on-line IPR repository at)100.8 287.2 Q
(http://www)100.8 300.2 Q(.ietf.or)-.715 E(g/ipr)-.198 E(.)-.605 E
(The IETF in)100.8 316.8 Q(vites an)-.44 E 2.75(yi)-.165 G
(nterested party to bring to its attention an)-2.75 E 2.75(yc)-.165 G
(op)-2.75 E(yrights, patents or)-.11 E
(patent applications, or other proprietary rights that may co)100.8
329.8 Q -.165(ve)-.165 G 2.75(rt).165 G(echnology that may be)-2.75 E
(required to implement this standard.)100.8 342.8 Q
(Please address the information to the IETF at ietf-)5.5 E(ipr@ietf.or)
100.8 355.8 Q(g.)-.198 E(Flo)72 769 Q 361.391(yd [P)-.11 F(age 19])-.165
E 0 Cg EP
%%Trailer
end
%%EOF