Managing server load by varying responses to requests for dynamically-generated web pages

• US 7,624,047 B1
• Filed: 03/14/2003
• Issued: 11/24/2009
• Est. Priority Date: 07/31/2002
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Round, Matthew J.
• Primary Examiner(s): Zeender; F.
• Assistant Examiner(s): ALMATRAHI, FARIS S
• Application Number:
  US10/390,067
• Time in Patent Office: 2,447 Days
• Field of Search: 705/27
• US Class Current: 705/26.1
• CPC Class Codes: G06F 16/958 Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
• CPC Class Codes: G06F 9/50 Allocation of resources, e.g. of the central processing unit [CPU]
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/1012 based on compliance of requirements or conditions with available server resources
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/60 Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources admission control or resource allocation H04L47/70
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]

Abstract

A web server system regulates its own load level by dynamically varying its responses to requests for like web pages. The responses are preferably varied at least in-part using variables that directly or indirectly specify the type and/or quantity of processing to be performed in responding to requests for dynamically-generated web pages. The type of response generated by the web server system is preferably selected based upon one or more of the following: (a) the current load level of the server system as a whole, (b) the current load levels of specific computing devices or components invoked by the request, and (c) a profile of the user who requests the page.

Claims

1. A web server system, comprising:

a plurality of physical servers that generate personalized web page content in response to page requests from users, wherein the physical servers generate the personalized web page content using user profile data associated with particular users;

a level-of-service module that generates level-of-service parameters based at least in-part on current load levels of the plurality of physical servers; and

a dynamic page generator that responds to the page requests according to the level-of-service parameters such that an amount of processing performed by the physical servers per page request in generating requested web pages decreases as said load level increases, and increases as said load level decreases, wherein the dynamic page generator controls said amount of processing performed by the physical servers at least partly by using the level-of-service parameters to vary a quantity of user profile data used by the physical servers to generate the personalized web page content;

wherein the dynamic page generator is responsive to an increase in said load level by causing the physical servers to reduce, to a non-zero level, an amount of user profile data used to personalize web pages;

whereby load levels are regulated in a feedback configuration.

2. The system of claim 1, wherein the dynamic page generator responds to the page requests according to the level-of-service parameters at least in-part by selecting a depth of processing to be performed in generating said personalized web page content.

3. The system of claim 1, wherein the dynamic page generator responds to the page requests according to the level-of-service parameters at least in-part by selecting a size of an input data set used to generate personalized recommendations.

4. The system of claim 1, wherein the dynamic page generator additionally regulates said load levels by using the level-of-service parameters to select between alternative web page templates used to generate web pages.

5. The system of claim 1, wherein the dynamic page generator additionally regulates said load levels by using the level-of-service parameters to select services to disable for purposes of responding to specific page requests.

6. The system of claim 1, wherein the level-of-service module generates the level-of-service parameters based further upon profiles of users, such that different levels of service are provided to different users of the system.

7. The system of claim 6, wherein the level-of-service module generates the level-of-service parameters such that a higher level of service is generally provided to frequent users over infrequent users.

8. The system of claim 1, wherein the personalized web page content included in web pages generated by the dynamic page generator is dependent upon said level-of-service parameters.

9. The system of claim 1, wherein the dynamic page generator responds to the page requests according to the level-of-service parameters such that a load placed on the web server system by a page request is generally inversely proportional to a current load level of the web server system at a time of said page request.

10. The system of claim 1, wherein the dynamic page generator controls the generation of the personalized web page content by the physical servers such that a quantity of processing resources used by said physical servers to generate personalized content of a given web page is generally inversely proportional to said current load level.

11. The system of claim 10, wherein the dynamic page generator controls the generation of the personalized content, at least in part, by passing the level-of-service parameters to the physical servers.

12. The system of claim 10, wherein said personalized content comprises search results ranked according to information about a particular user.

13. The system of claim 1, wherein the dynamic page generator responds to the page requests such that the web pages generated by the dynamic page generator are generally personalized to a greater extent when said load levels are low.

26. The system of claim 1, wherein the dynamic page generator additionally uses the level of service parameters to vary a caching policy used to cache user profile data in non-volatile memory of the physical servers.

27. The system of claim 1, wherein the user profile data used to generate personalized web pages includes purchase history data reflective of purchases made by particular users.

14. A method, comprising:

providing a plurality of physical servers that generate personalized web page content in response to page requests from users, said physical servers operative to generate the personalized web page content using user profile data associated with particular users;

generating level-of-service parameters using a level of service module, said level-of-service parameters based at least in-part on current load levels of the plurality of physical servers; and

using the level-of-service parameters to vary amounts of user profile data used by the physical servers to generate the personalized web page content such that an amount of computing resources used by the physical servers per page request decreases as a load level of the physical servers increases, and increases as the load level of the physical servers decreases;

wherein using the level of service parameters to vary said amounts of user profile data comprises responding to an increase in said load levels by causing the physical servers to use a decreased but non-zero amount of user profile data to generate the personalized web page content;

whereby load levels are regulated in a feedback configuration.

15. The method of claim 14, wherein the user profile data includes purchase history data reflective of purchases of particular users.

16. The method of claim 14, wherein the method comprises using the level-of-service parameters to select a size of an input data set of user profile data used to generate personalized recommendations.

17. The method of claim 14, wherein the method additionally comprises using the level-of-service parameters to select between alternative web page templates used to generate web pages.

18. The method of claim 14, wherein the method additionally comprises using the level-of-service parameters to select services to disable in responding to specific page requests.

19. The method of claim 14, wherein the method comprises generating the level-of-service parameters based further upon profiles of users, such that different levels of service are provided to different users.

20. The method of claim 14, wherein the method comprises generating web pages that include personalized web page content that is dependent upon said level-of-service parameters.

21. The method of claim 14, wherein the method comprises varying said amounts of user profile data according to the level-of-service parameters such that a load placed on the physical servers by a page request is generally inversely proportional to a current load level at a time of said page request.

22. The method of claim 14, wherein the method comprises controlling the generation of the personalized web page content by the physical servers such that a quantity of processing resources used by said physical servers to generate personalized content of a given web page is generally inversely proportional to said current load level.

23. The method of claim 22, wherein the method comprises controlling the generation of the personalized web page content, at least in part, by passing the level-of-service parameters to the physical servers.

24. The method of claim 22, wherein said personalized web page content comprises search results ranked according to information about a particular user.

25. The method of claim 14, wherein the method comprises varying web page responses such that web pages generated with the physical servers are generally personalized to a greater extent when said load levels are low.

28. The method of claim 14, further comprising using the level of service parameters to vary a caching policy used to cache said user profile data in non-volatile server memory.

29. The method of claim 14, wherein said generation and use of the level-of-service parameters is automated by a computerized system.