Dear all,
The minutes taken from the GNI BOF are shown below.
T.Kudoh
presents the agenda of the GNI BOF.
G.
Zervas presents the Grid Network Interface (proposal to create a working
group).
2
Drafts are proposed for the GNI: 1st Requirements and Architecture,
2nd interface specification.
Short
discussion about the standards track at OGF (standard requires two independent
implementations)
Questions
about submitting the standard to other bodies like IETF and OIF. These bodies
try to address similar issues but with a different scope (e.g. OIF standardizes
the UNI interface to a GMPLS control plane). It is also noted that these are
completely different disconnected communities (vendors/telcos/...)
G.
Zervas points out that some vendors are involved and collaborating in the GNI
activity (Nortel, KDDI, Alcatel – Lucent). They need to be part of the
effort, and also there should be some grid people represented.
F.
Travostino asks about name of the GNI BOF middleware and mentions that can be
adopted by a non grid middleware, because the name of the interface suggests
that is only for grid applications interfacing the network. Maybe MNI
(Middleware NI) would be a better name to attract communities other than the
grid one?
G.Zervas
answers that the GNI covers all the use cases (not just grid, also other
middlewares and plain applications), but the name Grid Network Interface is
because of the OGF.
F.
Dijkstra asks if it is only an interface to a single domain in a network or to
a control plane. G. Zervas answers that it can be both, but that it is the
network part that handles these details.
E.
Grasa presents the ongoing GLIF efforts to define a common API to request
network services (based on the work of existing projects like Phosphorus, IDC,
Autobahn, G-Lambda, ...)
Discussion
about GLIF being an alternative track to the OGF GNI initiative. It is
clarified that the GLIF work has a shorter timeframe and a more experimental
focus: the goal is to have a first implementation that can be used by the
community in a relatively short period of time (about 6 months) and then
iterate to improve the interface. However the work of the GLIF will be done in
coordination with the OGF GNI activity and can provide valuable feedback for
the GNI specification. It is pointed out that the GLIF is not a standards body,
so it cannot and does not pretend to create any alternative standard.
Joan
Antoni presents Harmony (Phosphorus WP1 system) interfaces, focusing on the
northbound interface (which allows grid applications to reserve network
resources).
S.
Figuerola encourages other people to participate in the process of defining the
interface between Harmony and other projects with a similar scope (Phosphorus
G2MPLS, IDC, Autobahn, G-Lambda, Enlightened).
I.Monga
asks if, in case the GNI was standardized, this would meet the requirements of
such interface. S. Figuerola answers that they have to finish gathering the
requirements for the interface between Harmony and other projects.
I.
Monga makes a presentation about bridging the gap between grids and networks.
G.
Zervas comments that Inder’s presentation is very consistent with the GNI
because it must be generic enough to accommodate different network services
(circuits, L1/2/3 VPNs, etc).
David
asks if the interface has mixed operations (some for the administrator and some
for the applications). It is clarified that both the applications and users can
access the GNI.
J.
Vollbrecht presents the DCN network controlled by the IDC.
T.
Miyamoto presents the needs and requirements for the GNI from a telecom
operator point of view.
F.
Travostino says he will wait for the other BOF before having any judgement.
End
of GNI BOF
You can find the GNI BOF minutes also in a doc format on the
link below:
http://www.ogf.org/gf/event_schedule/attendees.php?id=1279
Thanks again for your participation, interesting and
valuable comments and views that are taken on board and considered during the
next phase of creating the joint WG.
If you feel that something is missing or is not accurate
please don’t hesitate to point it out.
Best Regards,
George
---------------------------------------------------------------------------
Georgios
Zervas MIET, MIEEE
Senior
Research Officer
Photonic
Network Research Lab
Department
of Electronic Systems Engineering
University
of Essex,
Colchester,
CO4 3SQ, UK
Tel
: +44 (0) 1206 874233
web
: http://privatewww.essex.ac.uk/~gzerva/
---------------------------------------------------------------------------