Browse Prior Art Database

A Method or System to Validate the Applications Running on Multiple Nodes

IP.com Disclosure Number: IPCOM000238132D
Publication Date: 2014-Aug-04
Document File: 4 page(s) / 79K

Publishing Venue

The IP.com Prior Art Database

Abstract

A system or method to help the tester quickly validate the application on multiple nodes is disclosed. Our solution is to provide the "propagate mode" to the dispatcher (workload balancer): if the propagate mode is enabled on the dispatcher, the requests sent to certain port will be replicated & propagated to all nodes which the application is running on. After the requests are processed, the responses from each node will be collected & grouped by the dispatcher, and sent back to the user. With our solution, the user/tester will be able to access & validate the application on all nodes quickly with any test cases, and identify the problematic nodes easily.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 50% of the total text.

Page 01 of 4

A Method or System to Validate the Applications Running on Multiple Nodes

1


1.

.. The problem and known solutions

The problem and known solutions:

To ensure the availability & capability of Enterprise Applications, most of them are running on the Middleware (eg. Application Servers) which are hosted on multiple nodes, either managed by Middleware itself or controlled by a standalone manage system (eg. Cloud System). Typically, one or more workload balancer(s) dispatch the requests to different nodes per each node's workload. This architecture is proved to be effective & feasible, but it also brings the contingency & difficulty to testing: the tester's requests will be routed to any of the active nodes randomly (per workload), which may not hit the node on which the application is not working correctly.

A well-known solution to validate the applications running on multiple nodes is: to ask the user to provide the sample requests & expected responses, when the application is activated on the node, an agent will send the sample requests, and verify whether the responses match the expected ones. The drawback of this solution is, the sample requests & responses are provided by the user, but executed by an agent, which is not flexible enough. If the user wants to do some tests for new features, which have different test cases other than existing one-time verification test cases, the test requests still may not hit the problematic nodes.

Other similar solutions include:


(1) Automated testing of cluster data services https://www.google.com.hk/patents/US7370101?dq=validate+cluster+node&hl=en&sa=X&ei=N3Z5U4u0Hsr08QXRiYGQBA&ved=0CD4Q6AEwATgK

This invention discloses a system or method to automated test data service which are hosted in a cluster environment. The test package is needed to installed on each node of the cluster, and used to monitor the data service on each node. Its purpose & solution are different from our invention.


(2) Monitoring distributed software health and membership in a compute cluster https://www.google.com.hk/patents/US8108733?dq=validate+cluster+node&hl=en&sa=X&ei=N3Z5U4u0Hsr08QXRiYGQBA&ved=0CFMQ6AEwBDgK

This invention discloses a system or method to monitor software health in a computing cluster. This solution requires a watchdog component running on each node, which is used to monitor the neighboring node's health, and take actions accordingly if neighboring node is not responding. This solution could not resolve our pain points, and is different from our invention.

2


2.

.. Summary of our invention

Summary of our invention:

To help the tester quickly validate the application on all nodes, and identify the problematic nodes if any, our solution is to provide the "propagate mode" to the dispatcher (workload balancer): if the propagate mode is enabled on the dispatcher, the requests sent to certain port will be replicated & propagated to all nodes which the application is running on. After the requests are processed...