1 / 27

Towards a Bell-Curve Calculus For e-Science

Towards a Bell-Curve Calculus For e-Science. Lin Yang Supervised by Alan Bundy, Dave Berry and Conrad Hughes. Content. Background Bell-Curve Calculus (BCC) Importance Definition Methodology Result and Analysis Evaluation Future Work and Conclusion. Background. Why QoS Properties

gracie
Download Presentation

Towards a Bell-Curve Calculus For e-Science

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. Towards a Bell-Curve Calculus For e-Science LinYang Supervised by Alan Bundy, Dave Berry and Conrad Hughes

  2. Content Background Bell-CurveCalculus(BCC) Importance Definition Methodology ResultandAnalysis Evaluation Future Work and Conclusion

  3. Background WhyQoSProperties Prediction,descriptionandevaluation Likelihoodofeachdatavalue Aims Define a suitable calculus for runtime Apply it to e-Science workflows

  4. BCC(1)–Importance WhyBellCurve An average case analysis: likely or unlikely Easy to store,calculate and propagate Deal with complex workflows efficiently

  5. BCC(2)–Importance Bell Curve = Normal Distribution Commonly occurs in the real world Evidence Experimental evidence Central Limit Theorem

  6. BCC(3)–Definition BellCurve

  7. BCC(4)–Definition QoSProperty: Runtime FourwaysofcombiningGridServices Sequential Parallel_All Parallel_First Conditional Fourfundamentalcombinationfunctions: sum, max, min & cond

  8. BCC(5)–Definition–FourCombinations Sequential (sum) Parallel_All (max) Parallel_First (min) Conditional (cond)

  9. BCC(6)–Methodology Twoinputbellcurves and Oneoutputbellcurve Thecombinationmethod = ParametersCalculation e.g.and

  10. BCC(7)–Methodology Twomaintasks Tofindasatisfactoryformula foreachcombinationmethod Toevaluateaccuracy and efficiency Agrajag DevelopedbyConradHughes Define classic distribution functions, operations and numeric approximation of function combinations

  11. BCC(8)–Result&Analysis–Max

  12. BCC(9)–Result&Analysis–Max

  13. BCC(10)–Result&Analysis–Refine Perfectparameters DefinedinAgrajag Approximatetheperfectvalues Fixoneparameter Uselinearfunctiontoapproachtheperfectvaluesintermsoftheotherparameterastheasymptote Derivethelinearparameters

  14. BCC(11)–Result&Analysis–Refine Useexponentialcompensation Getexponentialparameters Findtheregularpatternofthelinearandexponentialparametersintermsofthefirstparameter Combineanddescribetheperfectvalues intermsofthetwoBCCparameters

  15. BCC(12)–Result&Analysis–Max–Ref

  16. BCC(13)–Result&Analysis–Max–Ref

  17. BCC (14) – Result & Analysis – Max – Ref

  18. BCC (15) – Result & Analysis – Max -- Ref

  19. BCC (16) – Evaluation Two Methods Comparison with Agrajag Accuracy Efficiency Apply to use cases Brain atlas Extended workflows

  20. BCC (17) – Evaluation

  21. BCC (18) – Evaluation – Accuracy

  22. BCC (19) – Evaluation – Efficiency

  23. BCC (20) – Evaluation – Extended Use Case sum0 max0 sum1 max2 cond1 sum4 sum2 +perc*softmean0 +(1-perc)*softmean1 max1 sum3 warp+reslice sum8 sum5 min0 sum9 sum6 min1 sum10 sum7 +convert +slicer

  24. BCC (21) – Evaluation – Extended Use Case

  25. Future Work (1) Embed In Frameworks More Evaluation More complex workflows Real data More Calculi e.g. log-normal distribution More QoS Properties e.g. accuracy and reliability

  26. Future Work (2) Extended TwelveFundamentalCombinationFunctions

  27. TheEnd

More Related