Web serving system
AssigneeXcelera
InventorMark E. Kriegsman
Life of this patent
Estimating the term…
Abstract
A method for distributing files among a set of servers in a content-delivery system, includes selecting a file stored on a first server from the set of servers and associating a first replication value with that selected file. Then, at least in part on the basis of the first replication value, a determination is made as to whether the selected file is to be replicated at a second server selected from the set of servers.
Claims17
In a network comprising a set of servers forming a content-delivery system, a computer-implemented method for distributing files among the servers in the set of servers, the method comprising:
- selecting a file stored on a first server from the set of servers;
- associating a first replication value with the selected file;
- and
- determining, at least in part on the basis of the first replication value, that the selected file is to be replicated at a second server selected from the set of servers, comprising
- defining a file set on the second server;
- associating a second replication value with the file set;
- comparing the first replication value with the second replication value;
- and
- on the basis of the comparison, determining whether the selected file is to replace the file set;
- wherein associating the first replication value comprises determining the first replication value at least in part on the basis of a file usage pattern associated with the selected file, and
- determining the first replication value further comprises building a statistical model representative of an extent to which transmission of the selected file consumes transmission capacity.
A method for causing a content delivery comprising a set of servers interconnected in a network to serve content, the method comprising:
- on a first server, identifying a file to be replicated, the file including the content;
- associating a first replication value with the file;
- replicating the file on servers selected from a set of secondary servers, comprising replacing a file set on one of the selected servers with the file to be replicated based on a comparison of the first replication value associated with the file and a second replication value associated with the file set;
- in response to a request for the content, selecting a server for serving the file;
- and
- serving the file from the selected server;
- wherein associating the first replication value comprises determining the first replication value at least in part on the basis of a file usage pattern associated with the selected file, and
- determining the first replication value further comprises building a statistical model representative of an extent to which transmission of the selected file consumes transmission capacity.
A content-delivery system comprising a plurality of servers interconnected in a computer network, the servers including:
- a first server having a file to be replicated, the file having a first replication value in excess of a threshold;
- and
- a set of secondary servers, each secondary server configured to replace a file set on the secondary server based on a comparison of a second replication value associated with the file set and the first replication value associated with the file from the first server;
- wherein the first server includes computer hardware and each secondary server includes computer hardware, and
- wherein the first replication value is associated with the file to be replicated based at least in part on the basis of a file usage pattern associated with the file, and
- the first replication value is determined at least in part by building a statistical model representative of an extent to which transmission of the file consumes transmission capacity.
Dependencies are inferred from the claim text.