[net2-wg] TEP 119 : MAY vs SHOULD

Rodrigo Fonseca rfonseca at cs.berkeley.edu
Tue Aug 19 20:49:58 PDT 2008


But to clarify, the TEP says this:

"If CollectionC receives a data packet that a different node
is supposed to forward, it MAY signal Snoop.receive."

The way this is stated it must be that the node received a packet that
is not destined to it at the link layer, because there is no other way
for a node to know that it should not forward a packet, but it is
confusing.

how about:

--8<-------
If CollectionC receives a data packet that it is not going to forward,
it MAY signal Snoop.receive. This can happen when the network layer
delivers packets that were destined at the link layer to another node.
A received packet can signal at most one of the Snoop.receive and
Intercept.receive events.
--8<-------

Is there any other case in which Snoop would be signaled?
Retransmitted packets? Loop packets?

Rodrigo


On Tue, Aug 19, 2008 at 8:33 PM, Philip Levis <pal at cs.stanford.edu> wrote:
>
> On Aug 19, 2008, at 8:04 PM, Omprakash Gnawali wrote:
>
>> On Tue, Aug 19, 2008 at 5:16 PM, Philip Levis <pal at cs.stanford.edu> wrote:
>> ...
>>>
>>> Or I dunno -- should we just forget this whole thing and say that it MUST
>>> signal Receive? It seems like we're designing for flexibility which
>>> hasn't
>>> yet really been important at all.
>>>
>>
>> Lets do that. When the flexibility becomes necessary, we will revise
>> the specification.
>>
>> One last discussion is Snoop vs Intercept+Snoop before sending the
>> response to the shepherd. I sent an email about this on a different
>> thread.
>
> Snoop is for overhearing packets you're not going to forward. Intercept is
> for packets you're going to forward. If you're computing a MAX, for example,
> then snoop lets you suppress if you hear something higher, while intercept
> actually lets you suppress.
>
> Phil
>


More information about the net2-wg mailing list