[ndnSIM] (no subject)
alexander.afanasyev at ucla.edu
Sun Apr 28 13:27:27 PDT 2013
I'm not sure that I can follow you completely. It doesn't really matter whether a "temporary" PIT entry is created or not. In any case, with the first interest that doesn't have a match in CS, you will get full hop count (i.e., towards data producer and back). For all other interest, hop count will be limited by the caching node.
On Apr 27, 2013, at 11:25 PM, "xyhu" <xyhu at njnet.edu.cn> wrote:
> Hi Alex,
> No problem. It is just because that I did some modification about the forwarding
> strategy. For example, I want to express some information to the nodes in the delivery
> path about the hop count that the Interests for the content object has experienced.
> From your process, when interest(s) get hits in content store, the returned object may
> for several Interests from different consumers and I can pick the hop count experienced
> by only one of the Interests to the content objects such that the hop count is not the
> actual value for the left interests.
> Hope that I describe it correctly.
>> Hi Xiaoyan,
>> The main reason for this processing is to properly calculate metrics for incoming
>> interest/satisfied pending interests.
>> Is there any problem?
>> On Apr 27, 2013, at 9:54 PM, "xyhu" <xyhu at njnet.edu.cn> wrote:
>>> Hi Alex,
>>> Sorry to bother you again.
>>> I found that in ForwardingStrategy::OnInterest(),
>>> you first check PIT (and get a PIT entry for the Interest anyway) and then CS, which is
>>> different from the description in Van's NDN paper.
>>> But if there is the matching data in CS, it is unnecessary to create a PIT entry for this
>>> Interest, right?
>>> So I am curious about that you may have special meaning for processing incoming
>>> interests in this way. Do you?
>>> ndnSIM mailing list
>>> ndnSIM at lists.cs.ucla.edu
>> XFilter-NENC-Signature: 517cadee00360d69
More information about the ndnSIM