Method and apparatus for providing decentralized load distribution
View Patent ↗The present invention discloses a method and apparatus for providing decentralized load sharing, e.g., providing a protocol that can be used to address load balancing without the need of a central proxy device. Specifically, in one embodiment, the present invention addresses the decentralized load sharing problem, by allowing a group of servers to detect when more than one server attempts to process the same “work request” from a client. To address the potential conflict, each server will implement a back-off strategy for negotiating which server will actually process the work request.
1. A method for servicing a request by a first server, comprising:
receiving a broadcasted request from a client;
delaying for a period of time, wherein said period of time is calculated from a measure of busyness of a server;
announcing a claim for servicing the entire request that includes a first ranking of the first server;
receiving a second claim from a second server for servicing the entire request that includes a second ranking of the second server;
determining, by the first server, whether a potential collision has been detected for serving the entire request by evaluating the first ranking of the first server against the second ranking of the second server;
servicing the request if said potential collision is resolved in the first server's favor; and
not servicing the request if said potential collision is not resolved in the first server's favor.
2. The method of claim 1 , wherein said period of time is based on a randomly generated period of time.
3. The method of claim 1 , further comprising: announcing a commencement of servicing the request.
4. The method of claim 1 , wherein said broadcasted request is received in accordance with Internet Group Management Protocol.
5. The method of claim 1 , wherein said broadcasted request is received from a broadcast channel.
6. A non-transitory computer-readable medium having stored thereon a plurality of instructions, the plurality of instructions including instructions which, when executed by a first processor, cause the first processor to perform a method for servicing a request, comprising:
receiving a broadcasted request from a client;
delaying for a period of time, wherein said period of time is calculated from a measure of busyness of a server;
announcing a claim for servicing the entire request that includes a first ranking of the first processor;
receiving a second claim from a second processor for servicing the entire request that includes a second ranking of the second processor;
determining, by the first processor, whether a potential collision has been detected for serving the entire request by evaluating the first ranking of the first processor against the second ranking of the second processor;
servicing the request if said potential collision is resolved in the first processor's favor; and
not servicing the request if said potential collision is not resolved in the first processor's favor.
7. The computer-readable medium of claim 6 , wherein said period of time is based on a randomly generated period of time.
8. The computer-readable medium of claim 6 , further comprising announcing a commencement of servicing the request.
9. The computer-readable medium of claim 6 , wherein said broadcasted request is received in accordance with Internet Group Management Protocol.
10. The computer-readable medium of claim 6 , wherein said broadcasted request is received from a broadcast channel.