Folks: Pursuant to the "State of the WG" thread [1], a number of
WG participants (Richard Y., Sabine R., Wendy R., Jensen Z., Gao, K)
exchanged some thoughts as a prelude to a larger WG discussion. This
email serves as a summary of the the ideas exchanged by them and the
start of a larger WG discussion.
Clearly there is a need to reboot ALTO.
The protocol is stable, the abstractions in the protocol (PID, endpoint
cost service, network maps) are sound and very useful, but perhaps what
is lacking is a proselytizing push of the protocol to other WGs who may
have an interest in the capabilities offered by ALTO but may not be
aware of it.
Richard Yang noted that there is a large body of interest in ALTO
outside of the IETF:
- The ALTO project in OpenDaylight [2]
- Use of ALTO at Yale, Caltech, and other institutions (Richard,
please fill in with more communities of interest ...)
- Early use of ALTO in service provider networks
This is a good sign in the generality of the protocol and the acceptance
of it.
Beyond this, there are other avenues that hold promise to raise the
profile of ALTO:
1) At the risk of wading into politics, the WG has to avoid the "low-
energy" label, which has known to be detrimental in other non-technical
spheres. To this extent, the larger working group has to be more
engaged in list discussions, reviewing documents, and evangelizing the
protocol within IETF. The WG has seen strong contributions from Lyle
Bertz, Haibin Song, Lingli Deng, among others, and these contributions
are very much appreciated and encouraged. A number of WG participants
have been reviewing documents over the last few weeks. A note of
appreciation to Gao Kai, Jensen Zheng, and Qiao Xiang is in order ...
thanks to all.
2) Take a look at other WGs to see how ALTO may help. Sdnrg, i2rs and
supa come to mind, although this is by no means an exhaustive list.
The work in ALTO on multi-cost and alto-calendar may could be of
interest to sdnrg and should be shared with them more formally.
Regarding i2rs, while ALTO will not --- and should not --- serve as a
southbound interface, it has the capacity and capability to serve as
one of the northbound interfaces [3].
Supa develops models to express policies using YANG, could it also
develop policies using ALTO? Clearly, YANG and ALTO are oriented
towards two different layers, the former is designed for controlling
and querying network devices while the latter has strong high-level
abstractions (PID, endpoint cost service, network maps) that aid in
reasoning at higher levels. While YANG is predominantly focused on
fixed networks, ALTO abstractions are useful regardless of fixed,
wireless, or cellular networks. Can supa use these abstractions to
develop policies for applications?
3) Evangelize/proselytize/market ALTO in area-wide meetings, such as
tsvarea or opsarea. The plan is to prepare an overview of the ALTO
protocol and its advantages in time for a slot at the Berlin IETF
tsvarea meeting.
So ... what to you folks think? How do we reboot ALTO and maintain its
relevance? Any further thoughts and ideas expressing them would be
great.
Thanks for reading so far!
[1] https://www.ietf.org/mail-archive/web/alto/current/msg02996.html
[2] https://wiki.opendaylight.org/view/ALTO:Main
[3] Gurbani, V.K., Scharf, M., Lakshman, T.V. and Hilt, V.,
"Abstracting network state in Software Defined Networks (SDN) for
rendezvous services," In Workshop on Software Defined Networks (SDN),
held in conjunction with IEEE International Conference on
Communications (ICC), June 2012.
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Nokia Networks
1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60563 (USA)
Email: vkg@{bell-labs.com,acm.org} / [email protected]
Web: http://ect.bell-labs.com/who/vkg/ | Calendar: http://goo.gl/x3Ogq
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto
