[Ndn-interest] HopLimit=1 vs Interest aggregation

Teng Liang philoliang at email.arizona.edu
Mon May 7 09:49:01 PDT 2018



On 5/7/18 8:00 AM, Junxiao Shi wrote:
> Dear folks
>
> I have a question on how to process an Interest with HopLimit=1 when 
> the Interest matches an existing PIT entry.
>
> Topology:
>
> A--\
>     R---P
> B--/
>
> A and B are two downstream nodes. P and Q are two upstream nodes. All 
> faces are non-local.
>
> Event sequence:
>
>  1. R receives an Interest with HopLimit=32 from A, and forwards it to P.
>  2. P is preparing a reply and not responding yet.
>  3. R receives an Interest with HopLimit=1 from B.
>
> Per v0.3 protocol:
>
>     if the HopLimit value is larger than or equal to 1, a node should
>     accept the packet and decrease the encoded value by 1.
>     If the HopLimit value becomes 0, a node can satisfy this Interest
>     locally (cache or applications bound to local faces), but must not
>     forward the Interests to any non-local faces.
>
>
> R is prohibited from forwarding B's Interest to P. My question is: 
> Does R add B to the PIT entry as a downstream? Or should R respond 
> Nack-NoRoute to B?
 From my point of view, the HopLimit is just a mechanism to prevent an 
Interest from being forwarded further than the expected. Therefore, 
except the forwarding part, I don't see why a forwarder by default 
should treat an Interest with HopLimit = 0 differently, so in your case 
R will add B to the PIT entry as a downstream. You can always have 
customized forwarding strategies for different behaviors.

Teng
>
> My opinion is:
> If B's Interest would arrive after P responds, it would have matched 
> the cached Data. Thus, R can add B to the PIT entry as a downstream, 
> and send the Data to B when it arrives.
> However, B's Interest should not be considered as a consumer 
> retransmission, so the strategy should not attempt to resend A's 
> Interest to any non-local face.
>
> Yours, Junxiao
>
>
> _______________________________________________
> Ndn-interest mailing list
> Ndn-interest at lists.cs.ucla.edu
> http://www.lists.cs.ucla.edu/mailman/listinfo/ndn-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndn-interest/attachments/20180507/e6a40098/attachment.html>


More information about the Ndn-interest mailing list