1 / 13

KPC-Toolbox Demonstration

KPC-Toolbox Demonstration. Eddy Zheng Zhang, Giuliano Casale, Evgenia Smirni Computer Science Department College of William & Mary. What is KPC-Toolbox for?. KPC-Toolbox: MATLAB toolbox Workload Traces  Markovian Arrival Process (MAP) Why MAP? Very versatile

ugo
Télécharger la présentation

KPC-Toolbox Demonstration

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. KPC-ToolboxDemonstration Eddy Zheng Zhang, Giuliano Casale, Evgenia Smirni Computer Science Department College of William & Mary

  2. What is KPC-Toolbox for? • KPC-Toolbox: MATLAB toolbox • Workload Traces  Markovian Arrival Process (MAP) • Why MAP? • Very versatile • High variability & temporal dependence in Time Series • Easily incorporated into queuing models • Friendly Interface • Departure from previous Markovian fitting tools • Fit the automatically (no manual tuning)

  3. User Interface • Requirement: Matlab installed • Input • A trace of inter-event times • Or a file that already stores the statistics of the trace • E.g., a file stores the moments, autocorrelations and etc • Help Information • Type “help FunctionName”, • E.g., “help map_kpcfit” • Website Keeps Up-To-Date Tool version • http://www.cs.wm.edu/MAPQN/kpctoolbox.html

  4. A Simple Example of MAP‏ • Two state jumps c -a-c Background Jumps D0 = Jumps With Arrivals d -b-d c 2 1 b a a 0 D1 = d b 0 Arrivals: Time: I1 I2 I3

  5. Challenges • How large is the MAP? • MAP(n): determine n? • Which trace descriptors are important? • Literature: Moments of interval times, lag-1 autocorrelation • But, for long range dependent traces? • Need temporal dependence descriptors • MAP Parameterization • Construct MAP(n) with matrices D0 and D1 (2n2– n entries)

  6. 1 2 Example: Important Trace Statistics • Seagate Web Server Trace Queue Prediction, 80% Utilization Fit With MAP(2) First, second, third moment and lag-1 autocorrelation accurately fit The queuing prediction ability is not satisfactory!

  7. 1 13 ……… ……… 2 14 ……… 3 15 4 16 Example: Higher Order Statistics Matter A grid of joint moments and a sequence of autocorrelations fitted, E[XiXi+kXi+k+h] • Seagate Web Server Trace Queuing Prediction, 80% Utilization Fit with MAP(16) Much Better Results!

  8. Fitting Guidelines • Higher Order Correlations V.S. Moments • Correlations capture sequence in the time series • Correlations are very important • Summary: • Matching up to the first three moments is sufficient • Matching higher order correlations with priority Ref: "KPC-Toolbox: Simple Yet Effective Trace Fitting Using Markovian Arrival Processes", G. Casale, E.Z. Zhang, E. Smirni, to appear in QEST’08

  9. Challenge (1): Determine MAP Size • Definition: •  lag-k ACF coefficient • MAP(n) Property: • Linear Recursive Relationship of n consecutive ACF coeffs • BIC Size Selection: • Linear regression model on estimated ACF coeffs • BIC value assesses goodness of model size MAP(8) MAP(16) MAP(32)

  10. Challenge (2): Trace Descriptor Matching • Kronecker Product Composition (KPC) • KPC Properties: Composition of Statistics • Moments are composed from moments of small MAPs • MAP Parameterization by KPC to Match • Mean and SCV Exactly • Higher order correlations as Close as Possible

  11. KPC Tool Overview Moments ACF Correlations …… Size of MAP N Trace Size Selection Extract Statistics Optimization J = log2N MAP(2)s MAP(2) MAP(2) MAP(2) MAP(2) …… KPC MAP(N) This work is supported by NSF grants ITR-0428330 and CNS-0720699

  12. Thank you! 

  13. Appendix • What are higher order correlations? • Joint moments of a sequence of inter-arrival times in the time series • Which higher order correlations to fit in KPC? • E[XiXi+jXi+j+k], where i can be arbitrary without loss of generality, and [j,k] chose from a grid of values • E.g., [10 100 1000 10000] × [10 100 1000 10000] = {[10,10], [10,100], [10,10000], …}

More Related