System and method of cost oriented software profiling

• US 8,578,348 B2
• Filed: 09/01/2011
• Issued: 11/05/2013
• Est. Priority Date: 09/02/2010
• Status: Active Grant

• Current Assignee: Code Value Ltd.
• Original Assignee: Code Value Ltd.
• Inventors: Fliess, Alon Mordechai, Fliess, Erez, Reuben, Joshua, Etchin, Aaron
• Primary Examiner(s): Khatri, Anil
• Application Number:
  US13/223,498
• Publication Number:
  US 20120060142A1
• Time in Patent Office: 796 Days
• Field of Search: 717127-130, 717134-136, 717151-152, 706 45- 46
• US Class Current: 717/135
• CPC Class Codes: G06F 11/3428 Benchmarking
• CPC Class Codes: G06F 11/3457 Performance evaluation by simulation
• CPC Class Codes: G06F 2201/865 Monitoring of software
• CPC Class Codes: G06F 8/70 Software maintenance or management

Abstract

A cost oriented profiler (COP) mechanism that analyzes the behavior of input application source code with regard to the software total cost of ownership (TCO). The cost analysis tool provided by the mechanism analyzes the behavior of the source code and generates a cost report with indications as to the portions of the source code that have the most impact on the TCO of the application. Based on simulations and by comparing multiple versions of the source code, the COP mechanism determines if a particular change to the source code will increase or decrease software TCO. Behavior analysis, including static and dynamic analysis of the source code, is used to generate one or more code recommendations to reduce the TCO.

Claims

1. A method for use on a computer of cost oriented profiling of input software application code executing in a cloud computing environment to generate an economic cost estimate therefrom, said method comprising:

performing, on said computer, a static analysis of said input application code with regard to the total economic cost of ownership (TCO) using a static analysis engine and associated rules to find economically costly code therein with first indications as to the portions of said software application code that have the most impact on the TCO thereof;

performing, on said computer, a simulation of and dynamic analysis of said input application code with regard to the TCO using a dynamic analysis engine and associated rules to find economically costly code therein; and

performing, on said computer, a dynamic analysis of said input application utilizing a dynamic analysis engine and associated rules to find economically costly code therein with second indications as to the portions of said software application code that have the most impact on the TCO thereof;

said dynamic analysis including simulating, on said computer, the usage of said input application code utilizing a cost oriented simulator to generate a dynamic analysis with third indications of any costly code found thereby providing a measure to a user of the total cost of ownership of said input application code; and

generating, on said computer, one or more economic cost estimate reports based on the results of said behavior static analysis, simulation and dynamic analysis, wherein said economic cost estimate reports include those portions of the application code having the most impact on the TCO, said first and second indications and one or more recommendations for reducing the TCO.

2. The method according to claim 1, wherein analyzing the behavior of said input application code comprises determining one or more lines of code, blocks or functions of said input application code that can potentially be optimized for cost.

3. The method according to claim 1, wherein analyzing the behavior of said input application code comprises determining whether storage utilization costs can be optimized according to the usage of said input application code.

4. The method according to claim 1, further comprising simulating the usage of said input application code utilizing a cost oriented simulator, said simulator operative to generate a dynamic analysis indicating any costly code found thereby providing a measure to a user of the total cost of ownership of said input application code comparing multiple versions of said source code and determining if a particular change to said source code increases or decreases software TCO.

5. The method according to claim 4, wherein said cost oriented simulator comprises tracking and recording user input to said input application thereby enabling recorded scenarios to be repeated.

6. The method according to claim 4, wherein said cost oriented simulator comprises enabling said user to define and record a set of application messages that trigger different scenarios within said input application.

7. The method according to claim 4, wherein said cost oriented simulator comprises providing a scheduling interface wherein scenarios can be defined while controlling or configuring one or more settings.

8. The method according to claim 1, wherein analyzing the behavior of said input application code comprises analyzing said input application code for violations of cost related programming rules and conventions related to performance and service costs.

9. The method according to claim 1, wherein analyzing the behavior of said input application code comprises comparative profiling of a cloud storage service mechanism used by said input application code.

10. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises finding specific storage access smells and comparing costs of a plurality of cloud storage options.

11. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises aggregating one or more relevant monitored events for each instance of cloud application storage for subsequent reporting to a user.

12. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises filtering event traces by a relevant subset of event data.

13. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises replaying captured event data against a local instance of said cloud storage service thereby re-executing saved events as they occurred originally.

14. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises monitoring an instance of a cloud database and identifying slow executing queries.

15. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises analyzing Structured Query Language (SQL) statements and stored procedures in a development fabric and determining cost optimization in accordance therewith.

16. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises auditing activity on a cloud database instance and providing a cost analysis breakdown of live usage over a sampling period.

17. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises utilizing one or more predefined trace templates to support a generic trace for recording cloud database events and activities that affect the cost of operations.

18. The method according to claim 9, wherein comparative profiling of said cloud storage service mechanism comprises utilizing one or more custom profiling trace templates for recording cloud database events and activities which affect the cost of operations.

19. The method according to claim 1, wherein analyzing the behavior of said input application code comprises determining one or more code optimizations that yield cost reductions.

20. The method of cost oriented profiling according to claim 19, wherein determining one or more code optimizations comprises utilizing a cost saving optimization algorithm adapted to one or more cloud application business cases.

21. The method of cost oriented profiling according to claim 19, wherein determining one or more code optimizations comprises:

utilizing pricing rules that identify and assess performance factors affecting one or more cloud billing decisions; and

recommending a course of action by applying a maximum expected utility action axiom to said one or more cloud billing decisions.

22. The method of cost oriented profiling according to claim 19, wherein determining one or more code optimizations comprises:

utilizing a decision tree representing one or more available price package alternatives; and

evaluating probability distributions for said one or more available price package alternatives.

23. The method of cost oriented profiling according to claim 19, wherein determining one or more code optimizations comprises:

estimating a compute hours metric; and

generating one or more code optimization recommendations that minimize said compute hours metric.

24. The method of cost oriented profiling according to claim 19, wherein determining one or more code optimizations comprises applying one or more accounting techniques to provide amortized cost information about one or more cloud operations.

25. The method of cost oriented profiling according to claim 19, wherein determining one or more code optimizations comprises applying time series predictions of past performance and previous demand to estimate an amount to be spent on cloud service resources over a given time period.

26. The method according to claim 1, wherein analyzing the behavior of said input application code comprises applying one or more analytical methods to determine a plurality of optimal code decisions that minimize the cost of deploying said input application to said cloud.

27. The method according to claim 26, wherein applying one or more analytical methods comprises evaluating one or more price/performance tradeoffs between different storage/computation alternatives.

28. The method according to claim 26, wherein applying one or more analytical methods comprises utilizing a routing service to dynamically determine an application instance to utilize in a distributed cloud application.

29. The method according to claim 26, wherein applying one or more analytical methods comprises scheduling cloud operations to optimize use of queues and to balance service access requests so as to minimize operational costs.

30. The method according to claim 26, wherein applying one or more analytical methods comprises load balancing an algorithm instance between different cloud computing host applications.

31. The method according to claim 26, wherein applying one or more analytical methods comprises moving services from local servers exhibiting high usage to idle cloud services.

32. The method according to claim 26, wherein applying one or more analytical methods comprises performing capacity provision analysis to determine the amount of storage used by an algorithm thereby enabling provisioning at an optimum cost/capacity point.

33. The method according to claim 1, wherein analyzing the behavior of said input application code comprises approximating the cost of said input application under different loads.

34. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises estimating the complexity of an operation asymptotically to determine a number of compute steps per billable hour.

35. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises estimating the computational complexity of an algorithm.

36. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises measuring the complexity of an algorithm to provide a performance analysis under relatively small random perturbations of worst case inputs.

37. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises reducing computationally expensive combinatorial calculations by applying a pruning algorithm and one or more cutoff thresholds.

38. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises determining the average running time per algorithmic operation over a worst case sequence of operations.

39. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises analyzing online algorithms exposed in said cloud whereby the performance of a service is compared to the performance of an optimal offline algorithm that processes the same sequence of requests.

40. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises inspecting an algorithm and recommending one or more general optimization heuristics that reduce costs.

41. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises identifying the processes in an algorithm which affect the overall duration of a cloud operation.

42. The method of cost oriented profiling according to claim 33, wherein approximating the cost of said application under different loads comprises analyzing properties of an algorithm to determine the level of resources it consumes and the operational cost associated therewith.

43. The method according to claim 1, wherein analyzing the behavior of said input application code comprises tracking changes in said application code and comparing cost oriented profiler results before and after said changes.

44. The method according to claim 1, wherein analyzing the behavior of said input application code comprises determining whether said application code passes or fails one or more total cost of ownership (TCO) verification tests.

45. The method according to claim 1, wherein said one or more cost reports comprises mapping projected total cost of ownership (TCO) of said input application to one or more cloud pricing packages.

46. The method according to claim 1, wherein analyzing the behavior of said input application code comprises providing guidance and recommendations based on said one or more cost reports thereby enabling a user to reduce costs and improve efficiency.

47. The method according to claim 46, wherein providing guidance and recommendations comprises utilizing one or more business rules that match the business demands of said input application.

48. A method for use on a computer of cost oriented profiling of input software application code executing in a cloud computing environment to generate an economic cost estimate therefrom, said method comprising:

performing, on said computer, a static analysis of said input application code with regard to the total economic cost of ownership (TCO) using a static analysis engine and associated rules to find economically costly code therein;

performing, on said computer, a simulation of and dynamic analysis of said input application code with regard to the TCO using a dynamic analysis engine and associated rules to find economically costly code therein;

determining one or more lines of code, blocks or functions that can potentially be optimized for cost;

determining whether economic costs associated with storage utilization costs can be optimized according to usage of said application code; and

simulating the usage of said input application code using a cost oriented simulator to generate a dynamic analysis indicating any costly code found thereby providing a measure of the total cost of ownership of said application code.

49. An apparatus for economic cost oriented profiling of input software application code executing in a cloud computing environment and generating an economic cost estimate therefrom, comprising:

one or more economic cost oriented static analysis engines;

one or more economic cost oriented dynamic analysis engines;

a static rules database;

a dynamic rules database;

a behavior analysis module operative to perform a static analysis of said input application code with regard to the total economic cost of ownership (TCO) using said one or more static analysis engines and associated static rules database to find economically costly code therein;

said behavior analysis module operative to perform a dynamic analysis of said input application code with regard to the TOC using said one or more dynamic analysis engines and associated dynamic rules database to find economically costly code therein; and

a cost oriented simulator coupled to said behavior analysis module and operative to simulate the usage of said input application code to generate dynamic analysis results indicating costly code found by said behavior analysis module thereby providing a measure of the economic TCO of said input application code.

50. The apparatus according to claim 49, wherein said behavior analysis module is operative to determine one or more lines of code, blocks or functions that can potentially be optimized for cost.

51. The apparatus according to claim 49, wherein said behavior analysis module is operative to determine whether storage utilization costs can be optimized according to usage of said input application code.

52. The apparatus according to claim 49, wherein said behavior analysis module comprises a code inspection rule engine operative to analyze said input application code for violations of cost related programming rules and conventions related to performance and service costs.

53. The apparatus according to claim 49, wherein said behavior analysis module comprises a storage utilization profiler engine operative to comparatively profile a cloud storage service mechanism used by said input application code.

54. The apparatus according to claim 49, wherein said behavior analysis module comprises a cost decision analysis engine operative to determine one or more code optimizations that yield cost reductions.

55. The apparatus according to claim 49, wherein said behavior analysis module comprises an operation analysis engine operative to apply one or more analytical methods to determine a plurality of optimal code decisions that minimize the cost of deploying said input application to said cloud.

56. The apparatus according to claim 49, wherein said behavior analysis module comprises an algorithm analysis engine operative to approximate the cost of said input application under different loads.

57. The apparatus according to claim 49, wherein said behavior analysis module comprises a code comparison engine operative to track changes in said application code and compare cost oriented profiler results before and after said changes.

58. The apparatus according to claim 49, wherein said behavior analysis module comprises a total cost of ownership (TCO) verifier operative to determine whether said input application passes or fails one or more TCO verification tests.

59. A computer program product for cost oriented profiling of input software application code executing in a cloud computing environment to generate an economic cost estimate therefrom, the computer program product comprising:

a tangible, non-transitory computer usable medium having computer usable code embodied therewith, the computer usable program code comprising;

computer usable code configured for performing a static analysis of said input application code with regard to the total economic cost of ownership (TCO) using a static analysis engine and associated rules to find economically costly code therein;

computer usable code configured for performing a simulation of and dynamic analysis of said input application code with regard to the TCO using a dynamic analysis engine and associated rules to find economically costly code therein; and

computer usable code configured for performing a dynamic analysis of said input application utilizing a dynamic analysis engine and associated rules to find economically costly code therein with second indications as to the portions of said software application code that have the most impact on the TCO thereof;

computer usable code configured for performing said dynamic analysis by simulating the usage of said input application code utilizing a cost oriented simulator to generate a dynamic analysis with third indications of any costly code found thereby providing a measure to a user of the total cost of ownership of said input application code; and

computer usable code configured for generating one or more economic cost estimate reports based on the results of said behavior static analysis, simulation and dynamic analysis, wherein said economic cost reports include those portions of the application code having the most impact on the TCO, said first and second indications and one or more recommendations for reducing the TCO.

60. The computer program product according to claim 59, further comprising computer usable code configured for determining one or more lines of code, blocks or functions of said input application code that can potentially be optimized for cost.

61. The computer program product according to claim 59, further comprising computer usable code configured for determining whether storage utilization costs can be optimized according to the usage of said input application code.

62. The computer program product according to claim 59, further comprising computer usable code configured for simulating the usage of said input application code utilizing a cost oriented simulator, said cost oriented simulator operative to generate a dynamic analysis indicating any costly code found thereby providing a measure of the total cost of ownership of said input application code comparing multiple versions of said source code and determining if a particular change to said source code increases or decreases software TCO.