Nokia Internship - Utility-based Congestion Control 

Belgium 

408083169

Low-Latency Low-Loss Scalable Throughput, or in short L4S (IETF RFCs 9330-9332), has been developed to offer low latency to traffic flows sharing the same bottleneck. In the L4S framework, the bottleneck marks packets it forwards depending on how close it is to being congested and the Prague congestion control determines the sending rates of the flows based on the percentage of marks they see, so that the aggregate (i.e., the sum of all throughputs of all flows) comfortably fits in the capacity of the bottleneck, and hence, the packets do not experience queueing latency. While default Prague congestion control offers the same throughput to every flow, an extension, referred to as Utility-Based Congestion Control (UBCC), in which each flow is allowed to interpret the measured marking ratio in a differentiated way, enables an (e.g., an augmented, mixed or virtual reality) application to (temporarily) assign more throughput to objects that the user is momentarily paying attention to. UBCC currently operates well when the users' attention is stable, i.e., when the users do not switch their attention too rapidly from one object to another.

You will be responsible 1) for further modifying UBCC so that it is able to respond to more frequent user attention switching and 2) for performing experiments with this extended version of UBCC in this dynamic context across different networks (PON, Wi-Fi, 5G, etc.).

What you need for this internship is:

• Good understanding of relevant network protocols such as TCP/IP, RTP, etc.
• Good understanding of congestion control algorithms.
• Demonstrated software programming proficiency (C/C++/Scripting/Linux).
• Basic understanding of multimedia codec technology.
• Fluency in English (both written and spoken).

Nokia has received the following recognitions for its commitment to inclusion & equality:

• One of the World’s Most Ethical Companies by Ethisphere
• Gender-Equality Index by Bloomberg
• Workplace Pride Global Benchmark