[ros-users] ROS & DDS

[Geoffrey Biggs]

On 18 February 2014 12:14, Brian Gerkey <gerkey at osrfoundation.org> wrote:

> I have to say that I'm a little surprised by your experience, as it
> seems to run contrary to what I know about ROS combined with what I've
> read about DDS.   As we experiment, we'll pay close attention to the
> issues you raised (discovery responsiveness, large message
> transmission, and reliability over wifi).
>

Large message transmission requires altering the maximum message size from
its default. In RTI's implementation, this makes internal buffers larger in
order to handle the potential larger messages without affecting real-time
responsiveness through the need for memory allocations.

Discovery is always going to be tricky in a changing network. ROS does not
have this issue because it does not use a discovery protocol. DDS's
discovery protocol can be tuned to match the environment in which it will
be used. There are DDS users successfully working in an environment where
network changes through loss of signal is a common occurrence. I have
spoken to someone at NASA who worked on the DDS setup used between the
Robonaut 2 and the ground, and he said that they had to convince RTI to
support much longer timeouts to handle their communication latency and loss
of signal events.

Geoff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ros.org/pipermail/ros-users/attachments/20140218/1030059a/attachment.html>