Description
Date depot: 1 janvier 1900
Titre: A network traffic measurement infrastucture for the characterization of Internet Multimedia Content.
Directeur de thèse:
Dario ROSSI (LTCI (EDMH))
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Non defini
Resumé:
The ever increasing amount of multimedia content delivered to Internet users is changing significantly the properties of traffic carried by network operators. The need to better understand the nature of such traffic is fundamental to build analytical models, either for network performance prediction and capacity planning but also to design traffic control and content placement mechanisms based on such knowledge [1-7].
Traffic engineer practitioners have had a number of tools to deploy telecommunication networks, like the Erlang formula for phone networks and Internet traffic theory to design, provision and operate IP based networks at large scale to carry data services. However, such knowledge is not enough to design content delivery networks where bandwidth based practices turn out to be a partial solution to the whole problem.
The advent of CDN infrastructures, from the very beginning operated by over-the-top players, has shown how much performance gain the whole content delivery chain, from the content provider to the end user can obtain. Telco CDNs are the first attempt for that a traditional network operator to better control its own resources (assets) and deliver novel services at lower cost carrying greater value.
The ambition of the thesis is to define a rigorous framework to collect and analyze today multimedia traffic and to propose a simplified characterization traffic modeling framework than can feed performance modeling of next generation networks.
The main preliminary tasks are the following:
-definition of a scalable distributed network measurement -infrastructure to collect information about the delivered media content, with special care to video delivery
-definition of required metrics to measure and the associated methodology to infer such metrics.
-Collection of large database of network traffic metrics
analysis of such metrics vis à vis the impact on next generation traffic engineering, based on network bandwidth and storage management.
Doctorant.e: Imbrenda Claudio