Browse Prior Art Database

System, Method and Apparatus to predict, learn and adapt to the best network strategy and configuration in a cloud.

IP.com Disclosure Number: IPCOM000244665D
Publication Date: 2016-Jan-06
Document File: 7 page(s) / 97K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a System method and apparatus to predict, learn and adapt to the best network strategy and configuration in a cloud. Prediction and Adaptation is made based on identified scenarios and rules which works based on the data about the cloud and user input

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 34% of the total text.

Page 01 of 7

System, Method and Apparatus to predict , learn and adapt to the best network strategy and configuration in a cloud.

When the cloud administrator deploys a new cloud, the selection of the type of network required for the cloud has to be made. This choice of the network type differs based on the business purpose for which this cloud is created and optimal network selection will improve the efficiency of the resources used in cloud.

Often the cloud administrator will not be aware of the technical benefits or disadvantages of selecting a particular network type for the cloud and this will be uncovered in case of performance issues when a cloud is already deployed and many virtual servers are already running critical business applications.

Various problems in this context


a) Current cloud management systems are not having capability to predict a best fit network strategy based on user/administrator inputs and the environmental analysis at the time of deployment.


b) Current cloud management systems are not having capability to adapt and flip between network strategies based on changes in the environment.


c) Cloud environment also are incapable of predicting and reacting to network topology and strategy adjustments based on ongoing current usage of the cloud and its resource trends.

This article aims at suggesting ideas to optimize these aspects.

This article suggests method to predict best fit network strategy and identified scenarios and rules that can be used to do this prediction.

It also suggests method to adapt to best fit network strategy and identified scenarios and rules that can be used to do this adaptation.

The proposed solution is divided into two parts like predictive and adaptive algorithms.

For the discussion below we will use example from openstack (

1


Page 02 of 7

https://en.wikipedia.org/wiki/OpenStack) however these ideas are adaptable to any cloud.

Predictive Algorithm:

The predictive algorithm works by asking some user-friendly objective queries to the user during cloud deployment. The answers selected by the user are processed by the algorithm to determine the ideal type of network (flat, GRE, vlan) needed for the cloud yet to be deployed.

Below given (Fig.1) is the flow chart which summarizes the prediction process

Note: Take user inputs will use queries listed in the section "Standard scenarios identified and associated queries for prediction"

Predictive Analysis will arrive into conclusion/result based on the answers to the queries as explained in "Standard scenarios identified and associated queries for

prediction"

Standard scenarios identified and associated queries for prediction

2


Page 03 of 7

Below given is the set of standard scenarios identified and associated queries, which we are proposing as part of the predictive algorithm , through which a best fit prediction can be made

Note : In future arts - more queries can be added to improve the accuracy of the algorithm and more network strategies can be add...