Method for determining load balancing weights using application instance topology information

• US 7,493,380 B2
• Filed: 12/02/2003
• Issued: 02/17/2009
• Est. Priority Date: 12/02/2003
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Baskey, Michael Edward, Aman, Jeffrey David, Bostjancic, David Vincent, Bivens, John Alan II, Dillenberger, Donna N., Yocom, Peter Bergersen, Arwe, John E.
• Primary Examiner(s): Avellino; Joseph E
• Application Number:
  US10/725,635
• Publication Number:
  US 20050120095A1
• Time in Patent Office: 1,904 Days
• Field of Search: 709217-219, 709223-229, 709238-248, 718100-108
• US Class Current: 709/223
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/101 based on network conditions
• CPC Class Codes: H04L 67/1023 based on a hash applied to IP addresses or costs

Abstract

An apparatus and method for distributing traffic across a group of machines using application instance statistics. In order to perform load balancing in accordance with the present invention, a method of generating weights to bias load balancing distributions is provided. The application instances to which traffic is being distributed, or the application middleware, are instrumented to establish certain metrics about the application while running. The application instance instrumentation will provide application statistics such as number of successful transactions, application response times, application topology, importance of transactions being processed, time the application is blocked waiting for resources, resource consumption data, and the like. These metrics are collected, processed, and then presented as a set of weights to the load balancing apparatus to govern its distribution of traffic. With such application metrics available, traffic can be disbursed based on the current state of the application instances and other application instances in the transaction'"'"'s path, the application instance'"'"'s likelihood to complete the request, or even higher level business-oriented goals.

Claims

1. A method, in a data processing system, of distributing traffic to application instances on one or more computing devices, comprising:

obtaining application instance specific operational information identifying operational characteristics of an application instance on a computing device of the one or more computing devices, wherein the application instance specific operational information includes an application instance topology;

comparing the application instance specific operational information to one or more other application instance specific operational information for one or more other application instances based on the application instance specific operational information obtained;

generating a load balancing weight based on a relationship between the application instance specific operational information and the one or more other application instance specific operational information;

attributing weight points to the application instance and the one or more other application instances based on a relative difference between the application instance specific operational information and the one or more other application instance specific operational information;

providing the load balancing weight to a load balancing device;

the load balancing device distributing the traffic to the application instance based on the load balancing weight;

wherein obtaining application instance specific operational information includes retrieving the application instance specific operational information from the application instance using an agent application residing on the computing device, and wherein the agent application identifies the application instance topology by sending a correlation in a request to an agent application associated with a second application instance, wherein application instance information is provided by the agent application associated with the second application; and

wherein the method is implemented in a weight management system that is separate from the computing devices and from the load balancing device.

2. The method of claim 1, wherein the application instance is instrumented to include code for communicating with the agent application and sending the application instance specific operational information to the agent application from the application instance.

3. The method of claim 1, wherein the relationship is a relative difference between the application instance specific operational information and the one or more other application instance specific information.

4. The method of claim 1, wherein retrieving the application instance specific operational information from the agent application is performed periodically.

5. The method of claim 1, wherein generating the load balancing weight based on the relationship between the application instance specific operational information and the one of more other application instance specific information includes:

assigning a base weight to each of the application instance and the one or more other application instances; and

increasing a weight value associated with the application instance or the one or more other application instances based on one or more of the following;

which of the application instance and the one or more other application instances has a relatively higher transaction success rate;

which of the application instance and the one or more other application instances operates on an underutilized system;

which of the application instance and the one or more other application instances processes the least significant transactions; and

which of the application instance and the one or more other application instances passes transactions on to higher performing computing systems.

6. The method of claim 1, wherein generating the load balancing weight based on the relationship between the application instance specific operational information and the one of more other application instance specific information includes:

assigning a base weight to each of the application instance and the one or more other application instances; and

increasing a weight value associated with the application instance or the one or more other application instances based on which of the application instance and the one or more other application instances processes the least significant transactions.

7. The method of claim 1, wherein generating the load balancing weight based on the relationship between the application instance specific operational information and the one of more other application instance specific information includes:

assigning a base weight to each of the application instance and the one or more other application instances; and

increasing a weight value associated with the application instance or the one or more other application instances based on which of the application instance and the one or more other application instances passes transactions on to higher performing computing systems.