Mixed Criteria Packet Scheduling

Packet scheduling in networks with quality of service constraints has been extensively studied as a single criterion scheduling problem. The assumption underlying single criterion packet scheduling is that the value of all packets can be normalized to a single scale, even in cases when packets have different requirements. We demonstrate that this approach can lead to inefficient utilization of network resources. To improve network efficiency, we model packet scheduling as a mixed criteriascheduling problem where there are two distinct sets of jobs: deadline jobs which represent real-time packets in a network and flow jobs which represent other packets in the network. As the names imply, the jobs in these two sets differ by the criteria associated with them.For this problem, the flow jobs are scheduled to minimize the sum of their flow times, and the deadline jobs are scheduled to maximize the value of jobs that complete by their deadlines. We demonstrate that even when there is only a single deadline job, this mixed criteria scheduling problem is NP-Complete. We give a polynomial time optimal algorithm Slacker for the variant where all jobs have unit size and the value of deadline jobs processed by the deadline must be maximized. Given this constraint Slacker minimizes the total flow time. Furthermore, we show that online Slacker is optimal for flow time while being 2-competitive with respect to the deadline jobs when compared to an optimal algorithm like Slacker that maximizes the value of deadline jobs.

[1]  Eric Torng,et al.  Existence theorems, lower bounds and algorithms for scheduling to meet two objectives , 2002, SODA '02.

[2]  Marek Chrobak,et al.  Improved Online Algorithms for Buffer Management in QoS Switches , 2004, ESA.

[3]  Michael Pinedo,et al.  Scheduling n Independent Jobs on m Uniform Machines with both Flowtime and Makespan Objectives: A Parametric Analysis , 1995, INFORMS J. Comput..

[4]  David S. Johnson,et al.  Computers and Intractability: A Guide to the Theory of NP-Completeness , 1978 .

[5]  Jonathan Cole Smith,et al.  A Multiple-Criterion Model for Machine Scheduling , 2003, J. Sched..

[6]  Marco Spuri,et al.  Scheduling aperiodic tasks in dynamic priority systems , 1996, Real-Time Systems.

[7]  Nodari Vakhania,et al.  Preemptive scheduling of equal-length jobs to maximize weighted throughput , 2002, Oper. Res. Lett..

[8]  Han Hoogeveen,et al.  Single-Machine Scheduling to Minimize a Function of Two or Three Maximum Cost Criteria , 1996, J. Algorithms.

[9]  Clifford Stein,et al.  On the existence of schedules that are near-optimal for both makespan and total weighted completion time , 1997, Oper. Res. Lett..

[10]  Fei Li,et al.  An optimal online algorithm for packet scheduling with agreeable deadlines , 2005, SODA '05.

[11]  Yishay Mansour,et al.  Improved Competitive Guarantees for QoS Buffering , 2003, Algorithmica.

[12]  Baruch Schieber,et al.  Further Improvements in Competitive Guarantees for QoS Buffering , 2004, ICALP.

[13]  Yuval Rabani,et al.  Linear Programming , 2007, Handbook of Approximation Algorithms and Metaheuristics.

[14]  Robin Milner,et al.  On Observing Nondeterminism and Concurrency , 1980, ICALP.

[15]  Maryline Chetto,et al.  Some Results of the Earliest Deadline Scheduling Algorithm , 1989, IEEE Transactions on Software Engineering.

[16]  Susanne Albers,et al.  Algorithms – ESA 2004 , 2004, Lecture Notes in Computer Science.

[17]  Yishay Mansour,et al.  Competitive queue policies for differentiated services , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[18]  Alessandro Agnetis,et al.  Scheduling Problems with Two Competing Agents , 2004, Oper. Res..

[19]  Yishay Mansour,et al.  Competitive queueing policies for QoS switches , 2003, SODA '03.

[20]  Karl Ernst Osthaus.  Van de Velde , 1920 .

[21]  Torben Hagerup,et al.  Algorithm Theory - SWAT 2004 , 2004, Lecture Notes in Computer Science.

[22]  Marek Chrobak,et al.  Online competitive algorithms for maximizing weighted throughput of unit jobs , 2006, J. Discrete Algorithms.

[23]  Bruce Hajek.  On the Competitiveness of On-Line Scheduling of Unit-Length Packets with Hard Deadlines in Slotted Time , 2001 .

[24]  Boaz Patt-Shamir,et al.  Buffer Overflow Management in QoS Switches , 2004, SIAM J. Comput..

[25]  John P. Lehoczky,et al.  An optimal algorithm for scheduling soft-aperiodic tasks in fixed-priority preemptive systems , 1992, [1992] Proceedings Real-Time Systems Symposium.

[26]  Kirk Pruhs,et al.  Getting the Best Response for Your Erg , 2004, SWAT.

[27]  Han Hoogeveen,et al.  Minimizing total completion time and maximum cost simultaneously is solvable in polynomial time , 1994, Oper. Res. Lett..

[28]  Clifford Stein,et al.  Improved bicriteria existence theorems for scheduling , 2002, SODA '99.

[29]  Ludo Gelders,et al.  Solving a bicriterion scheduling problem , 1980 .