On resource sharing and careful overbooking for network virtualisation

Authors:
Markus Fiedler
Affiliations:
Blekinge Institute of Technology, School of Computing, SE-371 79 Karlskrona, Sweden
Venue:
International Journal of Communication Networks and Distributed Systems
Year:
2011

Citing 8
Cited 0

Resource-Sharing and Service Deployment in Virtual Data Centers

ICDCSW '02 Proceedings of the 22nd International Conference on Distributed Computing Systems
Network Virtualization: Breaking the Performance Barrier

Queue - Virtualization
A Distributed and Autonomic Virtual Network Mapping Framework

ICAS '08 Proceedings of the Fourth International Conference on Autonomic and Autonomous Systems
Resource overbooking and application profiling in a shared Internet hosting platform

ACM Transactions on Internet Technology (TOIT)
The Optimization of Xen Network Virtualization

CSSE '08 Proceedings of the 2008 International Conference on Computer Science and Software Engineering - Volume 03
Characterizing User-Level Network Virtualization: Performance, Overheads and Limits

ESCIENCE '08 Proceedings of the 2008 Fourth IEEE International Conference on eScience
On combining temporal partitioning and sharing of functional units in compilation for reconfigurable architectures

IEEE Transactions on Computers
Bandwidth allocation and admission control in ATM networks with service separation

IEEE Communications Magazine

Quantified Score

Hi-index	0.00

Visualization

Abstract

Network virtualisation implies the need for sharing network resources. However, exclusive resource allocation to users drives blocking probabilities and cost while potentially leaving precious resources underutilised in constrained environments. Motivated by these observations, this paper analyses careful overbooking according to Service Level Agreements that specify desired degrees of availability. Besides of full availability of the requested resources, a second level of limited availability, implying a well-defined reduction of the allocated resource, is taken into account. Particular attention is paid to the gain borderline, representing the possibility of accommodating one extra user beyond exclusive allocation without violating the SLAs. Simple but telling formulae provide insights into requirements for careful overbooking, worst-case capacity reduction factors from full to limited availability, and the conditions under which it is sensible to integrate users with different activity levels. We also link capacity reductions to potential decreases in quality of experience.