ICNRG Meeting at IETF-87, Berlin, Sunday, July 28





  1. ICN Challenges
  2. ICN Management
  3. Scenarios
  4. Other Topics (have to see how we can address them)
  5. Open Discussion (15:30)
  1. Notes:

-- Dirk presents agenda for the day

-- Dirk gives an overview oif the research challenges draft

-- question (Christian): what is the intended overlap with the survey draft; answer Dirk: intention of document is not a survery, and yes, there needs to be some alignment; Börje: survey document will be discussed at the Honkong meeting as key people are not here today

-- comment (Jan): need to explain a bit mroe why TLS cannot be used to establish trust with endpoints

-----Börje: What I think we want to  say is that if you have the assumption that you can't trust the  endpoints/sources from where you retrieve objects solutions like TLS  does no make sense

-- question (Christian): we need to write something about name lifecylcles

(Elwyn: maybe covered by challenge bullet 4 on p8.. updating and versioning of names)

-- comment (Nacho): how can we enumerate all possible research challenges, we do not know all of them yet, how do we scope this?; answer (Dirk): true, it may be possible that we cut out challenges later on; question (Nacho): so when is the cut off date for deciding what challenges we include?

-- question (Jan): what about access control for NDOs as a security challenge?

-- short discussion on mobility management challenges; potentially, ICN has less challenges here than IP; Kostas: however, for certain real-time applications, anchor-based solutions perform better than with no network support for mobility, so also for ICN mobility management solutions probably need to be studied; Jan: agreed, but only for certain types of applictaions with e.g. tight delay constraints

-- comment: what about writing more about adhoc scenario benefits and research challenges?

- comment (Christian): WIreless Networking could be associated to the mobility challenges section; Kostas: it could be so, pending contributions

- comment (Christian): this could be broader than wireless; Kostas: but no one doesn't on cables anymore; Nacho: Satelites (and uni-directional links)!

- comment (Kostas): there are two topics on this: the centralized repository Dirk focused on, and also the wireless aspects below; Nacho: and they all relate to mobility; Kostas: but I think we should separate wireless from mobility, because we are addressing media here.

- comment (Spiros): challenges vs opportunities vs technical work. Should we try to explicitly make this distinction? (popular challenges vs newly discovered ones); Kostas: but we need contributions; Jörg: this should not be a newcomer document; Jan: true, but sometimes we need more text to explain the challenges in areas where not everybody is working on; Dirk: objective is to document ackknowldeged challenges in the ICN area Kostas: not everyone is an expert at the cutting edge of all topics discussed here. So experts in one area could get familiar with issues in other areas too (and detect new challenges).  Broadcast (besides wireless) is not really common (my textbook defintion at least); yes, it may be that it is defined in copper/other but not really in use at scale.

(?) Three topics covered now in wireless section: wireless aspects (next revision will discuss more than broadcast); serverless/non-centralized repository; and (?). Also this section is the only one that covers lower layers (media). Do we need to add text for optical/other, for example?

-comment (Dirk): (caching) Object sizes need to be mentioned here for caching challenges

-comment (Dirk): (management) for the managent part we need to consider how we split this between the Challenges draft and the on-going management draft; Daniel - the split could be in identifying what is challenging in the Challenges draft, and showcasing how which mechanisms ICN provides for management in the management draft

-comment (ElwynE): do we want to provide video distribution? Dirk: Isn't it a transport issue? Börje: It's a good topic.Dirk: this is important enough for a document on its own, but we could see how to best fit a section in the challenges doc. Mayutan: Application-driven document can be a possibility.

Topics to be discussed later: Kostas: whether and why to limit the size of the challenges document. What belongs to which document. 

------------------- Daniel's presentation -----------------------

comment: Is this document a mobile centric document. Daniel: We envision that it should be  general, but we are providing mobile centric examples. kostas: it should be a more than mobile management document and should go in direction of general ICN centric management.  

comment: Other management issues: Cache management

dirk: Taking this as an example, we need to learn about what needs to be done for CCN mamangement. 

Kostas: is there interest in this document and moving it forward? Daniel: The document is based on the contributions we get. Currently, we have a mobility-oriented management contributions. We are happy to add more contributions from other aspects of mobility, so all feel free to contribute. In fact, we can even go back and revise what was already done, to assertain (in the end) a list/description of common considerations of different management aspects.

-comment: The naming does not reflect the current content of the draft, it should have other management issues concerning federated caches, naming, .....

Kostas: The draft contains more than what the presentation touched upon

------------------- Scenarios -----------------------

------------------- ICN Baseline Scenarios -----------------------


Kostas presenting

*********** Elwyn presenting on ICDTN ************

*********** Daniel presenting on IoT in ICN ************

*********** ??? presenting on Smart City ************

*********** Spiros presenting on Relevant Metrics ************

------------------- Discussion Items -----------------------

------------------- CCNlite -----------------------

Christian presenting

fragmentation potentially interesting as a general ICNRG topic

------------------- named-function.net -----------------------

The presentation proposed that ICN objects could also be active objects such as functions or programs. The ICN could then be extended to match data and function objects and evaluate the result. It was agreed this is an interesting research topic to investigate further even though some people expresses some sceptsm based on experiences from active networking research.