Browse Prior Art Database

Data-centric predictive cloud migration based on user modelling

IP.com Disclosure Number: IPCOM000238359D
Publication Date: 2014-Aug-20
Document File: 3 page(s) / 115K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed are a system and method for the migration of virtual appliances across data centers based on user-application modelling.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 3

Data-

Virtual appliances are indifferent to the type and location of the physical hardware on

which the appliances run. This characteristic feature enables the users to install and configure any application stack in the virtual appliance (e.g., configure a virtual machine as a web service solution stack), and be able to migrate to any physical node at any data center for execution.

This article addresses users in need of a migration solution for virtual appliances across data centers as opposed to on-the-spot instantiation. These end-users have installed and configured applications and processes on virtual appliances that enclose trade secrets on processes and methodology, which, if disclosed, could severely affect the user's revenue stream.

Migration of virtual appliances is not new. Current work includes methods for optimizing migration across data centers with metrics around performance , input/output (I/O), and cost. The majority of this work relates to High Performance Computing (HPC) applications. However, no work relates to the migration of virtual appliances that are based on the combined logic of application -usage modelling and data-locality with respect to a series of sequential blocks of tasks as in a workflow .

The novel contribution is a system and method for the migration of virtual appliances across data centers based on user-application modelling. The proposed method leverages existing techniques for migration, such as live and offline migration. It also uses existing techniques to identify data associated with an application , the location of such data sources, the users of applications and data, and the various quality parameters (e.g., cost, time, bandwidth, etc.) for use.

The method comprises modelling of end -user application use, end-user location, and input and intermediate data produced by applications . Based on these models, the method executes a set of computational units within a virtual appliance by optimally scheduling the migration of the appliance, and/or the data. The virtual appliance is a solitary computing unit or a set of computing units with required states ; therefore, the virtual appliance must be migrated and cannot be otherwise provisioned at the destination. The data and/or virtual appliance migration is decided based on the optimality of various metrics (e.g...