This is the top of the page.
Displaying present location in the site.
  1. Home
  2. About NEC
  3. Research and Development @NEC Laboratories Europe
  4. Research Areas
  5. ICT Virtualization Platform
Main content starts here.

ICT Virtualization Platform

Research on Software-Defined Networking (SDN) and Network Function Virtualization (NFV) at NEC Laboratories Europe (NLE) is based on two visions: A super-fluid ICT infrastructure where resources are fine grained enough to optimally use available resources and to provide customized individual per-user services, and a network operating system (NetOS) which gives full programmability to distributed compute and network resources.

Super-fluid ICT Infrastructure

Some of the biggest challenges for realizing a super-fluid ICT infrastructure are today's limitations of granularity of virtual machines and traffic flows and limitations of high-speed software switching. At present, virtual machines (VMs) are heavy-weight instances that are far too heavy to use a single VM per user for encapsulating the individual service. At the same time, SDN hardware switches have scalability limitations concerning the number of flows that they can distinguish and limitations regarding the dynamics of flow creations. These restrictions force network operators to aggregate flows instead of treating each user's flows individually. Our research on SDN and NFV is targeted at overcoming these limitations and significantly improving user plane performance of virtualization platforms to a level where (virtualized) network functions in data centers can run at full network speed. Core technologies developed at our lab include:

  • high speed virtual switches able to i) offer a virtual switching capacity up to 150 Gigabit per second (Gbps) ii) go beyond 20 Gbps when forwarding from virtual machines to physical network interfaces and iii) accelerate standard switching technologies like Open vSwitch by 200% at minimum.
  • very light-weight high-performance virtual machines able to run many network functions (e.g., firewall, load balancer, tunnel end point, DPI, Carrier Grade NAT, Software BRAS) while processing packets at rates beyond 10 Gbps
  • highly scalable OpenFlow switches that combine high throughput with high scalability towards millions of individual flows and flow dynamics 1-2 orders of magnitude beyond state of the art

The combination of these performance boosting technologies allows for highly efficient fine-grained use of resources in data centers and enables individual per-user services even in large operator networks, as well customized industrial applications in resource-constrained environments.

Network Operating System

SDN and NFV are basic technologies that enable software control of network and compute resources. For non-trivial deployments, this software is complex and needs proper software engineering. In order to make software for SDN and NFV more structured and increase the speed of development and the quality of code, we develop a Network Operating System (NetOS), that structures software into the application-specific code, libraries, demons, drivers, and a kernel.

Application-specific code specifies services that might be composed of multiple functions at multiple locations in the network. Libraries represent those functions, for example, in the form of a virtual machine that implements a gateway. Demons are active components providing a service to other services. For example, a topology demon provides up-to date information of the network and server topology. Drivers connect to the hardware infrastructure, such as servers and switches. Network services will progressively build upon the NetOS framework and NFV platform. One example is NLE's software Broadband Remote Access Server (BRAS) application that displays the capabilities of this new approach.


  • This research area has received funding from the European Union under the H2020 Grant Agreement n. 671648 (Project "VirtuWind", Virtual and programmable industrial network prototype deployed in operational Wind park,
  • This research area has received funding from the European Union under the H2020 Grant Agreement n. 644866 (Project "SSICLOPS", Scalable and Secure Infrastructures for Cloud Operations,
  • We launched our open source initiative “Cloud Networking Platform” featuring high speed virtual switches and very light-weight high-performance virtual machines. Please check it out at A list of our talks is also available at



  • Unikernels Everywhere: The Case for Ephemeral CDNs. Simon Kuenzer, Anton Ivanov, Filipe Manco, Yuri Volchkov, Florian Schmidt, Kenichi Yasukata, Michio Honda, Felipe Huici. ACM SIGPLAN/SIGOPS International Conference on Virtual Execution Environments (VEE’17)


  • Building Ephemeral CDNs with MiniCache. Simon Kuenzer, Anton Ivanov, Filipe Manco, Jose Mendes, Iurii Volchkov, Florian Schmidt, Kenichi Yasukata, Felipe Huici. USENIX Annual Technical Conference (USENIX ATC '16), Poster Session,


  • Rollback Recovery for Middleboxes . Justine Sherry (UC Berkeley), Peter Gao (UC Berkeley), Soumya Basu (UC Berkeley), Aurojit Panda (UC Berkeley), Arvind Krishnamurthy (University of Washington), Christian Macciocco (Intel Research), Maziar Manesh (Intel Research), Joao Martins (NEC Labs), Sylvia Ratnasamy (UC Berkeley), Luigi Rizzo (University of Pisa), Scott Shenker (UC Berkeley and ICSI). [best student paper award]
    ACM SIGCOMM 2015, London, United Kingdom, August 17-21, 2015.
  • IN-NET: In-Network Processing for the Masses. Radu Stoenescu, Vladimir Olteanu, Matei Popovici, Mohamed Ahmed, Joao Martins, Roberto Bifulco, Filipe Manco, Felipe Huici, Georgios Smaragdakis, Mark Handley, Costin Raiciu. European Conference on Computer Systems (EuroSys 2015), April 2015,
  • Lost in Network Address Translation: Lessons from Scaling the World's Simplest Middlebox. Vladimir Olteanu, Felipe Huici, Costin Raiciu. ACM SIGCOMM Workshop on Hot Topics in Middleboxes and Network Function Virtualization (HotMiddlebox 2015), August 2015,
  • The Case for the Superfluid Cloud. Filipe Manco, Joao Martins, Kenichi Yasukata, Jose Mendes, Simon Kuenzer, Felipe Huici. USENIX Workshop on Hot Topics in Cloud Computing (HotCloud 2015), July 2015,
  • mSwitch: A Highly-Scalable, Modular Software Switch. Michio Honda, Felipe Huici, Giuseppe Lettieri, Luigi Rizzo. [best paper award] ACM Sigcomm Symposium on SDN Research (SOSR 2015), June 2015,
  • Dietz, Thomas; Bifulco, Roberto; Manco, Filipe; Martins, Joao; Kolbe, Hans-Joerg; Huici, Felipe, "Enhancing the BRAS through virtualization," Network Softwarization (NetSoft), 2015 1st IEEE Conference on , vol., no., pp.1,5, 13-17 April 2015, doi: 10.1109/NETSOFT.2015.7116144,


  • M. Honda, F. Huici, C. Raiciu, J. Araujo, L. Rizzo: “Rekindling Network Protocol Innovation with User-Level Stacks”, ACM SIGCOMM Computer Communications Review (CCR), April 2014. Available at
  • J. Martins, M. Ahmed, C. Raiciu, V. Olteanu, M. Honda, R. Bifulco, F. Huici: “ClickOS and the Art of Network Function Virtualization”, 11th USENIX Symposium on Networked Systems Design and Implementation (NSDI), January 2014. Available at
  • M. Dusi, R. Bifulco, F. Gringoli, F. Schneider: Reactive Logic in Software - Defined Networking: Measuring Flow - Table Requirements , Proceedings of 5th International Workshop on TRaffic Analysis and Characterization (TRAC)
  • R. Bifulco, M. Dusi: Reactive Logic in Software - Defined Networking: Accouting for the limitations of the Switches , Proceedings of EWSDN 2014
  • J. Rückert, R. Bifulco, R. Ul Haq, H.J. Kolbe, D. Hausheer: Flexible Traffic Management in Broadband Access Networks using Software Defined Networking, Proceedings of IEEE NOMS 2014



  • M. Honda, F. Huici, L. Rizzo: “MiniStack: Operating System Support for Fast User-space Network Protocols”, Poster at 10th USENIX Symposium on operating Systems Design and Implementation (OSDI),
  • R. Bifulco, M. Brunner, R. Canonico, P. Hasselmeyer, F. Mir : A practical experience in designing an OpenFlow controller, European Workshop on Software - Defined Networking (EWSDN’12)

Top of this page