Smart Federation Platform for Mobile Cloud Computing Infrastructures

Project Details

Project Lead
Vladimir Getov 
Project Manager
Vladimir Getov 
Project Members
Vlad Cocoru  
Supporting Experts
Tak-Lon Wu, Fugang Wang  
Institution
University of Westminster, School of Electronics and Computer Science  
Discipline
Computer Science (401) 
Subdiscipline
11.04 Information Sciences and Systems 

Abstract

In recent years the concepts and implementations of modern distributed computing infrastructures have been developing rapidly in the area of cloud computing. However, in order to tackle many of the application grand challenges we need qualitatively new large-scale autonomous systems with smart properties. Such properties together with support for virtualization and distributed intelligence can change completely the quality of computerised services provided to mobile units in various defence scenarios. In particular an ad-hoc network federation of static data-centres and mobile cloud infrastructures designed and integrated in a smart cloud platform can be particularly useful and important for the middle level of command, bringing high quality location-aware services right at the field of operation. The cloud computing paradigm is quite suitable for this purpose but it needs significant improvements towards providing full autonomy.

Intellectual Merit

The objective of the proposed project is to design, implement and evaluate a platform integrating services to automatically adapt applications running on top of an infrastructure spanning multiple clouds. Federation composed of mobile clouds and static data centres will be considered. In this context, the proposed platform should enable application resilience by securely adapting the application to the evolving federation composition despite transient lack of connectivity between clouds within the federation and stringent resource constraints in mobile clouds.

Broader Impacts

The design and development of a Platform-as-a-Service layer on top of a cloud federation will take into account the fact that services/applications need to tolerate intermittent connectivity between the clouds. The idea is to have a federation of clouds with one (or several) large clouds accessible over a WAN and smaller mobile clouds close to the theatre of operations. Since the large cloud may not be always accessible the applications should be able to execute on the mobile clouds in degraded mode. This means that we would have to deal with replication and aggregation of services, security and consistency issues, as well as adaptation mechanisms. In addition, the project will consider and investigate further the potential for significant improvements in the area of security in our future autonomous cloud system.

Scale of Use

We want to run a set of relatively short experiments involving up to 64 VMs but typically not more than 10. The expected frquency is a couple of experimental runs per week.