Tight Bounds for Online Vector Scheduling

Modern data centers face a key challenge of effectively serving user requests that arrive online. Such requests are inherently multi-dimensional and characterized by demand vectors over multiple resources such as processor cycles, storage space, and network bandwidth. Typically, different resources require different objectives to be optimized, and Lr norms of loads are among the most popular objectives considered. Furthermore, the server clusters are also often heterogeneous making the scheduling problem more challenging. To address these problems, we consider the online vector scheduling problem in this paper. Introduced by Chekuri and Khanna (SIAM J. of Comp. 2006), vector scheduling is a generalization of classical load balancing, where every job has a vector load instead of a scalar load. The scalar problem, introduced by Graham in 1966, and its many variants (identical and unrelated machines, makespan and Lr-norm optimization, offline and online jobs, etc.) have been extensively studied over the last 50 years. In this paper, we resolve the online complexity of the vector scheduling problem and its important generalizations - for all Lr norms and in both the identical and unrelated machines settings. Our main results are: · For identical machines, we show that the optimal competitive ratio is Θ(log d/ log log d) by giving an online lower bound and an algorithm with an asymptotically matching competitive ratio. The lower bound is technically challenging, and is obtained via an online lower bound for the minimum mono-chromatic clique problem using a novel online coloring game and randomized coding scheme. Our techniques also extend to asymptotically tight upper and lower bounds for general Lr norms. · For unrelated machines, we show that the optimal competitive ratio is Θ(log m + log d) by giving an online lower bound that matches a previously known upper bound. Unlike identical machines, however, extending these results, particularly the upper bound, to general Lr norms requires new ideas. In particular, we use a carefully constructed potential function that balances the individual Lr objectives with the overall (convexified) min-max objective to guide the online algorithm and track the changes in potential to bound the competitive ratio.

[1]  Jirí Sgall,et al.  On-line Scheduling , 1996, Online Algorithms.

[2]  Csaba D. Tóth,et al.  Selfish Load Balancing and Atomic Congestion Games , 2004, SPAA '04.

[3]  Yossi Azar,et al.  Load Balancing in the L P Norm , 2012 .

[4]  Amos Fiat,et al.  On-line routing of virtual circuits with applications to load balancing and machine scheduling , 1997, JACM.

[5]  Allan Borodin,et al.  Online computation and competitive analysis , 1998 .

[6]  Yossi Azar,et al.  Ancient and New Algorithms for Load Balancing in the lp Norm , 1998, SODA '98.

[7]  Csaba D. Tóth,et al.  Selfish load balancing and atomic congestion games , 2004, SPAA '04.

[8]  Erik D. Demaine,et al.  Energy-Efficient Algorithms , 2016, ITCS.

[9]  Yossi Azar,et al.  Tight bounds for online vector bin packing , 2013, STOC '13.

[10]  György Turán,et al.  On the performance of on-line algorithms for partition problems , 1989, Acta Cybern..

[11]  Ioannis Caragiannis,et al.  Better bounds for online load balancing on unrelated machines , 2008, SODA '08.

[12]  Eric Torng,et al.  Generating adversaries for request-answer games , 2000, SODA '00.

[13]  Adam Meyerson,et al.  Online Multidimensional Load Balancing , 2013, APPROX-RANDOM.

[14]  Ioannis Caragiannis,et al.  Tight Bounds for Selfish and Greedy Load Balancing , 2006, Algorithmica.

[15]  Reto Spöhel,et al.  Coloring random graphs online without creating monochromatic subgraphs , 2011, SODA '11.

[16]  Ajay Gulati.  VMware distributed resource Management : design , Implementation , and lessons learned , 2022 .

[17]  Joseph Naor,et al.  Online Primal-Dual Algorithms for Covering and Packing Problems , 2005, ESA.

[18]  Susanne Albers,et al.  Better bounds for online scheduling , 1997, STOC '97.

[19]  Scott Shenker,et al.  Choosy: max-min fair sharing for datacenter jobs with constraints , 2013, EuroSys '13.

[20]  Calton Pu,et al.  Generating Adaptation Policies for Multi-tier Applications in Consolidated Server Environments , 2008, 2008 International Conference on Autonomic Computing.

[21]  Amos Fiat,et al.  New algorithms for an ancient scheduling problem , 1992, STOC '92.

[22]  B. Awerbuch,et al.  Load Balancing in the Lp Norm , 1995, FOCS 1995.

[23]  Kirk Pruhs,et al.  Online scheduling , 2003 .

[24]  Jing Tao,et al.  An Energy-Aware Task Scheduling Algorithm for a Heterogeneous Data Center , 2013, 2013 12th IEEE International Conference on Trust, Security and Privacy in Computing and Communications.

[25]  Srikanth Kandula,et al.  Multi-resource packing for cluster schedulers , 2014, SIGCOMM.

[26]  Yuval Rabani,et al.  A Better Lower Bound for On-Line Scheduling , 1994, Inf. Process. Lett..

[27]  T. N. Vijaykumar,et al.  Tarazu: optimizing MapReduce on heterogeneous clusters , 2012, ASPLOS XVII.

[28]  Russ Bubley,et al.  Randomized algorithms , 1995, CSUR.

[29]  Marek Karpinski,et al.  On-Line Load Balancing for Related Machines , 1997, J. Algorithms.

[30]  FiatAmos,et al.  On-line routing of virtual circuits with applications to load balancing and machine scheduling , 1997 .

[31]  Chak-Kuen Wong,et al.  Worst-Case Analysis of a Placement Algorithm Related to Storage Allocation , 1975, SIAM J. Comput..

[32]  Vahab S. Mirrokni,et al.  Convergence and approximation in potential games , 2006, Theor. Comput. Sci..

[33]  Vahab S. Mirrokni,et al.  Convergence and approximation in potential games , 2012, Theor. Comput. Sci..

[34]  David R. Karger,et al.  A better algorithm for an ancient scheduling problem , 1994, SODA '94.

[35]  Ronald L. Graham,et al.  Bounds for certain multiprocessing anomalies , 1966 .

[36]  Leah Epstein,et al.  Multidimensional Packing Problems , 2018, Handbook of Approximation Algorithms and Metaheuristics.

[37]  Benjamin Hindman,et al.  Dominant Resource Fairness: Fair Allocation of Multiple Resource Types , 2011, NSDI.

[38]  Yossi Azar,et al.  The competitiveness of on-line assignments , 1992, SODA '92.

[39]  Edward G. Coffman,et al.  Record Allocation for Minimizing Expected Retrieval Costs on Drum-Like Storage Devices , 1976, J. ACM.

[40]  Yossi Azar.  On-line Load Balancing , 1996, Online Algorithms.

[41]  Magnús M. Halldórsson,et al.  Lower bounds for on-line graph coloring , 1992, SODA '92.

[42]  F. Frances Yao,et al.  A scheduling model for reduced CPU energy , 1995, Proceedings of IEEE 36th Annual Foundations of Computer Science.

[43]  W. Campbell,et al.  THE UNIVERSITY OF TEXAS AT DALLAS , 2004 .

[44]  Randy H. Katz,et al.  Heterogeneity-Aware Resource Allocation and Scheduling in the Cloud , 2011, HotCloud.

[45]  Rudolf Fleischer,et al.  Online Scheduling Revisited , 2000, ESA.