Method and apparatus for load balancing a distributed processing system

Abstract

Method and apparatus for load balancing a distributed processing system. A method for load balancing a distributed processing system having a plurality of processing nodes associated therewith is provided. A portion of a process is first received at one of the processing nodes for processing of that portion thereat. The amount of processing at that node is then determined that will be required to complete the received portion of the process as a processing load on that one processing node. A determination is then made if the amount of processing exceeds a predetermined threshold. If the amount of processing exceeds the predetermined threshold, then information is transferred to other nodes in the network that the one of the processing nodes is unavailable for further processing.

Description

CROSS REFERENCE TO RELATED APPLICATIONS [0001] This application is Continuation Application of U.S. patent application Ser. No. 09 / 915,910, filed on Jul. 25, 2001, entitled “Method and Apparatus for Load Balancing a Distributed Processing System” (Atty Dkt. No. 25,512), which is a Continuation-in-Part of U.S. patent application Ser. No. 09 / 887,494, filed Jun. 22, 2001, entitled “Method And Apparatus For Converting Data Between Two Dissimilar Systems” (Atty Dkt No. ATTA-25514), which is a Continuation-in-Part of U.S. patent application Ser. No. 09 / 879,571, filed Jun. 12, 2001, entitled “Method and Apparatus for Generating Unique Id Packets in a Distributed Processing System” (Atty Dkt No. ATTA-25,515), which is a Continuation-in-Part of U.S. patent application Ser. No. 09 / 841,135, filed Apr. 24, 2001, entitled “System and Method for Transmission of Information Between Locations on a ComputerNetwork with the Use of unique Packets,” (Atty Dkt. No. ATTA-25,441).TECHNICAL FIELD OF THE IN...

AI Technical Summary

Benefits of technology

Facilitates efficient data classification and transmission, reduces CPU load, and minimizes errors by using unique ID codes for data packet routing and conversion, enhancing search capabilities and data management efficiency.

Problems solved by technology

The classification and management of data is one of the most difficult tasks faced by corporations, government entities, and other large users of information.

Data exchanges face a bigger challenge in that they must work with multiple companies and develop a comprehensive classification system for their buyers.

However, no existing classification database is comprehensive enough to address all the issues associated with building a classification system.

These issues include: uniform numbers for products that cross multiple industries, restricting products from inclusion in classification, and non-usage of slang or industry standard language to access or classify products.

The classification databases frequently do not address all the products, thus resulting in inconsistencies even when companies use the same classification system.

Additionally, many of the various classification systems conflict with each other.

This program requires companies to pay multiple fees and go through a lengthy administrative process.

Even then it may not cover all products in an industry.

These efforts can be costly, and for this reason, compliance is generally not high.

However, under current practices, most exchanges offer barely adequate search capabilities for their buyers.

In many instances the buyer will fail to find the product that they seek.

These existing processes could therefore be characterized as cumbersome, time consuming, frustrating and ineffective.

Another challenging data management task is the transmission of data between dissimilar systems.

Transmitting data between such systems can be a time-consuming and expensive task.

Thus, in adapter-based systems, CPU load considerations may affect when and how often data pulls can be scheduled.

It is well known that when large amounts of data are being transmitted between systems, a system error (i.e., stoppage) and / or data loss (i.e., dropout) may occur.

With conventional adapter-based system architectures, debugging a system stoppage can be very challenging because of the large number of conversion processes involved, and because most systems do not have an integrated way to indicate the point at which processing stopped, relying instead upon error logs.

For example, if the host system is used for OLTP (on line transaction processing) during the day, pulling bulk data from the host system in order to replace data lost in a previous data transfer may be delayed until the late night hours.

Of course, the delay in processing the data can have an adverse impact of its own.

Images