Sharing bandwidth among multiple users of network applications
A system and method are provided that allocate bandwidth on a global large scale network. Bandwidth allocation is based on the predicted bandwidth demands of the network users. Each user may be assigned one of a plurality of different values that determines the amount of bandwidth allocated to that user. In instances where an application runs on behalf of a plurality of remote clients, a system and method is provided that allows for the allocation of bandwidth based each individual remote client.
1. A method of allocating bandwidth over a network, the method comprising:
receiving, on a network, a request for a shared-service application to run on a host device on behalf of a plurality of remote clients;
receiving an estimation of a volume of network traffic associated with running the shared-service application for each remote client;
classifying the plurality of remote clients, based at least in part on the received estimation, as either being in a first classification or a second classification, wherein classifying the plurality of remote clients includes:
aggregating the estimation of the volume of network traffic associated with each of the remote clients to create a network traffic total;
selecting at least some of the plurality of remote clients in an order of descending estimation of the volume of network traffic;
upon selecting each remote client, aggregating the estimation of the volume of network traffic for all selected remote clients;
stopping the selection of remote clients when the selected remote clients have an aggregated estimation of the volume of network traffic that is above a predetermined fraction of the network traffic total;
classifying all selected remote clients into the first classification; and
classifying all remote clients that have not been selected into the second classification;
transmitting classifications of the plurality of remote clients to the host device; and
allocating bandwidth for the shared-service application,
wherein remote clients classified with the first classification are allocated bandwidth individually, and wherein the remote clients classified with the second classification are provided bandwidth as part of a group.
2. The method of claim 1 , wherein classifying the plurality of remote clients includes:
determining whether the estimation of the volume of network traffic for each remote client is below a predetermined threshold;
classifying all remote clients having estimation of the volume of network traffic that is not below the predetermined threshold with the first classification; and
classifying all remote clients having estimation of the volume of network traffic that is below the predetermined threshold with the second classification.
3. The method of claim 1 , wherein allocating bandwidth includes providing bandwidth for each remote client having the first classification based at least in part on one of a plurality of weight values associated with each of the remote clients.
4. The method of claim 1 , further comprising:
receiving bandwidth usage information for a plurality of applications running on behalf of a plurality of clients, including the shared-service application running on behalf of the plurality of remote clients;
aggregating bandwidth usage information to determine a total bandwidth demand;
detecting congested links within the network; and
dividing the network's available bandwidth between the plurality of applications based, at least in part, on a weight value associated with the remote clients.
5. The method of claim 4 , wherein the bandwidth demand information is based on peak bandwidth usage of the plurality of applications over a predetermined time period.
6. The method of claim 4 , further comprising:
determining that all network bandwidth demand is being met;
determining that unused bandwidth is available on the network; and
dividing the unused bandwidth among the plurality of clients in proportion to the client's weight values.
7. The method of claim 4 , wherein a portion of network bandwidth remains exempt from the operations to create bandwidth allocations.
8. The method of claim 4 , wherein a subset of the clients receive a permanent allocation of bandwidth.
9. A network device for allocating bandwidth over a network, the device comprising:
a memory; and
a processor coupled to the memory, the processor configured to execute instructions to:
receive a request for a shared-service application to run on a host device on behalf of a plurality of remote clients;
receive an estimation of a volume of network traffic associated with running the shared-service application for each remote client;
classify the remote plurality of clients, based at least in part on the received estimation, as either being in a first classification or a second classification, wherein classifying the plurality of remote clients includes:
aggregating the estimation of the volume of network traffic associated with each of the remote clients to create a network traffic total;
selecting at least some of the plurality of remote clients in an order of descending estimation of the volume of network traffic;
upon selecting each remote client, aggregating the estimation of the volume of network traffic for all selected remote clients;
stopping the selection of remote clients when the selected remote clients have an aggregated estimation of the volume of network traffic that is above a predetermined fraction of the network traffic total;
classifying all selected remote clients into the first classification; and
classifying all remote clients that have not been selected into the second classification;
transmit classifications of the plurality of remote clients to the host device; and
allocate bandwidth for the shared-service application,
wherein remote clients classified with the first classification are allocated bandwidth individually, and wherein the remote clients classified with the second classification are provided bandwidth as part of a group.
10. The device of claim 9 , wherein the processor is further configured to execute instruction to:
determine whether each of the remote client's estimation of the volume of network traffic is below a predetermined threshold;
classify all remote clients having estimation of the volume of network traffic that is not below the predetermined threshold with the first classification; and
classify all remote clients having estimation of the volume of network traffic that is below the predetermined threshold with the second classification.
11. The device of claim 9 wherein instructions to allocate bandwidth includes providing bandwidth for remote clients having the first classification based, at least in part, on one of a plurality of weight values associated with each of the remote clients.