[Ndn-interest] [EXT] NDN name lookup procedure
shijunxiao at email.arizona.edu
Sun May 2 18:59:33 PDT 2021
Greetings. After reading the NDN concept and documents, still, I have some
> basic question about NDN naming. When a consumer searches/requests a
> word/data like "Baltik-Sea", with assumption that it is not in CS/PIT
Consumer cannot request a *word*. They must send Interest with a *name*.
It is certainly possible to implement a search service on an NDN network.
In that case, a client application could send an Interest such as
/yahoo/search/Baltik-Sea that is to be processed by the specified search
1- How the name hierarchy is constructed?
Top levels of the name hierarchy are assigned in a DNS-like fashion.
Middle levels are specified by a content publisher.
Bottom levels typically follow naming conventions.
For example, one of the video packets behind https://pushups.ndn.today is
The first component is assigned in a DNS-like fashion, in this case it's my
The second and third component are defined by me, to identify a video.
The fourth and fifth components are determined by the video packager.
The last two components are version number and segment number, following
NDN naming conventions.
2- How is the FIB entry selected to build further hierarchy (eg:
> /ucla/Black-sea, /ua/Black-sea, ... ?)
FIB entries are selected by longest prefix match, provided that the
consumer should send Interests with routable names.
3- What does "Prefix" in FIB mean? is it like IP prefix in IP architecture?
> If so, how many prefix/names are saved in FIB? (Millions/Billions/ ...?)
The FIB is expected to have as many names as active websites in the world.
As you see from the example, I would announce /yoursunny via routing
protocol of global NDN network, and every other website owner would be
doing so too.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Ndn-interest