Heterogeneous Congestion Control Protocols Steven Low CS, EE netlab.CALTECH.edu with A. Tang, J. Wang, D. Wei, Caltech M. Chiang, Princeton Outline Review: homogeneous case Motivating experiments Model Equilibrium Existence, uniqueness, local stability Slow timescale control Tang, Wang, Low, Chiang. Infocom March 2005 Tang, Wang, Hegde, Low. Telecommunications Systems, Dec 2005 Network model x y R F1 Network TCP G1 FN q AQM GL R T p Rli 1 if source i uses link l IP routing x(t 1) F ( RT p(t ), x(t )) Reno, Vegas p(t 1) G ( p(t ), Rx (t )) DT, RED, … Network model: example Reno: Jacobson 1989 for { for { every RTT W += 1 } every loss W := W/2 } (AI) (MD) AI 2 i x 1 xi (t 1) 2 Ti 2 R li pl (t ) MD l pl (t 1) Gl Rli xi (t ), pl (t ) i TailDrop Network model: example FAST: Jin, Wei, Low 2004 periodically { baseRTT W : W RTT } i xi (t 1) xi (t ) i xi (t ) Rli pl (t ) Ti l 1 pl (t 1) p l (t ) Rli xi (t ) cl cl i Duality model of TCP/AQM TCP/AQM x* F ( RT p * , x* ) p* G ( p* , Rx * ) Equilibrium (x*,p*) primal-dual optimal: max U i ( xi ) subject to Rx c x 0 F determines utility function U G guarantees complementary slackness Kelly, Maloo, Tan 1998 p* are Lagrange multipliers Low, Lapsley 1999 Uniqueness of equilibrium x* is unique when U is strictly concave p* is unique when R has full row rank Duality model of TCP/AQM TCP/AQM x* F ( RT p * , x* ) p* G ( p* , Rx * ) Equilibrium (x*,p*) primal-dual optimal: max U i ( xi ) subject to Rx c x 0 F determines utility function U G guarantees complementary slackness Kelly, Maloo, Tan 1998 p* are Lagrange multipliers Low, Lapsley 1999 The underlying concave program also leads to simple dynamic behavior Duality model of TCP/AQM Equilibrium (x*,p*) primal-dual optimal: max x 0 U ( x ) i i subject to Rx c Mo & Walrand 2000: log xi U i ( xi ) (1 ) 1 xi1 1 : 1.2: 2 : : if 1 if 1 Vegas, FAST, STCP HSTCP Reno XCP (single link only) Low 2003 Some implications Equilibrium Always exists, unique if R is full rank Bandwidth allocation independent of AQM or arrival Can predict macroscopic behavior of large scale networks Counter-intuitive throughput behavior Fair allocation is not always inefficient Increasing link capacities do not always raise aggregate throughput [Tang, Wang, Low, ToN 2006] FAST TCP Design, analysis, experiments [Jin, Wei, Low, ToN 2007] Some implications Equilibrium Always exists, unique if R is full rank Bandwidth allocation independent of AQM or arrival Can predict macroscopic behavior of large scale networks Counter-intuitive throughput behavior Fair allocation is not always inefficient Increasing link capacities do not always raise aggregate throughput [Tang, Wang, Low, ToN 2006] FAST TCP Design, analysis, experiments [Jin, Wei, Low, ToN 2007] Outline Review: homogeneous case Motivating experiments Model Equilibrium Existence, uniqueness, local stability Slow timescale control Throughputs depend on AQM FAST throughput buffer size = 80 pkts buffer size = 400 pkts FAST and Reno share a single bottleneck router NS2 simulation Router: DropTail with variable buffer size With 10% heavy-tailed noise traffic Multiple equilibria: throughput depends on arrival Dummynet experiment eq 2 eq 1 eq 2 Path 1 52M 13M path 2 61M 13M path 3 27M 93M eq 1 Tang, Wang, Hegde, Low, Telecom Systems, 2005 Multiple equilibria: throughput depends on arrival Dummynet experiment eq 2 eq 1 eq 2 Path 1 52M 13M path 2 61M 13M path 3 27M 93M eq 3 (unstable) eq 1 Tang, Wang, Hegde, Low, Telecom Systems, 2005 Some implications homogeneous heterogeneous equilibrium unique non-unique bandwidth allocation on AQM independent dependent bandwidth allocation on arrival independent dependent Duality model: max U i ( xi ) s.t. Rx c x 0 * * x Fi Rli pl , xi l * i Why can’t use Fi’s of FAST and Reno in duality model? They use different prices! i Fi xi i xi Rli pl Ti l delay for FAST xi2 1 Fi 2 Ti 2 loss for Reno R li l pl Duality model: max U i ( xi ) s.t. Rx c x 0 * * x Fi Rli pl , xi l * i Why can’t use Fi’s of FAST and Reno in duality model? They use different prices! i Fi xi i xi Rli pl Ti l 1 p l Rli xi (t ) cl cl i xi2 1 Fi 2 Ti 2 p l g l pl (t ), Rli xi (t ) i R li l pl Homogeneous protocol x y R F1 Network TCP G1 FN q GL R T xi (t 1) Fi Rli pl (t ), xi (t ) l pl (t 1) Gl pl (t ), AQM i Rli xi (t ) p same price for all sources Heterogeneous protocol x y R F1 G1 Network TCP FN GL q p R T xi (t 1) Fi Rli pl (t ), xi (t ) l heterogeneous prices for type j sources xi (t 1) Fi Rli mlj pl (t ) , xij (t ) l j AQM j Heterogeneous protocols Equilibrium: p that satisfies j xi ( p) f i Rli ml ( pl ) l cl j j yl ( p) : Rli xi ( p) i,j cl if pl 0 j j Duality model no longer applies ! pl can no longer serve as Lagrange multiplier Heterogeneous protocols Equilibrium: p that satisfies j xi ( p) f i Rli ml ( pl ) l cl j j yl ( p) : Rli xi ( p) i,j cl if pl 0 j j Need to re-examine all issues Equilibrium: exists? unique? efficient? fair? Dynamics: stable? limit cycle? chaotic? Practical networks: typical behavior? design guidelines? Heterogeneous protocols Equilibrium: p that satisfies j xi ( p) f i Rli ml ( pl ) l cl j j yl ( p) : Rli xi ( p) i,j cl if pl 0 j j Dynamic: dual algorithm j xi ( p (t )) f i Rli ml ( pl (t )) l p l l yl ( p (t )) cl j j Notation Simpler notation: p is equilibrium if y( p) c on bottleneck links Jacobian: J ( p) : y ( p) p Linearized dual algorithm: p J ( p* ) p(t) See Simsek, Ozdaglar, Acemoglu 2005 for generalization Outline Review: homogeneous case Motivating experiments Model Equilibrium Existence, uniqueness, local stability Slow timescale control Tang, Wang, Low, Chiang. Infocom 2005 Existence Theorem Equilibrium p exists, despite lack of underlying utility maximization Generally non-unique There are networks with unique bottleneck set but infinitely many equilibria There are networks with multiple bottleneck set each with a unique (but distinct) equilibrium Regular networks Definition A regular network is a tuple (R, c, m, U) for which all equilibria p are locally unique, i.e., y det J ( p) : det ( p) 0 p Theorem Almost all networks are regular A regular network has finitely many and odd number of equilibria (e.g. 1) Global uniqueness m lj [al ,21/ L al ] for any al 0 m lj [a j ,21/ L a j ] for any a j 0 Theorem If price heterogeneity is small, then equilibrium is globally unique Corollary If price mapping functions mlj are linear and linkindependent, then equilibrium is globally unique e.g. a network of RED routers almost always has globally unique equilibrium Local stability:`uniqueness’ stability j 1/ L ml [al ,2 al ] for any al 0 m lj [a j ,21/ L a j ] for any a j 0 Theorem If price heterogeneity is small, then the unique equilibrium p is locally stable Linearized dual algorithm: * p J( p ) p(t) Equilibrium p is locally stable if Re J( p) 0 Local stability:`converse’ Theorem If all equilibria p are locally stable, then it is globally unique Proof idea: For all equilibrium p: I ( p) (1) L Index theorem: L I ( p ) ( 1 ) eq p