Browse Prior Art Database

System method and apparatus to predict the best restore mechanism for over-cloud infrastructure by heuristic analysis Disclosure Number: IPCOM000244701D
Publication Date: 2016-Jan-06

Publishing Venue

The Prior Art Database


Disclosed is a System method and apparatus to predict the best fit restore mechanism for cloud infrastructure . Recommendation or Prediction is made based on analysis of reliability and speed of various backup types used previoulsy for restoring the same or similar infrastructure.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 40% of the total text.

Page 01 of 10

System method and apparatus to predict the best restore mechanism for over -cloud infrastructure by heuristic analysis

Undercloud specified in this art is a basic virtualization mechanism or (cloud with basic set of virtualization capabilities) running on one or more physical servers
designed to deploy, test, manage, and update the overcloud servers. where as overcloud is the functional cloud available to run guest virtual machines and workloads. Multiple overclouds could be available on an undercloud. A typical example for undercloud is a linux physical machine with KVM and virtualization capabilities and an cloud installed and running in one or more VMs in this machine could be considered as an overcloud

When the cloud administrator deploys a new overcloud, they will be able to set up various backup mechanisms depending upon the capability of undercloud and overcloud APIs and services. Fig. 1 shows a typical example with simple overcloud deployed on a single VM provisioned by underlying undercloud virtualization mechanisms. Now following are the backup options available to administrator

Backup Mechanism


Configuration Files Backup of VM1 using overcloud capabilities

1) One typical way is to ssh into VM1 and save all the required configuration files of critical services.

2) Store this as a Archived compressed file


Page 02 of 10

in external storage

VM Snapshot using undercloud capabilitiesUsing the undercloud APIs take the snapshot of VM1 at various scheduled times. Based on the capability of undercloud external live snapshot of VM1 is also possible without down time.

There are many other ways of backup mechanisms that can be implemented

On an event of overcloud service failure, the cloud administrator will need to be aware of the technical benefits or disadvantages of selecting a particular backup type from the available list of backups, for restoration of the overcloud . Otherwise all issues/merits may be uncovered on the failure/success of restoration of one or more overcloud services partially or completely.Current cloud management systems are not having capability to predict a best fit restore strategy for overcloud infrastructure and services based on past experience and user inputs.Disclosed method aims at optimizing this aspect

The main point here is to provide efficient and effective link between past experience and usage of backup types in current restore event.

Disclosed method proposes

a) a mechanism by which past data on various cloud restore events can be accumulated
b) use the above collected data and format it in such a way that following recommendations can be made

Recommendation 1: Recommendation of Backup Types That can restore the fastest

Recommendation 2: Recommendation of Most Reliable Backup Types

Recommendation 3: Recommendation of Most Reliable and Fastest Backup Types

Centralized Heuristic DB

Centralized Heuristic DB is the central part of the solution proposed.

Point 1:

This DB can accept Restor...