Size-Based Scheduling Policies with Inaccurate Scheduling Information

1
Size-Based Scheduling Policies with Inaccurate
Scheduling Information

• Dong Lu, Huanyuan Sheng, Peter A. Dinda
• Prescience Lab, Dept. of Computer Science
• Dept. of Industrial Engineering Management
  Science
• Northwestern University
• Evanston, IL 60201 USA

2
Outline

• Review of size-based scheduling
• Motivation
• Simulation Setup
• Simulation Results
• New applications
• Research Summary by subjects

3
Non-size-based scheduling

• FCFS, PS, etc.
• FCFS First Come First Serve
• Intuitive
• Easiest to implement
• PS Processor Sharing
• Fair all jobs accept equal resources
• Also easy to implement

Problem Unaware of job size information, which
results in big mean response time
4
Review of size-based scheduling

• SRPT, FSP, etc.
• Utilize the job size (processing time, service
  time) information for scheduling
• Optimal in mean response time
• Fair?
• Easy to implement?

We use Job Size to refer to the Processing Time
(Service Time) of the job
5
Shortest Remaining Processing Time (SRPT)

• Always serve the job with minimum remaining
  processing time first, Preemptive scheduling
• Yields minimum mean response time Schrage,
  Operations Research, 1968
• Performance gains of SRPT over PS do not usually
  come at the expense of large jobs, in other
  words, it is Fair for heavy-tail job size
  distribution Bansal and Harchol-Balter,
  Sigmetrics 01
• Easy to implement?
• With accurate a priori job size information, YES
• Otherwise, NO

6
Fair Sojourn Protocol (FSP)

• Combined SRPT with PS, preemptive scheduling
• Mean response time is close to that of SRPT and
  more fair than PS Friedman, et al, Sigmetrics
  03
• Easy to implement?
• With accurate a priori job size information, YES
• Otherwise, NO

7
Motivation

• Size-based scheduling requires accurate knowledge
  of job sizes

• In practice, a priori job size information is
  not always
• available

• All the previous work assumes perfect knowledge
  of job sizes a priori

• How does performance depend on
• quality of job size information?

8
Correlation
We study the performance of Size-based
schedulers as a function of the correlation
coefficient (Pearsons R) between actual job
sizes and estimated job sizes.
9
Outline

• Review of size-based scheduling
• Motivation
• Simulation Setup
• Simulation Results
• New applications
• Research Summary by subjects

10
Simulation Setup Trace generator
Correlation (Pearsons R)
Distribution A
Distribution B
Trace Generator

• X Y
• 100
• 300
• . .
• . .
• . .

• Correlated random pairs of X and Y
• X has distribution A
• Y has distribution B
• X and Y are correlated to R

11
Simulation Setup Trace generator

• Algorithm Normal-To-Anything
• First developed by Cario and Nelson, on INFORMS
  Journal on Computing 10, 1 (1998).
• We simplified the algorithm and first introduced
  it into the simulation studies of computer systems

12
Scatter plot of example traces
Y
Y
X
X
R0.78
R0.13
13
Simulation Setup Performance metrics

• Performance metrics
• Mean response time Sojourn time, Turn-around
  time
• Slowdown the ratio of response time to its size.
  Fairness metric

14
Simulation Setup Simulator

• Simulator
• Written in C
• Supports M/G/1 and G/G/n/m queuing model
• Simulator validation
• Littles law
• Repeat the simulations in the FSP paper
  Friedman, et al, Sigmetrics 03
• Compare with available theoretical results
  Bansal and Harchol-Balter, Sigmetrics 01

15
Simulation Setup Scheduling Policies

• PS Processor sharing
• Size-based scheduling policies
• SRPT Ideal SRPT scheduler
• SRPT-E SRPT scheduler using estimated job size
• FSP Ideal Fair Sojourn Protocol
• FSP-E FSP scheduler using estimated job size

Each simulation is repeated 20 times and we
present the average
16
Outline

• Review of size-based scheduling
• Motivation
• Simulation Setup
• Simulation Results
• New applications
• Research Summary by subjects

17
Simulation Results Mean response time
18
Simulation Results Slowdown (R0.0224)
19
Simulation Results Slowdown (R0.239)
20
Simulation Results Slowdown (R0.4022)
21
Simulation Results Slowdown (R0.5366)
22
Simulation Results Slowdown (R0.7322)
23
Simulation Results Slowdown (R0.9779)
24
Simulation Results Conclusions

• Performance heavily depends on correlation
• SRPT-E and FSP-E can outperform PS given an
  effective job size estimator
• Crossover point of performance metrics is a
  function of correlation
• Also of job size distributions (See TR
  NWU-CS-04-33)

25
Outline

• Review of size-based scheduling
• Motivation
• Simulation Setup
• Simulation Results
• New applications
• Research Summary by subjects

26
New Applications Web server scheduling (TR
NWU-CS-04-33)

• Is file size a good estimator of a jobs service
  time (processing time)? Not Really (R ? 0.14)

File Size
Service time (wall clock time)
27
New Applications Web server scheduling

• Domain-based estimator much more accurate
  prediction of the service time at low overhead

28
New Applications P2P server side scheduling (LCR
04)

• Server side of current file sharing P2P
  applications superficially similar to web server
• Both send back files upon requests.
• However, P2P application cant even know the file
  size accurately a priori
• Partial downloads
• Our ongoing work shows that SRPT-E performs well
  using our time-series based job size estimators.
  

29
New Applications Network backup system scheduling

• Incremental backup copies only the files that
  have been created or modified since a previous
  backup
• With Incremental backup, the actual job sizes is
  difficult to know until the backup finishes
• We believe that SRPT-E or FSP-E can be applied
  with time series based job size predictors

30
Summary

• Performance of size-based scheduling policies
  depends on correlation between size estimates and
  actual sizes
• Fairness, mean response time, etc.
• Estimator must preserve ordering of job sizes for
  high performance
• Performance degrades as correlation degrades
• Effective new estimators for Web and P2P

31
For MoreInformation

• Prescience Laboratory
• http//plab.cs.northwestern.edu
• Home Page of Dong Lu
• http//www.cs.northwestern.edu/donglu/

32
Outline

• Review of size-based scheduling
• Motivation
• Simulation Setup
• Simulation Results
• New applications
• Research Summary by subjects

33
Research Summary by subjects

• Grid Computing
• Internet Measurement and Prediction
• Queuing and Scheduling
• Fat-Tree Based End System Multicast
• Wireless Ad Hoc Networks
• Incentivized Protocol Design for Peer-to-Peer
  Systems
• Parallel Computing

34
Grid Computing

• Dong Lu, Peter Dinda, GridG Generating
  Realistic Computational Grids, ACM SIGMETRICS
  Performance Evaluation Review (Per), Volume 30,
  Number 4, 2003.
• Dong Lu, Peter Dinda, Synthesizing Realistic
  Computational Grids, Proceedings of the 15th
  ACM/IEEE Supercomputing (SC 2003), Phoenix, AZ,
  November 2003.
• Peter Dinda, Dong Lu, Nondeterministic queries
  in a relational Grid information service,
  Proceedings of the 15th ACM/IEEE Supercomputing
  (SC 2003), Phoenix, AZ, November 2003.
• Dong Lu, Peter Dinda, Jason Skicewicz Scoped and
  Approximated queries in a relational Grid
  Information Service, Proceedings of 4th IEEE/ACM
  International Workshop on Grid Computing (Grid
  2003), November, 2003, Phoenix, AZ.
• Bin Lin, Peter Dinda, Dong Lu, User-driven
  Scheduling of Interactive Virtual Machines,
  Proceedings of Grid 2004, PITTSBURGH, PA,
  November, 2004.

35
Internet Measurement and Prediction

• Dong Lu, Y. Qiao, Peter Dinda, and F. Bustamante,
  Characterizing and Predicting TCP Throughput on
  the Wide Area Network, Proceedings of the 25th
  IEEE International Conference on Distributed
  Computing Systems (ICDCS 2005), June 2005,
  Columbus, Ohio. To appear.
• Dong Lu, Yi Qiao, Peter Dinda, Fabian Bustamante,
  Modeling and Taming Parallel TCP on the Wide
  Area Network, Proceedings of the 19th IEEE
  International Parallel and Distributed Processing
  Symposium (IPDPS 2005), April 4-8, 2005, Denver,
  Colorado.

36
Queuing and Scheduling

• Dong Lu, Huanyuan Sheng, Peter Dinda, Size-Based
  Scheduling Policies with Inaccurate Scheduling
  Information, Proceedings of MASCOTS 2004,
  October 2004, Volendam, The Netherlands.
• Dong Lu, Peter Dinda, Yi Qiao, Huanyuan Sheng,
  Fabian Bustamante, Applications of SRPT
  Scheduling with Inaccurate Scheduling
  Information, (short paper) Proceedings of
  MASCOTS 2004, October 2004, Volendam, The
  Netherlands.
• Yi Qiao, Dong Lu, Fabian Bustamante, Peter Dinda,
  Looking at the Server-Side of Peer-to-Peer
  Systems, Proceedings of the 7th ACM Workshop on
  Languages, Compilers, and Run-time Systems for
  Scalable Computers (LCR 2004), October 2004,
  Houston, Texas.
• Dong Lu, Huanyuan Sheng, and Peter Dinda,
  Effects and Implications of File Size/Service
  Time Correlation on Web Server Scheduling
  Policies, Technical Report NWU-CS-04-33,
  Department of Computer Science, Northwestern
  University, April, 2004. In Submission.

37
Fat-Tree Based End System Multicast FatNemo

• Stefan Birrer, Dong Lu, Fabian Bustamante, Yi
  Qiao, Peter Dinda, FatNemo Building a Resilient
  Multi-Source Multicast Fat-Tree, Proceedings of
  the Ninth International Workshop on Web Content
  Caching and Distribution (WCW 2004), October
  2004, Beijing, China. Also appeared in LNCS, Vol.
  3293/2004, pp. 182-196.
• Long version in submission

38
Wireless Ad Hoc Networks

• Dong Lu, Haitao Wu, Qian Zhang, Wenwu Zhu, PARS
  Stimulating Cooperation for Power-Aware Routing
  in Ad-Hoc Networks. Proceedings of the 40th IEEE
  International Conference on Communications (ICC
  2005), May 2005, Seoul, Korea. To appear.

39
Incentivized Protocol Design for Peer-to-Peer
Systems

• Dong Lu, Yi Qiao, Peter Dinda, Fabian Bustamante,
  MultiTorrents Bandwidth Optimized Hybrid
  Multicast for Incentivized P2P File Sharing. In
  Submission.

40
Parallel Computing

• Dong Lu, Peter Dinda Virtualized Audio A Highly
  adaptive Interactive High Performance Computing
  Application, Proceedings of the 6th Workshop on
  Languages, Compilers, and Run-time Systems for
  Scalable Computers (LCR 2002), Washington, DC,
  2002. Also to appear in LNCS.