Content delivery network cache grouping

• US 8,683,002 B2
• Filed: 01/02/2013
• Issued: 03/25/2014
• Est. Priority Date: 10/02/2009
• Status: Active Grant

• Current Assignee: Limelight Networks, Inc.
• Original Assignee: Limelight Networks, Inc.
• Inventors: Harvell, Bradley B., McCarthy, Nils H.
• Primary Examiner(s): Chang, Jungwon
• Application Number:
  US13/732,570
• Publication Number:
  US 20130246555A1
• Time in Patent Office: 447 Days
• Field of Search: 709/214, 709/219, 709/203, 709/226, 709/224, 709/238, 370/351, 370/389, 718/105, 711/122
• US Class Current: 709/214
• CPC Class Codes: G06F 16/9574 of access to content, e.g. by caching
• CPC Class Codes: H04L 65/1026 at the edge
• CPC Class Codes: H04L 67/289 Intermediate processing functionally located close to the data consumer application, e.g. in same machine, in same home or in same sub-network
• CPC Class Codes: H04L 67/568 Storing data temporarily at an intermediate stage, e.g. caching

Abstract

One or more content delivery networks (CDNs) that deliver content objects for others is disclosed. Content is propagated to edge servers through hosting and/or caching. End user computers are directed to an edge server for delivery of a requested content object by a universal resource indicator (URI). When a particular edge server does not have a copy of the content object from the URI, information is passed to another server, the ancestor or parent server to find the content object. There can be different parents servers designated for different URIs. The parent server looks for the content object and if not found, will go to another server, the grandparent server, and so on up a hierarchy within the group. Eventually, the topmost server in the hierarchy goes to the origin server to find the content object. The origin server may be hosted in the CDN or at a content provider across the Internet. Once the content object is located in the hierarchical chain, the content object is passed back down the chain to the edge server for delivery. Optionally, the various servers in the chain may cache or host the content object as it is relayed.

Claims

1. A method to retrieve content objects in a content delivery network (CDN) having a plurality of points of presence (POPs) distributed geographically by checking one or more neighboring servers before checking a parent server, the method comprising:

receiving a first universal resource identifier (URI) at a first edge server having a first cache wherein;

the first edge server is in a first POP of the plurality of POPs, and

the first URI specifies a first content object;

determining that the first cache does not hold the first content object;

querying the one or more neighboring servers for the first content object, wherein the one or more neighboring servers each have a cache;

determining that one or more caches of the one or more neighboring servers stores the first content object;

serving the first content object from the one or more caches of the one or more neighboring servers;

receiving a second URI at the first edge server, wherein the second URI specifies a second content object;

determining that the first cache does not hold the second content object;

querying the one or more neighboring servers for the second content object;

determining that the one or more caches of the one or more neighboring servers do not hold the second content object;

analyzing the second URI to determine the parent server, and

requesting the second content object from the parent server after determining that the one or more caches of the one or more neighboring servers do not hold the second content object.

2. The method to retrieve content objects in the CDN having the plurality of POPs distributed geographically as recited in claim 1, wherein the one or more caches of the one or more neighboring servers are located within the first POP.

3. The method to retrieve content objects in the CDN having the plurality of POPs distributed geographically as recited in claim 2, wherein the one or more caches of the one or more neighboring servers comprise each cache, for storing content objects, within the first POP.

4. The method to retrieve content objects in the CDN having the plurality of POPs distributed geographically as recited in claim 1, wherein:

the parent server is in a second POP of the plurality of POPs, and

the second POP is different from the first POP.

5. The method to retrieve content objects in the CDN having the plurality of POPs distributed geographically as recited in claim 1, wherein the serving the first content object from the one or more caches of the one or more neighboring servers comprises serving the first content object to an end-user system from the first edge server acting as a proxy for the one or more neighboring servers.

6. The method to retrieve content objects in the CDN having the plurality of POPs distributed geographically as recited in claim 1, wherein serving of the first content object is reassigned from the first edge server to the one or more neighboring servers.

7. The method to retrieve content objects in the CDN having the plurality of POPs distributed geographically as recited in claim 1, wherein querying one or more neighboring servers for the second content object comprises:

querying a second edge server; and

querying a third edge server, wherein the querying the second edge server and the querying the third edge server are performed overlapping in time.

8. A content delivery network (CDN) for retrieving content objects from one or more neighboring servers before requesting content objects from a parent server, the CDN comprising:

a plurality of POPs distributed geographically;

a first edge server having a first cache, wherein;

the first edge server is located within a first POP of the plurality of POPs,

the first edge server receives a first universal resource identifier (URI),

the first URI specifies a first content object,

the first edge server determines that the first cache does not hold the first content object,

the first edge server queries one or more neighboring servers for the first content object, wherein the one or more neighboring servers each have a cache,

the first edge server receives an indication that the first content object is stored in one or more caches of the one or more neighboring servers,

the first edge server receives a second URI,

the second URI specifies a second content object,

the first edge server determines that the first cache does not hold the second content object,

the first edge server determines that the one or more caches of the one or more neighboring servers do not store the second content object, and

the first edge server requests the second content object from the parent server after determining that the one or more caches of the one or more neighboring servers do not hold the second content object; and

a second POP, wherein;

the second POP is different from the first POP, and

the second POP comprises the parent server.

9. The CDN for retrieving content objects from the one or more neighboring servers before requesting content objects from the parent server, as recited in claim 8, wherein the one or more neighboring servers are located within the first POP.

10. The CDN for retrieving content objects from the one or more neighboring servers before requesting content objects from the parent server, as recited in claim 8, wherein an end user is served the first content object from the one or more caches of the one or more neighboring servers by the first edge server acting as a proxy for the one or more neighboring servers.

11. The CDN for retrieving content objects from the one or more neighboring servers before requesting content objects from the parent server, as recited in claim 8, the CDN further comprising the one or more neighboring servers, wherein serving the first content object to an end-user system is reassigned from the first edge server to the one or more neighboring servers.

12. The CDN for retrieving content objects from the one or more neighboring servers before requesting content objects from the parent server, as recited in claim 8, the CDN further comprising:

a first neighboring server, wherein;

the first neighboring server is part of the one or more neighboring servers, and

the first edge server makes a first query of the first neighboring server for the first content object; and

a second neighboring server, wherein;

the second neighboring server is part of the one or more neighboring servers,

the first edge server makes a second query of the second neighboring server for the first content object, and

the first query and the second query overlap in time.

13. The CDN for retrieving content objects from the one or more neighboring servers before requesting content objects from the parent server, as recited in claim 8, wherein the first URI and the second URI correspond to different content objects.

14. The CDN for retrieving content objects from the one or more neighboring servers before requesting content objects from the parent server, as recited in claim 8, wherein the first URI is a portion of a third URI.

15. A content delivery network (CDN) having a plurality of points of presence (POPs) for retrieving content objects, the CDN comprising one or more servers programmed for:

receiving a first universal resource identifier (URI) at a first server having a first cache, wherein;

the first server is in a first POP of the plurality of POPs, and

the first URI specifies a first content object;

determining that the first cache does not hold the first content object;

querying one or more neighboring servers for the first content object, wherein the one or more neighboring servers each have a cache;

determining that one or more caches of the one or more neighboring servers stores the first content object;

serving the first content object from the one or more caches of the one or more neighboring servers;

receiving a second URI at the first server, wherein the second URI specifies a second content object;

determining that the first cache does not hold the second content object;

determining that the one or more caches of the one or more neighboring servers do not hold the second content object;

analyzing the second URI to determine a cache hierarchy for the second URI, and

requesting the second content object from the cache hierarchy after determining that the one or more caches of the one or more neighboring servers do not hold the second content object.

16. The CDN having the plurality of POPs for retrieving content objects as recited in claim 15, wherein the one or more neighboring servers are in the first POP.

17. The CDN having the plurality of POPs for retrieving content objects as recited in claim 15, wherein:

the cache hierarchy comprises a cache in a second POP; and

the second POP is different from the first POP.

18. The CDN having the plurality of POPs for retrieving content objects as recited in claim 15, wherein an end user is served the first content object from the one or more caches of the one or more neighboring servers by the first server acting as a proxy for the one or more neighboring servers.

19. The CDN having the plurality of POPs for retrieving content objects as recited in claim 15, wherein querying one or more neighboring servers for the first content object comprises:

querying a second server; and

querying a third server, wherein the querying the second server and the querying the third server is performed overlapping in time.

20. The CDN having the plurality of POPs for retrieving content objects as recited in claim 15, wherein the first content object is not the same as the second content object.