Distributed caching system with subscription based notification of cache invalidations

• US 9,807,190 B2
• Filed: 11/18/2016
• Issued: 10/31/2017
• Est. Priority Date: 10/04/2013
• Status: Active Grant

• Current Assignee: Akamai Technologies, Inc.
• Original Assignee: Akamai Technologies, Inc.
• Inventors: Flack, Martin T., Steiner, Moritz M., Ludin, Stephen L., Hatala, Jozef
• Primary Examiner(s): Tiv, Backhean
• Application Number:
  US15/356,070
• Publication Number:
  US 20170085667A1
• Time in Patent Office: 347 Days
• Field of Search: None
• CPC Class Codes: H04L 2209/76 Proxy, i.e. using intermediary entity to perform cryptographic operations network architectures or network communication protocols using hop-by-hop encryption H04L63/0464
• CPC Class Codes: H04L 63/0884 by delegation of authentication, e.g. a proxy authenticates an entity to be authenticated on behalf of this entity vis-à-vis an authentication entity
• 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 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 67/568 Storing data temporarily at an intermediate stage, e.g. caching
• CPC Class Codes: H04N 21/23106 involving caching operations prefetching while addressing of a memory level in which the access to the desired data or data block requires associative addressing means within memory systems or architectures G06F12/0862; caching at an intermediate stage in a data network H04L67/568

Abstract

Described herein are systems, devices, and methods for content delivery on the Internet. In certain non-limiting embodiments, a caching model is provided that can support caching for indefinite time periods, potentially with infinite or relatively long time-to-live values, yet provide prompt updates when the underlying origin content changes. In one approach, an origin server can annotate its responses to content requests with tokens, e.g., placing them in an appended HTTP header or otherwise. The tokens can drive the process of caching, and can be used as handles for later invalidating the responses within caching proxy servers delivering the content. Tokens may be used to represent a variety of kinds of dependencies expressed in the response, including without limitation data, data ranges, or logic that was a basis for the construction of the response.

Claims

1. A method performed by a distributed system comprising a plurality of servers including a first, a second and a third server operable to communicate over one or more computer networks, the method comprising:

at the first server;

receiving a request for particular content from one of a plurality of clients, the request comprising a domain name;

sending the request for particular content to the third server;

in response to the request, receiving a message that comprises the particular content and one or more tokens distinct from the particular content;

storing the particular content in a local cache of the first server in association with the one or more tokens;

sending a subscription request to the third server, the subscription request comprising a request to receive invalidation messages for tokens associated with the domain name over a subscription channel associated with the domain name;

at the second server;

receiving a write request from any of a same and another one of the plurality of clients, the write request comprising the domain name;

sending the write request to the third server;

receiving a response to the write request from the third server;

at the third server;

sending the response to the write request to the second server;

determining that the response to the write request is associated with a particular token;

upon said determination, sending an invalidation message for the particular token to the first server, the invalidation message sent over the subscription channel for the domain name;

at the first server;

receiving the invalidation message for the particular token from the third server, and upon said receipt, determining that the particular token is one of the one or more tokens associated with the particular content, and based at least in part upon said determination, marking the particular content stored in the local cache as invalid;

wherein each of the first, second, and third servers comprise at least one microprocessor and memory storing computer instructions for execution by the at least one microprocessor to perform the method as specified above.

2. The method of claim 1, wherein at least the first and second servers are two of a plurality of nodes in a content delivery network.

3. The method of claim 1, wherein the third server communicates with an origin server to obtain the particular content and the response to the write request.

4. The method of claim 1, where the request for particular content comprises one or more of:

(i) an HTTP request, and (ii) a request for service from an application programmer interface (API).

5. The method of claim 1, wherein the invalidation message comprises any of:

an explicit assertion that the particular token is invalid, and an implicit assertion that the particular token is invalid.

6. The method of claim 1, wherein the particular token comprises a moniker and any of a serial number and a timestamp, and the invalidation message indicates that the token is invalid because the invalidation message includes the moniker and any of a different serial number and a different timestamp.

7. A method performed by a distributed system comprising a plurality of servers including a first, a second and a third server operable to communicate over one or more computer networks, the method comprising:

at the first server;

receiving a request for particular content from one of a plurality of clients, the request comprising a domain name;

sending the request for particular content to an origin server;

in response to the request, receiving a message comprising the particular content and one or more tokens distinct from the particular content;

storing the particular content in a local cache of the first server in association with one or more tokens;

sending a subscription request to the origin server, the subscription request comprising a request to receive invalidation messages for tokens associated with the domain name over a subscription channel associated with the domain name;

at the second server;

receiving a write request from any of a same and another one of the plurality of clients, the write request comprising the domain name;

sending the write request to the origin server;

receiving a response to the write request from the origin server;

determining that the response to the write request includes a particular token;

upon said determination, sending a first invalidation message for the particular token to a third server;

at the third server;

receiving the first invalidation message and upon said receipt, sending to the first server a second invalidation message for the particular token, the second invalidation message sent over the subscription channel associated with the domain name;

at the first server;

receiving the second invalidation message for the particular token from the third server, and upon said receipt, determining that the particular token is one of the one or more tokens, and based at least in part upon said determination, marking the particular content stored in the local cache as invalid;

wherein each of the first, second; and

third servers comprises at least one microprocessor and memory storing computer instructions for execution by the at least one microprocessor to perform the method as specified above.

8. The method of claim 7, wherein at least the first and second servers are two of a plurality of nodes in a content delivery network.

9. The method of claim 7, where the request for particular content comprises one or more of:

(i) an HTTP request and (ii) a request for service from an application programmer interface (API).

10. The method of claim 7, wherein the second invalidation message comprises any of:

an explicit assertion that the particular token is invalid, and an implicit assertion that the particular token is invalid.

11. The method of claim 7, wherein the particular token comprises a moniker and any of a serial number and a timestamp, and the second invalidation message indicates that the particular token is invalid because the invalidation message includes the moniker and any of a different serial number and a different timestamp.

12. A system comprising a plurality of servers operable to communicate over one or more computer networks, the system comprising:

a first server comprising at least one microprocessor and memory storing computer instructions for execution by the at least one microprocessor, the computer instructions including;

instructions for receiving a request for particular content from one of a plurality of clients, the request comprising a domain name;

instructions for sending the request for particular content, the request being sent to an origin server;

instructions for, in response to the request, receiving a message comprising the particular content and one or more tokens distinct from the particular content;

instructions for storing the particular content in a local cache of the first server in association with one or more tokens;

instructions for sending a subscription request to the origin server, the subscription request comprising a request to receive invalidation messages for tokens associated with the domain name over a subscription channel associated with the domain name;

a second server comprising at least one microprocessor and memory storing computer instructions for execution by the at least one microprocessor, the computer instructions including;

instructions for receiving a write request from a same and another one of the plurality of clients, the write request comprising the domain name;

instructions for sending the write request to the origin server;

instructions for receiving a response to the write request from the origin server;

instructions for determining that the response to the write request includes a particular token;

instructions for, upon said determination, sending a first invalidation message for the particular token to a third server;

the third server comprising at least one microprocessor and memory storing computer instructions for execution by the at least one microprocessor, the computer instructions including;

instructions for receiving the first invalidation message and upon said receipt, sending a second invalidation message for the particular token to the first server, the second invalidation message sent over the subscription channel associated with the domain name;

the first server'"'"'s computer instructions further including;

instructions for receiving the second invalidation message for the particular token from the third server, and upon said receipt, determining that the particular token is one of the one or more tokens, and based at least in part upon said determination, marking the particular content stored in the local cache as invalid.