Capwap Mailing List: Re: Issue 156: Need more details on use of DNS for discovery

[Pat Calhoun (pacalhou) (pcalhoun]

OK, in which case we need to include text in 3.1 that says that if an AC
uses the non-default control port, then the data port MUST be control+1.
Specifically, I would propose keeing your original text, and adding the
following text to section 3.1:

<new text>
3.1.  UDP Transport
[...]
   CAPWAP protocol data packets sent from the WTP to the AC use the
   CAPWAP data channel, as defined in Section 1.4.  The CAPWAP data port
   at the AC is the well known UDP port 5247.  If an AC permits the
   administrator to change the CAPWAP control port, the CAPWAP data port
   MUST be the next consecutive port number.  The CAPWAP data port at
   the WTP can be any port selected by the WTP.
</new text>

OK?

PatC 

-----Original Message-----
From: Pasi.Eronen [at] nokia.com [mailto:Pasi.Eronen [at] nokia.com] 
Sent: Thursday, August 07, 2008 9:48 AM
To: Pat Calhoun (pacalhou); margaret [at] thingmagic.com
Cc: capwap [at] frascone.com
Subject: RE: [Capwap] Issue 156: Need more details on use of DNS for
discovery

Pat Calhoun wrote:
> 
> The IANA service name has already been taken care of, and is already 
> in the IANA Considerations section in -12 (as requested during WG LC).

> I do, however, have an issue with the last sentence in Pasi's proposed

> text, as it states that data channel is control+1. However, in the 
> previous sentence he clearly states that control MAY NOT be the 
> default (assigned) port. So I would prefer to change the last sentence

> to:
> 
> "Note that the SRV record MAY specify a non-default port number for 
> the control channel; the default port number for the data channel is 
> the next port number (control channel port + 1). However, in the case 
> where the control channel uses a non-default port number, the WTP MAY 
> perform a SRV lookup with Service name "capwap-data" and Proto "udp".

> Thus, the name resolved in DNS would be e.g., 
> "_capwap-data._udp.example.com"."
> 
> Comments?

How would the WTP know whether the data channel is control+1, or found
by another SRV lookup?

Something like "If the control channel port in the SRV record is 5246,
data port is 5247; otherwise, the WTP MUST perform another SRV lookup"
would work, but IMHO it sounds like unnecessary complexity -- is there
any reason why the AC can't just pick two consecutive ports?

Best regards,
Pasi