Re: [irtf-discuss] [Icn-interest] Draft ICN RG Charter

[Paulo Mendes <[e-mail]>]

Hi Dirk

see my comments inline...

On May 24, 2011, at 1:41 PM, Dirk Kutscher wrote:

> Hi Paulo,
> 
> Thanks for the suggestions. Please my some comments inline.
> 
>> Since the list of research areas is not fully analyzed (work proposed
>> to be done as first milestone) I suggest to change the first objective
>> to "The main objective of the ICNRG is to advance the state of ICN
>> research in AT LEAST the mentioned areas, focusing on solutions that
>> are relevant for evolving the Internet at large."
> 
> OK, I agree.
> 
>> As for the expression "Internet at large" I would be a little more
>> precise, including some examples in order to avoid misunderstandings..
>> I would suggest "... Internet at large, including the support for new
>> services and operation into disruptive scenarios based for instance in
>> the present of a considerable number of low-capacity (computational,
>> storage) devices". The idea is to keep as an objective the analysis of
>> the interaction of information-centric networking with Internet-of-
>> services and Internet-of-connected-objects.
> 
> Here, I really disagree. We did not want to create the notion of different "Internets", i.e., Information-Centric, Internet-of-Services, Internet-of-Things etc. The message is that we want to assess ICN's applicability as a general Internet technology ("Internet at large").

I was not suggestion the creation of the notion of different "Internets", Just to say something in the charter about the applicability scenario at large... For instance, is ICNRG going to look at ICN solutions for the Internet as is it, or for the "expected" Internet which may be extended to different types of portable devices. But is all of us have this perspective in mind, than "Internet at large" is fine :)


> 
>> I would put more emphasis to work based on implementations than
>> simulations.. Simulations should be pointed out as acceptable to prove
>> the feasibility of proposed solutions in large-scale scenarios. Given
>> more importance to implementations is inline with the other mentioned
>> objective of ICNRG "The ICNRG will foster the development of ICN
>> testbeds for performing experiments with running code."
> 
> Right, so IMO that aspect is already covered. We also said "work that is based on implementation and simulation experiences will be given preference.".

RIght.. The question is that this sentence givens equal preference to simulations and testbed experiences. I was suggesting to give higher preference to testbed experiences.

> 
>> I would be careful with the word "architecture" since IETF/IRTF does
>> not standardize architectures. Hence I would propose to use the
>> following sentence...
>> 
>> "The ICNRG will investigate components of a common protocol framework
>> for information-centric networking, aiming to: i) identify protocols
>> for standardization, which may or may not re; ii) support an ICN
>> architecture with large applicability."
>> 
>> ... to replace the following two sentences:
>> 
>> "The ICNRG will identify key ICN architecture invariants across
>> different specific approaches which could form the basis of a future
>> ICN architecture."
>> "The intention is that one result from this work could be a common
>> protocol framework that can be used to identify protocols for
>> standardization. These protocols may or may not re-use existing IETF
>> protocols."
> 
> OK, got your point. It's true that "form the basis of a future ICN architecture" can be misinterpreted.
> 
> Here, we wanted to strike a balance: it's true that *system architecture* work is not the IETF's core business. However, for an IRTF research group, we expect some architecture work to be required, e.g., to find out about common core elements that you need to agree on (for later protocol specifications etc.). For example, RFC 4838 is a good example of architecture work that describes fundamental architecture concepts for DTN, which has been a good basis for subsequent DTNRG work on the Bundle Protocol spec.
> 
> Will think of a better formulation.

Good.. If we are talking about RFC4838 type of work, and not a 3GPP type of description this is expectable.

> 
> 
>> 
>> 
>> B) ICNRG Milestones
>> 
>> The charter has milestones to describe the main concepts and research
>> challenges, to provide a survey of different approaches and to describe
>> an ICN architecture.
>> 
>> What is missing here is the common protocol framework mentioned before.
>> So I'll add the following milestone as third one: "Based on the
>> identified main research challenges and the analysis of different
>> approaches, ICNRG will devise a common protocol framework for ICN,
>> aiming to have high applicability".
>> 
>> I would re-write the third ICNRG milestone, since saying that "...
>> long-term goals would be documents describing an ICN architecture.. "
>> since a little strange to see within IETF/IRTF work. Instead of saying
>> that the goal will describe an architecture, I would suggest to say
>> that the group will described guidelines to device an ICN architecture
>> based on the common protocol framework to be devised within the ICNRG.
> 
> OK, we have to remember that this is not an IETF WG charter. For a research group, that is really highly contribution-driven, we should allow some flexibility. From that perspective, I think it would be better not promise delivering guidelines to devise an architecture. If course, that would not keep the WG from doing that, if people are interested.

OK

> 
>> Somewhere in this process we should have a milestone to check the
>> applicability of ICN, aiming to check our inline with potential
>> standardization effort is the work to be devised in the ICNRG
> 
> Again, normally "yes" for a tightly controlled activity, but for a research group, I rather see our contribution in advancing ICN and offering the result to the community. I would not like to have a milestone on that.

I was not proposing a very tight activity, but just a reality-check milestone to analyze what is exactly the probability of having ICNRG research work being ported to IETF... I like to see this reality check in any research work, from PhD thesis to EU projects :-)

Cheers

Paulo


> 
> Thanks again for the comments!
> 
> Best regards,
> 
> Dirk
> 
> 
> 
> 
>> 
>> Cheers
>> Paulo
>> 
>> 
>> On May 23, 2011, at 12:27 PM, Börje Ohlman wrote:
>> 
>>> Here is a proposal for a charter for an ICN RG. In producing this
>> draft we have had great help from Volker Hilt, Lixia Zhang, Martin
>> Vigoureux, Joerg Ott, Stephen Farrell and Bengt Ahlgren who has
>> commented and made proposals for improvements. We also got some first
>> positive feedback from the IRTF chair Lars Eggert.
>>> But this does not mean that this charter is ready. There is still
>> room for improvement, e.g. good research issues to be added. Please
>> feel free to contribute to the discussion on this mailing list(s).
>>> At the next IETF meeting we will present the proposal at the IRTF
>> Open Meeting and an RG-to-be side meeting.
>>> The side meeting could either be held in the week before the IETF
>> meeting, with the problem that interested people might not yet be there,
>> or during the IETF week, with the problem that interested people might
>> have other conflicting meetings. Any feedback on this dilemma is very
>> much appreciated.
>>> 
>>> 	Börje & Dirk
>>> 
>>> <Draft-ICN-RG-
>> Charter.txt>_______________________________________________
>>> Icn-interest mailing list
>>> [e-mail]
>>> https://listserv.netlab.nec.de/mailman/listinfo/icn-interest
>> 
>> -------------------------------------------------------------
>> Paulo Mendes, Ph.D
>> Scientific Director for Innovation of the Research Unit in Informatics
>> Systems and Technologies (SITI)
>> Coordinator of the Internet Architecture and Networking Lab (IAN Lab)
>> University Lusofona, Portugal
>> [e-mail]
>> http://siti.ulusofona.pt
>> Tel.: [phone] Fax: [phone]
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> Icn-interest mailing list
>> [e-mail]
>> https://listserv.netlab.nec.de/mailman/listinfo/icn-interest
> _______________________________________________
> Icn-interest mailing list
> [e-mail]
> https://listserv.netlab.nec.de/mailman/listinfo/icn-interest

-------------------------------------------------------------
Paulo Mendes, Ph.D
Scientific Director for Innovation of the Research Unit in Informatics Systems and Technologies (SITI)
Coordinator of the Internet Architecture and Networking Lab (IAN Lab)
University Lusofona, Portugal
[e-mail]
http://siti.ulusofona.pt
Tel.: [phone] Fax: [phone]