Load Balancing Web Service by Rejecting Connections

• US 20110161980A1
• Filed: 12/31/2009
• Published: 06/30/2011
• Est. Priority Date: 12/31/2009
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Rothschild, Jeffrey, English, Robert M.
• Granted Patent: US 8,769,541 B2
• Time in Patent Office: Days
• US Class Current: 718/105
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: H04L 43/08 Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
• CPC Class Codes: H04L 47/70 Admission control; Resource allocation

Abstract

A load balancer allocates requests to a pool of web servers configured to have low queue capacities. If the queue capacity of a web server is reached, the web server responds to an additional request with a rejection notification to the load balancer, which enables the load balancer to quickly send the rejected request to another web server. Each web server self-monitors its rejection rate. If the rejection rate exceeds a threshold, the number of processes concurrently running on the web server is increased. If the rejection rate falls below a threshold, the number of processes concurrently running on the web server is decreased.

Claims

1. A method of distributing requests among a pool of web servers, the method comprising:

receiving, at a load balancer, a request from a first client;

sending, by the load balancer, the request to a first web server;

receiving, at the load balancer, a rejection notification from the first web server; and

sending, by the load balancer, the request to a second web server.

2. The method of claim 1, wherein the request comprises a TCP request.

3. The method of claim 1, wherein the first and second web server are selected through round-robin scheduling by the load balancer from a pool of web servers.

4. The method of claim 1, further comprising:

responsive to receiving a threshold number of rejection notifications from the pool of web servers in response to the request, sending the request to a different port of a web server from the pool for priority treatment of the request.

5. A method of managing a web server, the web server having a queue for receiving requests, the method comprising:

receiving, at a web server, a request from a client via a load balancer;

responsive to space being available in the queue, placing the request in the queue; and

responsive to the queue being full, sending a rejection notification to the load balancer.

6. The method of claim 5, wherein the web server is configured to have queue capacity of one.

7. The method of claim 5, further comprising:

self-monitoring a rejection rate for requests to the web server.

8. The method of claim 7, further comprising:

responsive to the rejection rate exceeding a threshold, spawning a new process for processing requests.

9. The method of claim 7, further comprising:

responsive to the rejection rate being lower than a threshold, ending one of a plurality of processes executing on the web server.

10. A load balancer comprising:

a client interaction module that receives a request from a first client; and

a server interaction module that sends the request to a first web server, receives a rejection notification from the first web server, and sends the request to a second web server from a pool of web servers.

11. The load balancer of claim 10, wherein the request comprises a TCP request.

12. The load balancer of claim 10, further comprising a scheduling module that selects the first and second web server through round-robin scheduling from a pool of web servers.

13. The load balancer of claim 10, wherein the server interaction module is configured to send the request to a different port of a web server from a pool of web servers responsive to receiving a threshold number of rejection notifications from the pool of web servers.

14. A web server comprising:

a queue; and

a queue manager that receives a request from a client via a load balancer, and responsive to space being available in the queue, places the request in a queue, and responsive to the queue being full, sends a rejection notification to the load balancer.

15. The web server of claim 14, wherein the queue has a queue capacity of one.

16. The web server of claim 14, further comprising a process manager that monitors a rejection rate for requests to the web server.

17. The web server of claim 16, wherein responsive to a rejection rate exceeding a threshold, the process manager spawns a new process for processing requests.

18. The web server of claim 16, wherein responsive to a rejection rate being lower than a threshold, the process manager ends one of a plurality of processes executing on the web server.