Laboratoire d'informatique de l'École polytechnique


PhD defense of Mohammed Hawari

Speaker: Mohammed Hawari
Location: Online
Date: Tue, 6 Jul 2021, 09:00 - Fri, 9 Jul 2021, 11:00

Mohammed Hawariwill defend his doctoral thesis for the degree of doctor in Computer Science on Tuesday, July 6th at 9:00 AM (GMT+1). Due to the COVID pandemic, this defense will happen remotely.

This defense will discuss System and Networking Aspects of the Transition of High-Performance Applications from Dedicated to Commodity Hardware: the Example of Media Production for Professional Broadcast.

Abstract: Due to the increasing performance offered by commodity servers and to the general availability of multi-gigabit Ethernet-based networking hardware, a growing number of performance-intensive and network-intensive applications are being migrated from dedicated to commodity hardware. Examples thereof include scientific computing and network-packet processing, historically implemented by dedicated super-computing clusters and by dedicated packet-processing hardware, respectively. However, media production for professional broadcast (i.e., the process by which multiple audiovisual sources are mixed and processed, in real-time, to elaborate the audiovisual stream as it will be consumed by the final viewer) is still being implemented with dedicated hardware equipment, based on the Serial Digital Interface (SDI), an interconnection technology carrying the legacy of analog video. Despite an ongoing industrial effort to replace SDI with IP-based interconnection — as specified by the SMPTE 2022-6 and 2110 standards — the delay-sensitive nature of media production still challenges its total transition to software running on commodity servers. This thesis solves different aspects of that problem.

First, the high rates and low jitter-tolerance of media production packet streams have motivated a quantitative and qualitative study of the sources of jitter undergone by those streams when they are processed by commodity servers. In addition to results specific to Linux x86_64 servers, that work has yielded a general jitter exploration methodology, applicable to any operating system and hardware commodity servers. Second, a generic platform enabling the implementation of custom high-accuracy instrumentation for hardware-based packet timestamping has been developed. By exposing a high-level programming interface — relying on the P4 language — that platform, despite being FPGA-based, allow network and broadcast operators with little hardware design skills to specify custom logic for line-rate packet processing and timestamping. In particular, such instrumentation can be used to qualify the jitter properties of media production streams. Third, a system to perform packet-pacing — i.e., the transmission of a constant-rate packet stream with negligible jitter — has been proposed. By exclusively but cleverly relying on commodity hardware, that work invalidates the common belief according to which software-based media-production is impossible on commodity servers (due to the jitter they introduce). The proposed system has been formally and experimentally proven to yield a jitter, conforming to the requirements of media production streams. Finally, a software framework easing the implementation of media-production applications has been developed. That framework relies on a separation between media processing and media transport: the media processing logic receives and transmits full media frames (e.g., video frames) from the media transport logic, which handles high-performance packet processing with techniques such as zero-copy and kernel bypass networking. Those last techniques have been shown to notably increase the scalability of media production on commodity servers.

Link to the online conference

Poste d'administrateur-trice systèmes et réseaux (concours CNRS)

Un poste a été ouvert au concours externe pour recruter un administrateur ou une administratrice systèmes et réseaux au LIX.

Le candidat ou la candidate recruté·e rejoindra le pôle « Ingénierie système et réseau » du service informatique du laboratoire. Celui-ci assure l’administration de l’infrastructure et des services déployés au laboratoire. Il accompagne également les chercheurs et les chercheuses du LIX pour l’exploitation des moyens de calcul, de prototypage, de développement de logiciels, et d’expérimentation spécifiques aux recherches menées par les équipes.

Les détails du poste et la procédure pour candidater se trouvent sur le site du CNRS:

Date limite de candidature : Mercredi 30 juin 2021 à 13h00 (heure de Paris).

Talk by Sonia Vanier: «Distributed Denial of Service cyber-attacks in 5G networks: a robust approximation approach »

Speaker: Sonia Vanier
Location: Online
Date: Thu, 24 Jun 2021, 14:30-15:30

For a new seminar of the proofs and algorithms pole of LIX, we are happy to welcome Sonia Vanier for her talk entitled Distributed Denial of Service cyber-attacks in 5G networks: a robust approximation approach.

Abstract: Distributed Denial of Service (DDoS) cyberattacks represent a major security risk for network operators and internet service providers. They thus need to invest in security solutions to protect their network against DDoS attacks. The present work focuses on deploying a network function virtualization based architecture to secure a network against an on-going DDoS attack. We assume that the target, sources and volume of the attack have been identified. However, due e.g. to 5G network slicing, the exact routing of the illegitimate flow in the network is not known by the internet service provider. We seek to determine the optimal number and locations of virtual network functions in order to remove all the illegitimate traffic while minimizing the total cost of the activated virtual network functions. We propose a robust optimization framework to solve this problem. The uncertain input parameters correspond to the amount of illegitimate flow on each path connecting an attack source to the target and can take values within a predefined uncertainty set. In order to solve this robust optimization problem, we develop an adversarial approach in which the adversarial sub-problem is solved by a Branch & Price algorithm. The results of our computational experiments, carried out on medium-size randomly generated instances, show that the proposed solution approach is able to provide optimal solutions within short computation times

Link to the online conference

Talk by Krzysztof Ziemiański: « Tracks in Higher Dimensional Automata »

Speaker: Krzysztof Ziemiański
Location: Online
Date: Wed, 23 Jun 2021, 11:00-12:00

For a new seminar of the proofs and algorithms pole of LIX, we are happy to welcome Krzysztof Ziemiański.

Abstract: The main goal of my talk is to present two models of executions of Higher Dimensional Automata. A track in a HDA is a sequence of cells such that every cell is either an upper face or an upper coface of the preceding one. To every track we can assign the (evi)pomset of its events; it is a labeled interval order equipped with an additional secondary ordering. Then we consider the set of tracks between fixed source and target cells, up to certain equivalence. This set is naturally a presheaf over the category of evipomsets, which is our first model of executions of HDA. This model can be simplified. When we restrict to step-sequence executions, ie, tracks in which the consecutive cells have only one common vertex, we obtain a presheaf over permutahedral category. This model is combinatorially simpler. At first glance it seems too restrictive but there is some evidence that this is not the case: the space of topological executions is homotopy equivalent to the geometric realization of the permutahedral model.

Link to the online conference

The list of next seminars can be found at:

The calendar of seminars can be found at:

Talk by Alex Kavvos: « Client-Server Sessions in Linear Logic »

Speaker: Alex Kavvos
Location: BBB
Date: Wed, 23 Jun 2021, 14:00-15:00

For a new seminar of the proofs and algorithms pole of LIX, we are happy to welcome Alex Kavvos.

Abstract: We introduce coexponentials, a new set of modalities for Classical Linear Logic. As duals to exponentials, the coexponentials codify a distributed form of the structural rules of weakening and contraction. This makes them a suitable logical device for encapsulating the pattern of a server receiving requests from an arbitrary number of clients on a single channel. Guided by this intuition we formulate a system of session types based on Classical Linear Logic with coexponentials, which is suited to modelling client-server interactions. We also present a session-typed functional programming language for client-server programming, which we translate to our system of coexponentials.

Link to the online conference

The list of next seminars can be found at:

The calendar of seminars can be found at: