Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: [802.21] [DNA] Prefix information for link identification in DNA



Qiaobing,

If you imagine the case of a server on a network, it will normally have all its layer 3 parameters configured.  So when layer 2 comes up, layer 3 will come up at the same time (ignoring processing delay).

I think the PDP context case is best thought of in the same way - it establishes the layer 3 configuration, it then brings layer 2 up, and then as the layer 3 configuration is already complete, layer 3 comes up immediately.

From "outside the black box" it's not possible to distinguish the difference between L2 and L3 coming up in this case.

In the general case, when an Ln link comes up, the L(n+1) protocol elements will first establish communication with each other, and then begin to pass protocol data units as a service to L(n+2).  However, that communication establishement stage can be null in some protocol configurations, such as this one.

Mike.

> -----Original Message-----
> From: Qiaobing Xie [mailto:Qiaobing.Xie@MOTOROLA.COM] 
> Sent: Friday, October 07, 2005 4:57 PM
> To: STDS-802-21@listserv.ieee.org
> Subject: Re: [802.21] [DNA] Prefix information for link 
> identification in DNA
> 
> 
> Stefano,
> 
> Sorry for late reply. Still catching up my email after a vacation.
> 
> Stefano M. Faccin wrote:
> 
> ...
> > Link up (link_X) indicates that link_X is available to 
> carry data. Since the MIH users of this event are at IP layer 
> and above, they would not have use of an indication stating 
> that the UE has setup a L2 connection with a BSS or SGSN 
> (since that is only signalling, not user plane). Link up in 
> this case means that the UE has a usable PDP context with a 
> GGSN and therefore has been assigned and IP address, so that 
> the link can be used to carry data. 
> ...
> 
> So, your "link_up" here means the IP connectivity is up and ready. In
> other words, your "link" here refers to the IP link used in 
> DNA, instead
> of a L2 link. Do you think in general 802.21 link events should be
> defined for IP link?
> 
> regards,
> -Qiaobing
>