Re: Dormant Mode?

James Kempf wrote:

> No, I don't think "no active addresses" would properly characterize 
> it. The host is still prepared to take incoming packets at one of the 
> addresses its got (like for example an incoming phone call).
>
> A deeper mode might be "no active addresses", in which the host either 
> is turned off for the night, or isn't prepared to take any incoming 
> packets, for even more power saving.
>
> I think this behavior spans phones and laptops, the difference between 
> the two is what happens to packets that come while the mobile host is 
> in dormant mode. With laptops (802.11), the AP buffers them then 
> delivers them when the laptop periodically wakes up and polls. With 
> phones, the RNC gets on the paging channel and pages the mobile, then 
> delivers the packets when the mobile wakes up and re-connects.

Or informs the device on a signaling channel, forcing the device
to wake up.

In any case, its useful to avoid extra packets over wireless. NAT
traversal and its keepalives are btw probably a big contributor
here. Some of us have talked about adaptive keepalive behaviours,
but I'm not sure yet if that is technically possible.

--Jari