Information and Memory in Dynamic Resource Allocation

We propose a general framework, dubbed Stochastic Processing under Imperfect Information (SPII), to study the impact of information constraints and memories on dynamic resource allocation. The framework involves a Stochastic Processing Network (SPN) scheduling problem in which the scheduler may access the system state only through a noisy channel, and resource allocation decisions must be carried out through the interaction between an encoding policy (who observes the state) and allocation policy (who chooses the allocation). Applications in the management of large-scale data centers and human-in-the-loop service systems are among our chief motivations. We quantify the degree to which information constraints reduce the size of the capacity region in general SPNs, and how such reduction depends on the amount of memories available to the encoding and allocation policies. Using a novel metric, capacity factor, our main theorem characterizes the reduction in capacity region (under "optimal" policies) for all non-degenerate channels, and across almost all combinations of memory sizes. Notably, the theorem demonstrates, in substantial generality, that (1) the presence of a noisy channel always reduces capacity, (2) more memory for the allocation policy always improves capacity, and (3) more memory for the encoding policy has little to no effect on capacity. Finally, all of our positive (achievability) results are established through constructive, implementable policies.

[1]  Sekhar Tatikonda,et al.  Control under communication constraints , 2004, IEEE Transactions on Automatic Control.

[2]  Ion Stoica,et al.  Efficient Coflow Scheduling Without Prior Knowledge , 2015, SIGCOMM.

[3]  Kimberly A Davis,et al.  Information Loss in Emergency Medical Services Handover of Trauma Patients , 2009, Prehospital emergency care : official journal of the National Association of EMS Physicians and the National Association of State EMS Directors.

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

[5]  J. Harrison Correction: Brownian models of open processing networks: canonical representation of workload , 2000, math/0610352.

[6]  Alexander L. Stolyar Pull-based load distribution among heterogeneous parallel servers: the case of multiple routers , 2017, Queueing Syst. Theory Appl..

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

[8]  John N. Tsitsiklis,et al.  Delay, Memory, and Messaging Tradeoffs in Distributed Service Systems , 2018 .

[9]  James R. Larus,et al.  Join-Idle-Queue: A novel load balancing algorithm for dynamically scalable web services , 2011, Perform. Evaluation.

[10]  J. Michael Harrison,et al.  Stochastic Networks and Activity Analysis , 2002 .

[11]  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.

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

[13]  Sanjay Shakkottai,et al.  Regret of Queueing Bandits , 2016, NIPS.

[14]  Randy H. Katz,et al.  FastLane: making short flows shorter with agile drop notification , 2015, SoCC.

[15]  Madhu Sudan,et al.  Queuing with future information , 2012 .

[16]  Richard H. Rand,et al.  A Stochastic Analysis of Queues with Customer Choice and Delayed Information , 2020, Math. Oper. Res..

[17]  Alexander L. Stolyar,et al.  Scheduling Flexible Servers with Convex Delay Costs: Heavy-Traffic Optimality of the Generalized cµ-Rule , 2004, Oper. Res..

[18]  Elena Yudovina,et al.  Tightness of Invariant Distributions of a Large-scale Flexible Service System Under a Priority Discipline , 2012 .

[19]  Rakesh Gupta Above the Clouds: A View of Cloud Computing , 2012 .

[20]  J. Walrand,et al.  Sufficient conditions for stability of longest-queue-first scheduling: second-order properties using fluid limits , 2006, Advances in Applied Probability.

[21]  J. Michael Harrison,et al.  A Method for Staffing Large Call Centers Based on Stochastic Fluid Models , 2005, Manuf. Serv. Oper. Manag..

[22]  Michael Mitzenmacher,et al.  The Power of Two Choices in Randomized Load Balancing , 2001, IEEE Trans. Parallel Distributed Syst..

[23]  Mark Burgess,et al.  Dynamic pull-based load balancing for autonomic servers , 2008, NOMS 2008 - 2008 IEEE Network Operations and Management Symposium.

[24]  John N. Tsitsiklis,et al.  On the Power of (Even a Little) Resource Pooling , 2012 .

[25]  M. Zaharia,et al.  Above the Clouds : A View of Cloud Computing , 2009 .

[26]  Hong Chen,et al.  Empirical Evaluation of a Queueing Network Model for Semiconductor Wafer Fabrication , 1988, Oper. Res..

[27]  Sanjay Shakkottai,et al.  On Learning the cμ Rule in Single and Parallel Server Networks , 2018, 2018 56th Annual Allerton Conference on Communication, Control, and Computing (Allerton).

[28]  Kuang Xu,et al.  Necessity of Future Information in Admission Control , 2015, Oper. Res..

[29]  D. Pothier,et al.  Pilot study to show the loss of important data in nursing handover. , 2005, British journal of nursing.

[30]  Julia Adler-Milstein,et al.  The Effect of Health Information Technology on Health Care Provider Communication: A Mixed-Method Protocol , 2015, JMIR research protocols.

[31]  David Lipshutz,et al.  Heavy Traffic Limits for Join-the-Shortest-Estimated-Queue Policy Using Delayed Information , 2021, Math. Oper. Res..

[32]  Adam Wierman,et al.  This Paper Is Included in the Proceedings of the 11th Usenix Symposium on Networked Systems Design and Implementation (nsdi '14). Grass: Trimming Stragglers in Approximation Analytics Grass: Trimming Stragglers in Approximation Analytics , 2022 .

[33]  Balaji Prabhakar,et al.  The throughput of data switches with and without speedup , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[34]  R. L. Dobrushin,et al.  Queueing system with selection of the shortest of two queues: an assymptotic approach , 1996 .

[35]  Amy R. Ward,et al.  Blind Fair Routing in Large-Scale Service Systems with Heterogeneous Customers and Servers , 2013, Oper. Res..

[36]  Laurent Massoulié,et al.  Bandwidth sharing and admission control for elastic traffic , 2000, Telecommun. Syst..

[37]  Golshid Baharian,et al.  Stability analysis of parallel server systems under longest queue first , 2011, Math. Methods Oper. Res..

[38]  Bruce Hajek,et al.  Random Processes for Engineers , 2015 .

[39]  Alexander L. Stolyar Pull-based load distribution in large-scale heterogeneous service systems , 2015, Queueing Syst. Theory Appl..

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

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

[42]  Jean C. Walrand,et al.  Robust scheduling for flexible processing networks , 2017, Advances in Applied Probability.

[43]  Anant Sahai,et al.  The Necessity and Sufficiency of Anytime Capacity for Stabilization of a Linear System Over a Noisy Communication Link—Part I: Scalar Systems , 2006, IEEE Transactions on Information Theory.

[44]  Tamer Basar,et al.  Stochastic Networked Control Systems: Stabilization and Optimization under Information Constraints , 2013 .