Content delivery network (CDN) cold content handling

Patent metadata

Field	Value
Publication number	US-9680952-B2
Application number	US-201213470262-A
Country	US
Kind code	B2
Filing date	May 12, 2012
Priority date	Jul 19, 2007
Publication date	Jun 13, 2017
Grant date	Jun 13, 2017

Abstract

A method of content delivery in a content delivery network (CDN), where the CDN is deployed, operated and managed by a content delivery network service provider (CDNSP). The CDN comprises a set of content servers and a domain name system (DNS). For a given content provider, a determination is first made whether the content provider has “cold content” delivery requirements by evaluating one or more factors that include: total content size, size of content objects expected to be served, uniqueness of content, total number of content objects, and a percentage of the total content size that is expected to account for a given percentage of traffic. Upon a determination that the content provider has cold content delivery requirements, a subset of the CDN content servers are configured to implement a set of one or handling rules for managing delivery of the cold content from the CDN content servers.

First claim

Having described our invention, what we claim is as follows:

1. A method of content delivery in a content delivery network (CDN), the CDN deployed, operated and managed by a content delivery network service provider (CDNSP) and comprising a set of content servers and a domain name system (DNS) associated with a CDN namespace, comprising:
for a given content provider, determining whether the content provider has cold content delivery requirements;
upon a determination that the content provider has cold content delivery requirements, determining from a set of cold content CDN delivery service configurations a cold content delivery service configuration that, with respect to the content provider's cold content delivery requirements, trades off content provider footprint, traffic, and number of objects, wherein the cold content delivery service configuration for the content provider differs from the cold content delivery service configuration for at least one other content provider that has different cold content delivery requirements;
instantiating the cold content delivery service configuration that has been determined for the content provider by delivering to each of the set of content servers a configuration file that defines the determined cold content delivery service configuration for the content provider; and
at a given content server, and in response to receipt of a request for cold content of the content provider, retrieving the configuration file that defines the cold content delivery service configuration for the content provider, applying a cold content handling rule from the configuration file to the request for cold content, and serving a response, the content server implemented in a machine having a hardware component.

2. The method as described in claim 1 wherein the determining whether the content provider has cold content delivery requirements step evaluates a set of criteria for the content provider, the set of criteria being one of: total content size, size of content objects expected to be served, uniqueness of content, total number of content objects, and a percentage of total content size that is expected to account for a given percentage of traffic.

3. The method as described in claim 1 wherein the handling rule associates a given time-to-live (TTL) value with the cold content such that, if the cold content is in cache, an if-modified-since (IMS) check need not be performed at a content provider origin server during a time period specified by the TTL value.

4. The method as described in claim 1 wherein the handling rule provides an HTTP 304 response code in response to receipt at the server of an if-modified-since (IMS) request.

5. The method as described in claim 1 wherein the handling rule implements either a constrained set of cache hierarchy nodes or a constrained set of regions for the cold content.

6. The method as described in claim 1 wherein the handling rule hashes a serial number associated with the incoming request to generate a new serial number, wherein the new serial number is then associated with a new request that is forwarded to at least one other server.

7. The method as described in claim 1 wherein the handling rule applies a forward rate limit on requests to a content provider origin infrastructure.

8. The method as described in claim 1 wherein the handling rule configures a cache so that the cold content is only cached after a predetermined number of requests for the cold content have been received.

9. The method as described in claim 1 wherein the handling rule configures a cache so that the cold content is only cached in the cache if, following a given time period, the cold content is subsequently re-requested with the given time period.

10. The method as described in claim 1 wherein the handling rule configures a predetermined number of content objects that a given content provider may store in a cache.

11. The method as described in claim 1 wherein the handling rule configures the content server, upon a cache miss, to go forward to attempt to fetch the cold content from a first tier of a multi-tier cache hierarchy.

12. The method as described in claim 1 wherein the handling rule configures the content server, upon a cache miss, to go forward to attempt to fetch the cold content from a first tier of a multi-tier cache hierarchy with a given probability and from a second tier of the multi-tier cache hierarchy otherwise.

13. The method as described in claim 1 wherein the handling rule configures the content server to inhibit caching of the cold content in a cache if the cold content is received from a peer server.

Assignees

• Kovacs Laszlo
• Oslakovic Keith E
• Kasbekar Mangesh
• Chen Zewei
• Akamai Tech Inc

Inventors

• Kovacs Laszlo
• Oslakovic Keith E
• Kasbekar Mangesh
• Chen Zewei

Classifications

• H04L67/2842 Primary

  Electricity · mapped topic

• H04L67/568 Primary

  Storing data temporarily at an intermediate stage, e.g. caching · CPC title