Research Article
Virtualized Application Networking Infrastructure
@INPROCEEDINGS{10.1007/978-3-642-17851-1_29, author={H. Bannazadeh and A. Leon-Garcia and K. Redmond and G. Tam and A. Khan and M. Ma and S. Dani and P. Chow}, title={Virtualized Application Networking Infrastructure}, proceedings={Testbeds and Research Infrastructures. Development of Networks and Communities. 6th International ICST Conference, TridentCom 2010, Berlin, Germany, May 18-20, 2010, Revised Selected Papers}, proceedings_a={TRIDENTCOM}, year={2012}, month={10}, keywords={Networking Testbed Network Architecture Service-Oriented Architecture}, doi={10.1007/978-3-642-17851-1_29} }- H. Bannazadeh
A. Leon-Garcia
K. Redmond
G. Tam
A. Khan
M. Ma
S. Dani
P. Chow
Year: 2012
Virtualized Application Networking Infrastructure
TRIDENTCOM
Springer
DOI: 10.1007/978-3-642-17851-1_29
Abstract
In this paper, we present a new platform for experimenting with networked systems and distributed applications called Virtualized Application Networking Infrastructure (VANI). This infrastructure is designed as a converged communications and computing infrastructure that would facilitate operation of an open applications marketplace. VANI enables introduction of new network architectures that require in-network (hardware-accelerated) content processing and storage. We describe the VANI architecture and the resources it provides. VANI has two main planes; control and management plane, and applications plane. VANI resources are virtualized and made available to the researchers and application providers through a service-oriented control and management plane. The current VANI resources are processing, storage, networking and various software-based resources. VANI also includes a new reprogrammable hardware resource that enables experimenting with hardware-based or hardware-accelerated networking algorithms and protocols. We present performance evaluations of this reprogrammable hardware resource, and the VANI virtual networking mechanism. The results show that by using the reprogrammable hardware resource, researchers can evaluate high performance and high throughput networking algorithms as easily as evaluating software-based networking algorithms.
