Cross-layer Wireless Video Research
As video traffic over the internet continues to rapidly grow, network providers continue to look for improved solutions to support the demands on resources and client devices offer opportunities for improved differentiation. We are focusing on how to devise a network architecture and protocols for streaming video that leverages existing http and TCP/IP infrastructure support over the internet (CDNs, network caching, web servers, etc.), while developing a cross-layer optimization framework that incorporates context information from the users, applications, client devices and network infrastructure, etc. to improve the quality of experience (QoE).
The project involves building on ideas to exploit various contextual information from devices, network infrastructure, etc. to develop cross-layer optimizations for improved TCP performance in video streaming scenarios. The key tasks are to identify the various pieces of context information that are relevant in video streaming scenarios, and develop the means to interface with and exploit this information with the video streaming application, in the networking stack implementation on the client device, and if necessary, in the network infrastructure. The objective is to study the QoE improvements that result from such approaches in various wireless networking scenarios spanning low/high mobility, small/large RTT, etc.
This work would require the use of Opnet modeling, including incorporation of video streaming applications along with networking modules and RF link layer models in order to characterize the QoE performance for video in terms of video quality and e2e latency, as well as other metrics such as radio resource used (total bytes transferred, IP packets, etc.), client throughput, network capacity improvements, etc.. We would also investigate different mechanisms for gathering network, client device, and user state to provide context input to the algorithms for optimization of the QoE.
MS or PhD candidates in Computer Science or Electrical Engineering.
Candidate should have 6+ months of work or educational experience with internet video streaming solutions, relevant internet protocols and the use of modeling/simulation tools such as Opnet, Matlab and C/C++ for simulations that combine application layer, network transport layer and radio link layer behavior.
Feb 12, 2013
Mar 13, 2013
Intel Labs is the company's world-class, industry leading research organization, responsible for driving Intel's technology pipeline and creating new opportunities. The mission of Intel Labs is to deliver breakthrough technologies to fuel Intel's growth. This includes identifying and exploring compelling new technologies and high risk opportunities ahead of business unit investment and demonstrating first-to-market technologies and innovative new usages for computing technology. Intel Labs engages the leading thinkers in academia and industry in addition to partnering closely with Intel business units.
We will accept applications/resumes until 60 days after posting date or earlier at Intel’s discretion. Intel invites people of all ages currently enrolled in an academic institution (or graduated within the last 18 months) to apply.
A lot of companies like to talk about changing the world. But while others are making claims, Intel’s been making history—for...