[Ndn-interest] How is scalability done globally given current name structure

AnChe Kuo schwannden at gmail.com
Thu Jan 15 18:52:41 PST 2015


Dear Sir or Madam

I am an intern at Academia Sinica and a student at NCTU, Taiwan,
and I have been studying NDN for a few month now.

If I wrote an application and call it "Foo", then can I simply register
"Foo" into the global NDN network? Given current architecture, each router
administrator is allowed to register only names satisfying certain pattern
that is assigned to them, right? So if Academia Sinica now joins NDN
network, we might get a prefix "ndn/tw/sinica", which means I can only
register my application as "ndn/tw/sinica/Foo".

But this present a problem. If some one in UCLA decided that she loves this
application, and also runs a producer there, she would have to register it
as "/ndn/edu/ucla/Foo". Now how can a client in China get the content that
is closest to him if he only knows the application name "Foo"? If he sends
an interest named "Foo", then longest prefix match won't find anything. If
he has to know the full name "/ndn/edu/ucla/Foo", then we have lost the benefit
of NDN architecture, because it is almost equivalent to an exact source,
not just a content name.

So how does NDN intend to solve the problem?
1. If NDN only focus on a small network, with applications for IoT or M2M
can register their names freely, then what are the benefits of NDN other
then its more natural architecture for related application? If we look at
efficiency, then TCP/IP is probably enough for a small area network, will
NDN offer the improvement that is necessary?
2. If we release the top level namespace for all applications to register,
so that we may register "Foo" into global routers, with these top level
namespace managed by some organization (similar to ICANN), then will we
suffer from routing table explosion? Because now we have each unique URLs,
and millions of different applications registered at the top level to serve
the global users.

So far these are my concerns. I understand I might be having the wrong
question, because it might no be the intended use of NDN or the
current objective of the whole NDN project. But either way I would love to
hear from your feedback, or any critique on my question is welcomed.


-- 
Best Wishes,
郭安哲
Kuo, AnChe(Schwannden)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndn-interest/attachments/20150116/2fc2a7ba/attachment.html>


More information about the Ndn-interest mailing list