Cache validation using smart source selection in a data network

• US 8,650,266 B2
• Filed: 04/12/2002
• Issued: 02/11/2014
• Est. Priority Date: 03/26/2002
• Status: Active Grant

• Current Assignee: AT&T Intellectual Property II LP (AT&T, Inc.)
• Original Assignee: AT&T Intellectual Property II LP (AT&T, Inc.)
• Inventors: Cohen, Edith
• Primary Examiner(s): STRANGE, AARON N
• Application Number:
  US10/063,341
• Publication Number:
  US 20030187917A1
• Time in Patent Office: 4,323 Days
• Field of Search: 709/212, 709/213, 709/214, 709/216, 709/218, 709/219, 709/217, 709/223, 711/122, 711/133
• US Class Current: 709/219
• CPC Class Codes: H04L 67/02 based on web technology, e.g. hypertext transfer protocol [HTTP]
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• CPC Class Codes: H04L 67/5682 Policies or rules for updating, deleting or replacing the stored data
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]

Abstract

In accordance with aspects of the invention, “low-level” caches can utilize source selection while non-authoritative sources can take advantage of rejuvenation to alleviate what the inventor refers to as “age penalty” and thereby reduce validation traffic.

Claims

1. A method of processing a request for content in a cache server comprising:

receiving, by the cache server, the request for the content from a client;

determining, by the cache server, if a stale copy of the content is in the cache server;

estimating, by the cache server, a freshness metric and an estimate of a fetching time for each one of a group of servers;

selecting, by the cache server, a source server from the group of servers, based on a comparison of an expected remaining freshness of a copy of the content stored at each one of the group of servers computed based on the estimating, when the stale copy of the content is in the cache server, wherein the group of servers comprises a high level cache server and an origin server, wherein the high level cache server comprises a reverse proxy for the origin server, wherein the high level cache server performs a pre-term refresh from the origin server in accordance with a scheduled process to check for content that has a respective freshness metric below a threshold, wherein the content comprises popular content defined as content that has a number of requests above a request threshold per an interval, and wherein the origin server has a copy of the content with a zero age;

sending, by the cache server, a validation request to the source server; and

updating the freshness of the content based on a response from the source server where the copy of the content stored at the source server has not been modified.

2. The method of claim 1 wherein the origin server comprises an authoritative server and the high level cache server comprises a non-authoritative server.

3. The method of claim 2 wherein the non-authoritative server comprises a reverse proxy server.

4. The method of claim 1 wherein the source server is selected based on a likelihood of a cache miss at the source server.

5. The method of claim 1 wherein the freshness of the copy of the content is measured by a time-to-live value.

6. The method of claim 5 wherein the content is validated using a hyper text transfer protocol.

7. A memory unit storing program instructions which, when executed by a cache server, cause the cache server to perform operations of processing a request for in the cache server, the operations comprising:

receiving the request for the content from a client;

determining if a stale copy of the content is in the cache server;

estimating a freshness metric and an estimate of a fetching time for each one of a group of servers;

selecting a source server from the group of servers, based on a comparison of an expected remaining freshness of a copy of the content stored at each one of the group of servers computed based on the estimating, when the stale copy of the content is in the cache server, wherein the group of servers comprises a high level cache server and an origin server, wherein the high level cache server comprises a reverse proxy for the origin server, wherein the high level cache server performs a pre-term refresh from the origin server in accordance with a scheduled process to check for content that has a respective freshness metric below a threshold, wherein the content comprises popular content defined as content that has a number of requests above a request threshold per an interval, and wherein the origin server has a copy of the content with a zero age;

sending a validation request to the source server; and

updating the freshness of the content based on a response from the source server where the copy of the content stored at the source server has not been modified.

8. The memory unit of claim 7 wherein the origin server comprises an authoritative server and the high level cache server comprises a non-authoritative server.

9. The memory unit of claim 8 wherein the non-authoritative server comprises a reverse proxy server.

10. The memory unit of claim 7 wherein the source server is selected based on a likelihood of a cache miss at the source server.

11. The memory unit of claim 7 wherein the freshness of the copy of the content is measured by a time-to-live value.

12. The memory unit of claim 11 wherein the content is validated using a hyper text transfer protocol.