Re: [dtn-security] Newbie seeking some security related advice

Minor comment to Armando's email:

> -----Original Message-----
> From: dtn-security-bounces@maillists.intel-research.net [mailto:dtn-
> security-bounces@maillists.intel-research.net] On Behalf Of Armando
> Caro
> Sent: Wednesday, May 27, 2009 8:14 PM
> To: Ivancic, William D. (GRC-RHN0)
> Cc: dtn-security@maillists.intel-research.net
> Subject: Re: [dtn-security] Newbie seeking some security related advice
> 
> Ivancic, William D. (GRC-RHN0) wrote:
> > I believe DARPA is using DTN for voice networks but I believe the
> disruption is expected to be very quick fades and the like or perhaps
> that voice simply will not make sense at certain times.
> 
> I believe you are referring to the work that BBN is doing for the DARPA
> WNaN program. For this program, we run IP over DTN and use VoIP without
> modifications to the VoIP application. We are using the BPA as a
> forwarding engine that works in both connected and disconnected
> scenarios. VoIP traffic (and other delay sensitive traffic) are only
> routed/forwarded to destinations that are currently connected, since
> VoIP doesn't make sense for disconnected destinations. If a node
> receives VoIP traffic for a disconnected destination, the traffic is
> dropped.

As Armando describes, we are currently using streaming voice over BP.
Individual packets contain 20-100ms of voice, so sending those individual
buffers in an individual disrupted fashion would not make a whole lot of
sense.  Note that we are also working on providing voice-snippets of 1-10s
that could be sent in a disrupted fashion over the network.  But ultimately
that's just data like any other bundle.  The only real "work" there is in
the application, not the BP.   

The key point is that voice can either be streaming as little snippets or
buffered as a big blob.  If it is a big blob, it's just another bundle.  If
it is streaming, then one needs to make sure you have a BPA/router that adds
only a minimal amount of additional overhead to the forwarding process.  (as
Armando says below)

> One important take-away from this work... the motivation of DTN is to
> support delay/disrupted scenarios and apps that can operate in that
> regime, but it is possible to implement a DTN stack that can also deal
> with time-sensitive traffic.
> 
> Armando
> _______________________________________________
> dtn-security mailing list
> dtn-security@maillists.intel-research.net
> http://maillists.intel-research.net/mailman/listinfo/dtn-security