High-Throughput Inter-Domain Multicast on the Xen Hypervisor

Lieferzeit: Lieferbar innerhalb 14 Tagen

32,90 

Multiplexing sensor data streams to a dynamically changing set of guest operating systems hosted on the Xen hypervisor

ISBN: 3639478495
ISBN 13: 9783639478495
Autor: Landolt, Florian
Verlag: AV Akademikerverlag
Umfang: 104 S.
Erscheinungsdatum: 20.03.2015
Auflage: 1/2015
Format: 0.7 x 22 x 15
Gewicht: 173 g
Produktform: Kartoniert
Einband: Kartoniert
Artikelnummer: 7987938 Kategorie:

Beschreibung

This work is concerned with multiplexing sensor data streams to a dynamically changing subset of guest operating systems, hosted on the Xen hypervisor. The distribution of sensor data in virtualized environments constitutes an ambitious task. Especially one property of virtualization, namely isolation, impedes unrestricted access to sensor hardware. While being of major importance for domain security, several layers of isolation induce a decrease in overall performance by causing substantial communication overhead across isolation boundaries. A single sensor is able to produce several MBs of output per second. The throughput provided by Xens virtual network is not sufficient in this case. In order to realize rapid data transmission between domains, a socket-like, high-throughput IDC mechanism, named XenSocket, was forward-ported to current versions of Xen and Linux and a newly developed library, called libxmc, addresses the need for dynamic sensor-to-domain mapping. The here presented system was designed with general applicability in mind. Therefore, it is possible to distribute any kind of data stream.

Autorenporträt

The author worked at the Computational Systems Group of the University of Salzburg as a graduate student in the field of real-time and embedded systems. Currently he is leading a project concerned with securing mobile services conducted by Denuvo / Sony DADC.

Herstellerkennzeichnung:


BoD - Books on Demand
In de Tarpen 42
22848 Norderstedt
DE

E-Mail: info@bod.de

Das könnte Ihnen auch gefallen …