SYSTEM AND METHOD FOR MODELING A SESSION-BASED SYSTEM WITH A TRANSACTION-BASED ANALYTIC MODEL

• US 20090119301A1
• Filed: 11/05/2007
• Published: 05/07/2009
• Est. Priority Date: 11/05/2007
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Zhang, Qi, Cherkasova, Ludmila
• Granted Patent: US 8,326,970 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Abstract

According to an embodiment of the present invention, a method for deriving an analytic model for a session-based system is provided. The method comprises receiving, by a model generator, client-access behavior information for the session-based system, wherein the session-based system comprises a plurality of interdependent transaction types. The method further comprises deriving, by the model generator, from the received client-access behavior information, a stateless transaction-based analytic model of the session-based system, wherein the derived transaction-based analytic model models resource requirements of the session-based system for servicing a workload. According to certain embodiments, the derived transaction-based analytic model is used for performing capacity analysis of the session-based system.

Claims

1. A method for deriving an analytic model for a session-based system, the method comprising:

receiving, by a model generator, client-access behavior information for the session-based system, wherein the session-based system comprises a plurality of interdependent transaction types; and

deriving, by the model generator, from the received client-access behavior information, a stateless transaction-based analytic model of the session-based system, wherein the derived transaction-based analytic model models resource requirements of the session-based system for servicing a workload.

2. The method of claim 1 wherein said client-access behavior information comprises a probability matrix of transition probabilities of the plurality of interdependent transaction types, and wherein said deriving further comprises:

determining, from the probability matrix, a vector that represents the steady-state probability for the plurality of interdependent transaction types.

3. The method of claim 2 wherein the determined vector, π

_i, gives the overall percentage of the interdependent transactions of type i in the workload of the session-based system.

4. The method of claim 3 wherein said workload of the session-based system comprises a historical workload previously encountered by the session-based system.

5. The method of claim 2 wherein said deriving further comprises:

determining a closed system of a network of queues based on the determined steady-state probability of the interdependent transaction types.

6. The method of claim 1 further comprising:

determining said client-access behavior information from a historical workload encountered by the session-based system.

7. The method of claim 1 wherein the interdependent transaction types are transaction types of an application hosted by the session-based system.

8. The method of claim 7 wherein the application comprises an e-commerce application, and wherein the interdependent transaction types include one or more of the following transaction types;

searching for a product;

selecting a product;

inputting shipping information;

arranging payment agreement;

receiving a confirmation;

home;

new products;

best sellers;

product detail;

search request;

search results;

shopping cart;

customer registration;

buy request;

buy confirm;

order inquiry;

order display;

admin request; and

admin confirm.

9. The method of claim 1 further comprising:

determining a corresponding resource cost for each of the interdependent transaction types.

10. The method of claim 9 wherein said determining said corresponding resource cost comprises:

receiving a representative workload of the session-based system; and

employing a regression-based analysis for approximating the corresponding resource cost for each of the interdependent transaction types.

11. The method of claim 10 wherein said resource cost includes CPU and memory demands.

12. The method of claim 9 further comprising:

parameterizing the stateless transaction-based analytic model of the session-based system based on the determined resource costs of the interdependent transaction types.

13. The method of claim 12 wherein the parameterized transaction-based model reflects capacity of the session-based system that is consumed for serving the workload under analysis.

14. The method of claim 1 wherein the stateless transaction-based model treats each transaction, of said plurality of interdependent transaction types, in the workload of the session-based system as if totally independent of other transactions.

15. A method for performing capacity analysis of a session-based system, the method comprising:

receiving, by a capacity analyzer, a representative workload of the session-based system, wherein the session-based system hosts at least one application comprising a plurality of interdependent transaction types;

determining, by said capacity analyzer, from a transaction-based analytic model of the session-based system, capacity of the session-based system for supporting the representative workload; and

performing at least one action based at least in part on or responsive to the determined capacity of the session-based system for supporting the representative workload.

16. The method of claim 15 wherein the performing said at least one action comprises at least one of:

outputting, by said capacity analyzer, information reflecting the determined capacity of the session-based system for supporting the representative workload;

dynamically provisioning resources for said session-based system based at least in part on the determined capacity of the session-based system for supporting the representative workload; and

communicating said determined capacity of the session-based system for supporting the representative workload from said capacity analyzer to another processor-based device.

17. The method of claim 15 wherein the representative workload comprises a historical workload previously encountered by the session-based system.

18. The method of claim 15 further comprising:

generating said transaction-based analytic model of the session-based system from a probability matrix of transition probabilities of the plurality of interdependent transaction types observed for the session-based system.

19. The method of claim 18 wherein said generating further comprises:

determining, from the probability matrix, a vector that represents the steady-state probability for the plurality of interdependent transaction types.

20. The method of claim 19 wherein said generating further comprises:

determining a closed system of a network of queues based on the determined steady-state probability of the interdependent transaction types.

21. The method of claim 15 further comprising:

employing a regression-based analysis for approximating a corresponding resource cost for each of the interdependent transaction types; and

parameterizing the transaction-based analytic model of the session-based system based on the determined resource costs of the interdependent transaction types.

22. The method of claim 21 wherein the parameterized transaction-based model reflects capacity of the session-based system that is consumed for serving the representative workload under analysis.

23. The method of claim 15 wherein the transaction-based model treats each transactions of said plurality of interdependent transaction types, in the representative workload of the session-based system as if totally independent of other transactions.

24. A system comprising:

computer-readable medium to which client-access behavior information for a session-based system is stored, wherein the session-based system hosts at least one application that comprises a plurality of interdependent transaction types;

a resource cost calculator for receiving a representative workload and determining from the representative workload a corresponding resource cost of the session-based system for each of the plurality of interdependent transaction types; and

an analytic model generator for deriving, based at least on the client-access behavior information and the corresponding resource cost of the plurality of interdependent transaction types, a stateless transaction-based analytic model of the session-based system, wherein the derived transaction-based analytic model models resource requirements of the session-based system for servicing a workload.

25. The system of claim 24 wherein said client-access behavior information comprises a probability matrix of transition probabilities of the plurality of interdependent transaction types, and wherein said analytic model generator determines, from the probability matrix, a vector that represents the steady-state probability for the plurality of interdependent transaction types;

and wherein said analytic model generator parameterizes the stateless transaction-based analytic model of the session-based system based on the determined resource costs of the interdependent transaction types.