Browse Prior Art Database

Method and mechanism to upgrade a distributed service originated system in production mode

IP.com Disclosure Number: IPCOM000237861D
Publication Date: 2014-Jul-17
Document File: 7 page(s) / 261K

Publishing Venue

The IP.com Prior Art Database

Abstract

This disclosure provides a way to upgrade the large systems step by step without halting the whole system. In this solution , there will be an upgrade service running as a third party service, its main responsibility is to work as an adapter to forward the request between the components in the whole system during the upgrade. The router and transformation rules can be configured and changed dynamically so the upgrade service knows the relationship between the components in the whole system, and the whole system can still work smoothly during the upgrade period. After the whole system upgrade is complete, the upgrade service can be shut down, and the upgraded system can start to work normally.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 01 of 7

Method and mechanism to upgrade a distributed service originated system in production mode

In a large distributed system, the software upgrade is a difficult job. It is difficult because usually each components in the whole system needs to be upgraded in the same time since the different version may have different interfaces which may make it only work with the other component with the specific version. More and more large systems require to work 24*7 , to halt the whole system to do the upgrade is not acceptable. It only affords a limited time frame for the upgrade. For example, in a banking system, the software upgrade can only take the time frame in 2:00AM to 5:00AM everyday. However, the system upgrade and migration is time consuming and hard to implement in the production environment. Usually the upgrade or migration should be separated into different short steps and finish them one by one in the different time frame. This may require to separate the whole job into several steps and the system need to be workable before the whole job get done.

This disclosure provides a way to upgrade the large systems step by step without halting the whole system. In this solution , there will be set up an upgrade service running as a third party service, its main responsibility is to work as an adapter to forward the request from component A to component B and transform the request to the valid type that can be accepted by the component B' after component B is upgraded to component B', and also transform the response data into the style that is accepted by component A. The router and transformation rules can be configured and changed dynamically so the upgrade service knows the relationship between the components in the whole system, and the whole system can still work smoothly during the upgrade period. After the whole system upgrade is complete, the upgrade service can be shut down, and the upgraded system can start to work normally. Compared to the existing method to do system upgrade and migration, the solution can avoid the system halt and is able to make the upgrade and migration progress more smoothly and stable .

The upgrade server combined with several components and a queue to deliver the request and messages. Below is the component diagram for the system.

1



Page 02 of 7

2



Page 03 of 7

The component of the distributed system, which could be a service and need to interact with other components of the system. In this diagram, we represent a process to m...