Minutes interim-2018-nmrg-02: Tue 16:00
||Minutes interim-2018-nmrg-02: Tue 16:00
NMRG 47th meeting
co-located with IEEE NOMS 2018, Taipei, Taiwan
NMRG Chairs: Lisandro Granville (UFRGS) and Laurent Ciavaglia (Nokia)
Tuesday, April 24th, 16:00-18:00 (Taiwan time)
Participants: 24 (see blue sheets)
1.Meeting introduction, NMRG chairs (5min.)
Laurent presented an overview of the NMRG scope, mission, meetings, research
topics, and recent activities.
2. Autonomic networking retrospective, Laurent Ciavaglia, NMRG co-chair (10
min.) Laurent presented an initial work done with Jéferson and Lisandro on on
historical perspective of autonomic networks comprising research and standards.
The analysis tries identifying gaps and potential research directions
considering new technologies (MEC, NFV, SDN
) or new environments (5G, IoT
Alex C.: clarify the goal of the effort. A Laurent: identify gaps (diff. from
RFC 7575/7576 and ANIMA) and areas where research can be done.
3. 5G Networks must be autonomic, Sven van der Meer, Ericsson (15 min.)
Sven presented issues for the management of 5G networks stressing missing
standardized and simple models for resource, network, service, coordination;
etc. and missing abstractions for e.g. control loops. Q Alex C.: now some YANG
models describing L2/L3 services. Q Alex G.: not only data models needed, but
most importantly common IM fit for the purpose of the autonomic networks (e.g.
semantics). Q Laurent: wouldnt the common denominator of generic resource
model be too small to be meaningful and usable? Q Alex C.: a useful point for
the multiple models would be to be mappable to workflows as currently the
various models are fragmented and there is no simple/documented way currently
4. Intent based networking, concepts and definitions, Alexander Clemm, Huawei
(15 min.) Alex C. presented a progress status on the IBN topics (incl. 2 IDs).
Q Alex G.: -Policy systems are not widely used because the over time the
management of the policy system becomes more complex than the managed system.
Suggestion to ass scalability in the list of issues to investigate in this
effort. -A monolithic approach will not work and is difficult to be adopted in
running/evolving systems. Need to fragment the approach. -Networks are
multi-domain and non-uniform. Need to consider these aspects in the effort from
5. Exploiting external event detectors to anticipate resource requirements for
the elastic adaptation of SDN/NFV systems, Pedro Martinez-Julia, NICT (15 min.)
Pedro presented on the inclusion of external events in the control-loop of
network management to improve the response of the system such as better
resource requirements anticipation and provisioning. Q Alex G.: suggestion to
add also service parameters/events in the sensing functionality.
6. Multi-MANO interworking for the management of multi-domains networks, Alex
Galis, UCL (15 min.) Alex G. presented work done at UCL on the integration of
single domain NFV orchestrator (SONATA project) and multi-domain orchestrator
(5GEx project) and lessons learned from this work, identifying some gaps, open
issues and priorities.
7. Towards data-driven management, Kohei Shiomoto, TCU (15 min.)
--> cancelled due to agenda conflict.
8. Analytics and security monitoring, Jerome Francois, INRIA (15 min.)
Jérôme presented work done in the HSL at INRIA comprising the use of AI/ML
techniques to improve insights and security responses, suggesting to develop
more largely network-dedicated toolbox for AI/ML, define better metrics,
features and algorithms.
9. Converging SDN and NFV at the network edges: the lightMANO approach, Roberto
Riggio, Create-Net / FBK (15 min.) --> cancelled due to agenda conflict.
10. Open discussion:
- Discussion for next steps on research items 4, 5, 6, 7, 8, 9
- Gaps, challenges, and propositions for an actualized road-map towards next
generation autonomic networking (items 2 and 3)
--> very limited time left for the discussion.
In summary, future actions will be to:
+ Extend the Autonomics 2.0 item with further inputs (e.g. Sven) and
propose an I-D to document the gaps and potential research items. +
Continue progress on the IBN work including remarks from the meeting. +
Further discuss, clarify and evaluate the new propositions (Pedro, Alex
G., Jérôme) and how they could create new research items for the NMRG.
Next meeting will be co-located with IETF 102, Montreal, Canada, in week 15-20
11. Demo Multi-MANO interworking, Alex Galis, UCL
--> cancelled due to time limit.