Re: [rtcweb] realiable data service

[Tim Panton <[e-mail]>]

On 18 Jul 2011, at 14:56, Magnus Westerlund wrote:

> On 2011-07-15 19:29, Tim Panton wrote:
>> I don't think it is very hard to implement, we could take RFC 5456's
>> sequencing and retry mechanism for example. or indeed Plan9's IL .
> 
> I would warn about thinking this is easy. It is easy to get something
> that works on sunny days, it is hard to build something that is secure,
> efficient, robust and fair.
> 

Agreed, getting all 4 in any protocol is hard, indeed I don't think I've seen it 
done yet.

> I would like to point out that a reliable protocol do need a number of
> functions. If those are achieve from a lower layer or need to be
> implemented in this protocol is a choice. However, functionalities such as:
> 
> - Congestion Control
> - Flow Control
> - Retransmission
> - RTT estimation
> - Buffer handling
> - Parameter and functionality negotiation

Agreed, but only Retransmission is specific to reliable datagrams (as opposed to
unreliable ones). 

> 
> Are likely all need for a safe and stable and future safe protocol.
> 
> There are good reasons for choosing something there is experience in
> that it works.
> 
> For example I took a quick look at RFC5456 which appears to not have
> congestion control. I think it is partly relying on that the transported
> flow will adopt them self if trunking, or if not, that they become
> useless and the call terminates because of that.

RFC5456 has 2 sorts of packet
 - miniframes, which typically form the bulk of the traffic are handled unreliably. 
 - fullframes which are mostly used for signalling  but some of which are reliable datagrams (e.g. the TEXT frame type).
I was saying we could (in theory) ignore the miniframes and adopt the fullframe sequence,ack,retry mechanisms for
the reliable datagram transport.
The core protocol has been 'repurposed' in this way before : http://www.dundi.com/dundi.txt so it is definitely possible.

I take your point about the lack of congestion control, but the ACK mechanism could be used to detect incipient congestion.


> 
> cheers
> 
> Magnus Westerlund
> 
> 


Tim.