[ndnSIM] Forwarding pipelines in ndnSIM 1

[Junxiao Shi]

Hi Sabet

ndnSIM v1 does not have forwarding pipelines. Most forwarding actions are
performed by the "strategy".
In contrast, ccnd 0.7.2 has very limited "strategy", and all other
forwarding actions are performed by the forwarding procedures.

When I designed NFD forwarding plane, one design choice is: *what is the
boundary of the strategy?*
ndnSIM v1's model is extremely flexible (which coincides with one of NFD's
design goals), but write a strategy for ndnSIM v1 without copy-and-pasting
code would be extremely complex. ccnd 0.7.2's model is very limited because
strategy cannot define when it's triggered and cannot store additional
information on table entries. I settled with letting the strategy handle
Interest forwarding, and observe (but not change) Data forwarding.
Recent requests from congestion control and wireless networking are making
me rethinking this design choice. In particular, strategy may be allowed to
control Data forwarding in the future.

Composable forwarding / composable strategy
<https://redmine.named-data.net/issues/2000> is an alternate design idea in
which the entire forwarding plane (with a similar or larger boundary than
ndnSIM v1's "strategy") is composed of many building blocks connected
together. This is inspired by Click modular router.
I didn't work further on this idea because I feel the configuration
file would become as complex as ndnSIM v1's strategy. Both compiled code
and configuration file are code, so I want to avoid going too far on that
direction.

Yours, Junxiao

On Sat, Mar 18, 2017 at 2:48 AM, Muhammad Hosain Abdollahi Sabet <
mhasabet at gmail.com> wrote:

> ​Hi there,
>
> It may seem an obsolete question, but I'm looking for the way forwarders
> acted in ndnSIM v1. Since at that time, there were no nfd, how ndn nodes
> behaved at time of receiving interests or data. I guess I'm looking for the
> equivalent for forwarder.h/cpp in ndnSIM v1. Could you refer me to the
> code? As far as I understood it seems in v1 forwarding strategies were
> responsible. Is that right?
>
> Thanks,
> Sabet
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20170319/36e0249c/attachment.html>