Routing under heavy loading
View Patent ↗According to the invention, a delivery network for assisting delivery of content objects over an Internet is disclosed. The delivery network includes a network outlet, an interface and a routing function. The network outlet is coupled to a plurality of full-route networks, where each of the plurality of full-route networks is capable of delivering content objects to a plurality of terminal networks. The plurality of terminal networks include a terminal network, where the plurality of terminal networks are coupled to a plurality of end user computers. The interface receives content objects for delivery to the plurality of end user computers. The routing function routes content objects in at least two modes, where a first mode routes content objects based upon a first route path from the network outlet to the terminal network, and a second mode routes at least some content objects using a second route path from the network outlet to the terminal network. The first route path is chosen based upon delivery efficiency. Switching from the first mode to the second mode is triggered when at least of a portion of the first route path reaches a predetermined level of use. The first and second route paths are different, and the second route path is less efficient than the first route path.
1. A network system that delivers content objects from one or more content providers over the Internet to one or more end users, comprising:
a content request interface for receiving content object requests for delivery of content objects to the one or more end users; and
a traffic distributor configured to
determine an efficient routing path from one of the content providers to one of the end users based on efficiency and without considering effects of loading, by selecting among a plurality of full-route networks of the Internet, comprising a first full-route network and a second full-route network, to establish the efficient routing path, wherein
the efficient routing path includes at least one of the first and second full-route networks, an egress point of the first full-route network to a terminal network that serves the one of the end users, and an egress point of the second full-route network to the terminal network;
process content object requests received from the one of the end users through the content request interface such that the at least one of the first and second full-route networks is utilized;
detect a routing bottleneck in the efficient routing path, wherein:
the routing bottleneck is associated with a routing element that comprises at least one of the first and second full-route networks, the egress point of the first full-route network, and the egress point of the second full-route network, and
the routing bottleneck is detected based on utilization of the routing element in excess of a threshold;
determine a sub-optimal routing path from the one of the content providers to the one of the end users that is less efficient under light load conditions than the efficient routing path, but does not include the routing bottleneck;
process content object requests received from the one of the end users through the content request interface while utilizing the sub-optimal routing path;
detect that the utilization of the at least one of the first and second full-route networks, the egress point of the first full-route network, and the egress point of the second full-route network has fallen below the threshold; and
resume processing content object requests received from the one of the end users through the content request interface along the efficient routing path.
2. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein the efficient routing path includes network hops, and the traffic distributor determines the efficient routing path based on a number of the network hops.
3. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein the efficient routing path includes network hops, and determining the efficient routing path comprises determining the efficient routing path based on speed of one or more routers between the network hops.
4. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein the efficient routing path connects one of the content providers directly with one full-route network and connects the egress point of the one full-route network directly with the terminal network.
5. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 4 , wherein the sub-optimal routing path connects the one of the content providers with the one full-route network, and connects the egress point of the one full-route network with at least a second full-route network before connecting with the terminal network.
6. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein the sub-optimal routing path replaces the routing element associated with the routing bottleneck with a more costly routing element.
7. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein the sub-optimal routing path is determined from a plurality of sub-optimal routing paths that differ by random assignment of differing full-route networks among the plurality of sub-optimal routing paths.
8. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein the traffic distributor determines the sub-optimal routing path after a content object request is received from the one of the end users through the content request interface.
9. The network system that delivers content objects from the one or more content providers over the Internet to one or more end users as recited in claim 1 , wherein determining the sub-optimal routing path includes determining a most efficient egress point for delivery to the one of the end users.
10. A method of optimizing network delivery of content objects from one or more content providers over the Internet to one or more end users, comprising:
determining, utilizing a traffic distributor, an efficient routing path from one of the content providers to one of the end users based on efficiency and without considering effects of loading, wherein:
the traffic distributor selects among a plurality of full-route networks of the Internet, comprising a first full-route network and a second full-route network, to establish the efficient routing path, and
the efficient routing path includes at least one of the first and second full-route networks, an egress point of the first full-route network to a terminal network that serves the one of the end users, and an egress point of the second full-route network to the terminal network;
processing content object requests from the end users such that the at least one of the first and second full-route networks is utilized;
detecting, utilizing the traffic distributor, a routing bottleneck in the efficient routing path, wherein:
the routing bottleneck is associated with a routing element that comprises at least one of the first and second full-route networks, the egress point of the first full-route network, and the egress point of the second full-route network, and
the routing bottleneck is detected based on utilization of the routing element in excess of a threshold;
determining, utilizing the traffic distributor, a sub-optimal routing path from the one of the content providers to the one of the end users that is less efficient under light load conditions than the efficient routing path, but does not include the routing bottleneck;
processing content object requests from the one of the end users while utilizing the sub-optimal routing path;
detecting, utilizing the traffic distributor, that the utilization of the at least one of the first and second full-route networks, the egress point of the first full-route network, and the egress point of the second full-route network has fallen below the threshold; and
resuming processing content object requests from the end users along the efficient routing path.
11. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein the efficient routing path includes network hops, and the traffic distributor determines the efficient routing path based on a number of the network hops.
12. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein the efficient routing path includes network hops, and determining the efficient routing path comprises determining the efficient routing path based on speed of one or more routers between the network hops.
13. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein the efficient routing path connects one of the content providers directly with one full-route network and connects the egress point of the one full-route network directly with the terminal network.
14. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 13 , wherein the sub-optimal routing path connects the one of the content providers with the one full-route network, and connects the egress point of the one full-route network with at least a second full-route network, before connecting with the terminal network.
15. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein the sub-optimal routing path replaces the routing element associated with the routing bottleneck with a more costly routing element.
16. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein determining the sub-optimal routing path comprises determining a plurality of sub-optimal routing paths that differ by random assignment of differing full-route networks among the plurality of sub-optimal routing paths.
17. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein determining the sub-optimal routing path occurs after a content object request is received from the one of the end users.
18. The method of optimizing network delivery of content objects from the one or more content providers over the Internet to one or more end users as recited in claim 10 , wherein determining the sub-optimal routing path includes determining a most efficient egress point for delivery to the one of the end users.
19. A computer readable device with instructions for optimizing network delivery of content objects from one or more content providers over the Internet to one or more end users, comprising computer-executable code for:
determining, utilizing a traffic distributor, an efficient routing path from one of the content providers to one of the end users based on efficiency and without considering effects of loading, wherein
the traffic distributor selects among a plurality of full-route networks of the Internet, comprising a first full-route network and a second full-route network, to establish the efficient routing path, and
the efficient routing path includes at least one of the first and second full-route networks, an egress point of the first full-route network to a terminal network that serves the one of the end users, and an egress point of the second full-route network to the terminal network;
processing content object requests from the end users such that the at least one of the first and second full-route networks is utilized;
detecting, utilizing the traffic distributor, a routing bottleneck in the efficient routing path, wherein:
the routing bottleneck is associated with a routing element that comprises at least one of the first and second full-route networks, the egress point of the first full-route network, and the egress point of the second full-route network, and
the routing bottleneck is detected based on utilization of the routing element in excess of a threshold;
determining, utilizing the traffic distributor, a sub-optimal routing path from the one of the content providers to the one of the end users that is less efficient under light load conditions than the efficient routing path, but does not include the routing bottleneck;
processing content object requests from the one of the end users while utilizing the sub-optimal routing path;
detecting, utilizing the traffic distributor, that the utilization of the at least one of the first and second full-route networks, the egress point of the first full-route network, and the egress point of the second full-route network has fallen below the threshold; and
resuming processing content object requests from the end users along the efficient routing path.