Agenda IETF100: nfvrg
agenda-100-nfvrg-03

* Welcome and administrivia ~ 7.5 min
  + Scribe(s) and notetaker(s)
  + Agenda bashing
  + RG matters and other announcements
* Introduction to research activities ~ 105 min
  + Rethinking NFV: Supporting Efficient Packet Processing
    (E. Jacob, UPV/EHU)
    The idea is to summarize the evolution of the original NFV in which
    every packet processing/manipulation needed to be done at the VM
    (computing) level, towards a SDN-Enabled NFV, in which part ot the
    processing is done at the switch (networking) level. 
    The difficult part is that although the interest is clear, there
    have been approaches that try to cover this objective from both sides,
    either trying to make computing based solutions to perform better
    and better, mainly by clever approaches regarding interfacing to NICs
    on x86 platforms or by improving the datapath to allow a more extensible
    packet processing.
    Every approach has their own particularities and costs, but at the
    end of the day it appears that a superior abstraction is needed for
    describing/managing a networking element, in the same way the computing
    element has. Clearly we need to elevate the role of networking in the
    SDN/NFV framework, to integrate packet processing in the networking
    infrastructure to be able to deploy efficient solutions.
  + Elastic Adaptation of SDN/NFV Systems to Dynamic Service Demands
    (P. Martinez-Julia, NICT)
    The incorporation of SDN/NFV to current computer and network system
    infrastructures is constantly increasing, becoming essential for the
    particular case of edge/branch network systems. The systems supported
    by these infrastructures require to be adapted to environment changes
    within a period of time that escapes human abilities. Therefore, we
    are developing the Autonomic Resource Control Architecture (ARCA) to
    achieve the fast detection and adaptation of virtual computer and
    network systems to changes in their operation environments, also keeping
    close-to-optimum resource allocations to reduce monetary and operative
    cost as much as possible. In this presentation we report our approach
    to evaluate such architecture by building a multi-domain SDN/NFV
    infrastructure based on OpenStack and deploying ARCA to adapt a virtual
    system based on the edge/branch network concept to the operational
    conditions of an emergency support service, which is rarely used
    but that cannot leave any user unattended.
  + Network slicing support by dynamic VIM instantatiation
    (S. Clayman, UCL)
    There are some scenarios in which it is important to have a separate
    Data Center slice within a full Network Slice. These include services
    for which there are critical attributes that need to be met. We
    present the case for creating a VIM on-demand for these DC slices,
    show some of the attributes of such a slice, and the increased value
    of a per slice VIM. The  architectural elements that are required to
    support such a model are shown. Finally a set of layered abstractions
    using slicing elements are presented, showing how they fit together
    for service provisioning. 
* Open mic on matters related to the RG ~ 7.5 min
  As a reminder, these are the matters the RG has agreed to focus on:
  - Re-architecting functions
  - New management frameworks
  - Techniques to guarantee dataplane features
  - Measurement and benchmarking