Delay analysis of the Max-Weight policy under heavy-tailed traffic via fluid approximations

We consider a single-hop switched queueing network with a mix of heavy-tailed (i.e., arrival processes with infinite variance) and light-tailed traffic, and study the delay performance of the Max-Weight policy, known for its throughput optimality and asymptotic delay optimality properties. Classical results in queueing theory imply that heavy-tailed queues are delay unstable, i.e., they experience infinite expected delays in steady state. Thus, we focus on the impact of heavy-tailed traffic on the light-tailed queues, using delay stability as performance metric. Recent work has shown that this impact may come in the form of subtle rate-dependent phenomena, the stochastic analysis of which is quite cumbersome. Our goal is to show how fluid approximations can facilitate the delay analysis of the Max-Weight policy under heavy-tailed traffic. More specifically, we show how fluid approximations can be combined with renewal theory in order to prove delay instability results. Furthermore, we show how fluid approximations can be combined with stochastic Lyapunov theory in order to prove delay stability results. We illustrate the benefits of the proposed approach in two ways: (i) analytically, by providing a sharp characterization of the delay stability regions of networks with disjoint schedules, significantly generalizing previous results; (ii) computationally, through a Bottleneck Identification algorithm, which identifies (some) delay unstable queues by solving the fluid model of the network from certain initial conditions.

[1]  A. Borovkov Factorization Identities and Properties of the Distribution of the Supremum of Sequential Sums , 1970 .

[2]  J. Cohen SOME RESULTS ON REGULAR VARIATION FOR DISTRIBUTIONS IN QUEUEING AND FLUCTUATION THEORY , 1973 .

[3]  A. Pakes ON THE TAILS OF WAITING-TIME DISTRIBUTIONS , 1975 .

[4]  N. Veraverbeke Asymptotic behaviour of Wiener-Hopf factors of a random walk , 1977 .

[5]  B. Hajek Hitting-time and occupation-time bounds implied by drift analysis with applications , 1982, Advances in Applied Probability.

[6]  J. Tsitsiklis Analysis of a multiaccess control scheme , 1987 .

[7]  P. R. Kumar,et al.  Dynamic instabilities and stabilization methods in distributed real-time scheduling of manufacturing systems , 1990 .

[8]  Leandros Tassiulas,et al.  Stability properties of constrained queueing systems and scheduling policies for maximum throughput in multihop radio networks , 1990, 29th IEEE Conference on Decision and Control.

[9]  David Williams,et al.  Probability with Martingales , 1991, Cambridge mathematical textbooks.

[10]  Ronald W. Wolff,et al.  A Review of Regenerative Processes , 1993, SIAM Rev..

[11]  Kiran M. Rege,et al.  A decomposition theorem and related results for the discriminatory processor sharing queue , 1994, Queueing Syst. Theory Appl..

[12]  J. Dai On Positive Harris Recurrence of Multiclass Queueing Networks: A Unified Approach Via Fluid Limit Models , 1995 .

[13]  Marshall L. Fisher,et al.  Reducing the Cost of Demand Uncertainty Through Accurate Response to Early Sales , 1996, Oper. Res..

[14]  Nick McKeown,et al.  A Starvation-free Algorithm For Achieving 100% Throughput in an Input- Queued Switch , 1999 .

[15]  Garrett van Ryzin,et al.  Optimal Control of a Multiclass, Flexible Queueing System , 1997, Oper. Res..

[16]  Sean P. Meyn,et al.  Piecewise linear test functions for stability and instability of queueing networks , 1997, Queueing Syst. Theory Appl..

[17]  Onno J. Boxma,et al.  Fluid queues with long-tailed activity period distributions , 1997, Comput. Commun..

[18]  A. Lazar,et al.  Asymptotic results for multiplexing subexponential on-off processes , 1999, Advances in Applied Probability.

[19]  Gennady Samorodnitsky,et al.  Activity periods of an infinite server queue and performance of certain heavy tailed fluid queues , 1999, Queueing Syst. Theory Appl..

[20]  S. Borst,et al.  Exact asymptotics for fluid queues fed by multiple heavy-tailed on-off flows , 2004, math/0406178.

[21]  Walter Willinger,et al.  Self-Similar Network Traffic and Performance Evaluation , 2000 .

[22]  A. P. Zwart,et al.  Sojourn time asymptotics in the M/G/1 processor sharing queue , 1998, Queueing Syst. Theory Appl..

[23]  R. Mazumdar,et al.  LOSS ASYMPTOTICS IN LARGE BUFFERS FED BY HETEROGENEOUS LONG-TAILED SOURCES , 2000 .

[24]  J. Harrison Brownian models of open processing networks: canonical representation of workload , 2000 .

[25]  S. Borst,et al.  Fluid queues with heavy-tailed M/G/infinity input , 2001 .

[26]  J. Tsitsiklis,et al.  Performance of Multiclass Markovian Queueing Networks Via Piecewise Linear Lyapunov Functions , 2001 .

[27]  Ravi R. Mazumdar,et al.  Asymptotic analysis of GPS systems fed by heterogeneous long-tailed sources , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[28]  Walter Willinger,et al.  Self‐Similar Network Traffic: An Overview , 2002 .

[29]  A. P. Zwart,et al.  Waiting-Time Asymptotics for the M/G/2 Queue with Heterogeneous Servers , 2002, Queueing Syst. Theory Appl..

[30]  Rudesindo Núñez-Queija,et al.  Queues with Equally Heavy Sojourn Time and Service Requirement Distributions , 2002 .

[31]  Sem C. Borst,et al.  A Reduced-Load Equivalence for Generalised Processor Sharing Networks with Long-Tailed Input Flows , 2002, Queueing Syst. Theory Appl..

[32]  Predrag R. Jelenkovic,et al.  Finite buffer queue with generalized processor sharing and heavy-tailed input processes , 2002, Comput. Networks.

[33]  Sem C. Borst,et al.  Generalized processor sharing with light-tailed and heavy-tailed input , 2003, TNET.

[34]  P. Box,et al.  Reduced-Load Equivalence and Induced Burstiness in GPS Queues with Long-Tailed Traffic Flows , 2003 .

[35]  Predrag R. Jelenkovic,et al.  Large Deviation Analysis of Subexponential Waiting Times in a Processor-Sharing Queue , 2003, Math. Oper. Res..

[36]  Lisa Zhang,et al.  Achieving stability in networks of input-queued switches , 2003, TNET.

[37]  Sem C. Borst,et al.  Reduced-Load Equivalence and Induced Burstiness in GPS Queues with Long-Tailed Traffic Flows , 2003, Queueing Syst. Theory Appl..

[38]  Sem C. Borst,et al.  The impact of the service discipline on delay asymptotics , 2003, Perform. Evaluation.

[39]  F. Baccelli,et al.  Moments and tails in monotone-separable stochastic networks , 2004, math/0405281.

[40]  A. Stolyar MaxWeight scheduling in a generalized switch: State space collapse and workload minimization in heavy traffic , 2004 .

[41]  Sem C. Borst,et al.  Tail asymptotics for discriminatory processor-sharing queues with heavy-tailed service requirements , 2005, Perform. Evaluation.

[42]  Sem C. Borst,et al.  Fluid Queues with Heavy-Tailed M/G/ Input , 2005, Math. Oper. Res..

[43]  F. Baccelli,et al.  Tails in generalized Jackson networks with subexponential service-time distributions , 2005, Journal of Applied Probability.

[44]  A. Gut Probability: A Graduate Course , 2005 .

[45]  J. G. Dai,et al.  Maximum Pressure Policies in Stochastic Processing Networks , 2005, Oper. Res..

[46]  Gennady Samorodnitsky,et al.  Limit Behavior of Fluid Queues and Networks , 2005, Oper. Res..

[47]  R. Núñez Queija,et al.  Discriminatory processor sharing revisited , 2005, Proceedings IEEE 24th Annual Joint Conference of the IEEE Computer and Communications Societies..

[48]  Sem C. Borst,et al.  Sojourn time asymptotics in processor-sharing queues , 2006, Queueing Syst. Theory Appl..

[49]  Leandros Tassiulas,et al.  Resource Allocation and Cross-Layer Control in Wireless Networks , 2006, Found. Trends Netw..

[50]  Bert Zwart,et al.  Tails in scheduling , 2007, PERV.

[51]  Eytan Modiano,et al.  Optimal Transmission Scheduling in Symmetric Communication Models With Intermittent Connectivity , 2007, IEEE Transactions on Information Theory.

[52]  Chita R. Das,et al.  Characterizing Network Traffic in a Cluster-based, Multi-tier Data Center , 2007, 27th International Conference on Distributed Computing Systems (ICDCS '07).

[53]  Vijay G. Subramanian,et al.  Large deviations of max-weight scheduling policies on convex rate regions , 2008, 2008 Information Theory and Applications Workshop.

[54]  D. Down,et al.  Stability of Queueing Networks , 1994 .

[55]  Michael J. Neely,et al.  Order Optimal Delay for Opportunistic Scheduling in Multi-User Wireless Uplinks and Downlinks , 2008, IEEE/ACM Transactions on Networking.

[56]  Marc Lelarge Asymptotic behavior of generalized processor sharing queues under subexponential assumptions , 2009, Queueing Syst. Theory Appl..

[57]  D. Gamarnik,et al.  ON DECIDING STABILITY OF MULTICLASS QUEUEING NETWORKS UNDER BUFFER PRIORITY SCHEDULING POLICIES , 2007, 0708.1034.

[58]  H. Vincent Poor,et al.  Queue Back-Pressure Random Access in Multihop Wireless Networks: Optimality and Stability , 2009, IEEE Transactions on Information Theory.

[59]  John N. Tsitsiklis,et al.  Qualitative properties of α-weighted scheduling policies , 2010, SIGMETRICS '10.

[60]  Devavrat Shah,et al.  Switched networks with maximum weight policies: Fluid approximation and multiplicative state space collapse , 2010, ArXiv.

[61]  Lei Ying,et al.  On Scheduling for Minimizing End-to-End Buffer Usage over Multihop Wireless Networks , 2010, 2010 Proceedings IEEE INFOCOM.

[62]  Ness B. Shroff,et al.  Delay-based Back-Pressure scheduling in multi-hop wireless networks , 2011, INFOCOM.

[63]  Ness B. Shroff,et al.  Delay-Based Back-Pressure Scheduling in Multihop Wireless Networks , 2011, IEEE/ACM Transactions on Networking.

[64]  Mihalis G. Markakis,et al.  Queue length asymptotics for generalized max-weight scheduling in the presence of heavy-tailed traffic , 2011, INFOCOM.

[65]  Alexander L. Stolyar,et al.  A Novel Architecture for Reduction of Delay and Queueing Structure Complexity in the Back-Pressure Algorithm , 2011, IEEE/ACM Transactions on Networking.

[66]  R. Srikant,et al.  Stochastic models of load balancing and scheduling in cloud computing clusters , 2012, 2012 Proceedings IEEE INFOCOM.

[67]  Sergey Foss,et al.  On Large Delays in Multi-Server Queues with Heavy Tails , 2011, Math. Oper. Res..

[68]  Ness B. Shroff,et al.  Throughput-Optimal Scheduling in Multihop Wireless Networks Without Per-Flow Information , 2013, IEEE/ACM Transactions on Networking.

[69]  Mihalis G. Markakis Scheduling in Switched Queueing Networks with Heavy-Tailed Trac , 2013 .

[70]  Delay Stability of Back-Pressure Policies in the Presence of Heavy-Tailed Traffic , 2016, IEEE/ACM Transactions on Networking.

[71]  Mihalis G. Markakis,et al.  Delay stability of back-pressure policies in the presence of heavy-tailed traffic , 2014 .

[72]  Mihalis G. Markakis,et al.  Max-Weight Scheduling in Queueing Networks With Heavy-Tailed Traffic , 2011, IEEE/ACM Transactions on Networking.