If you do not specify an explicit bandwidth for a link, you get a "native
speed" link, which for the d430s could be either 1Gb or 10Gb since these
machines have both 1Gb and 10Gb NICs. Which you get depends on the specific
nodes available at the time and the heuristics of the mapping algorithm.
You can specify a bandwidth of one or the other to ensure you get a native
link at the desired speed.
"Traffic shaping" is a different mechanism that was intended for BWs that
don't match the native speed (or if you need to induce delay or packet loss).
However, for bandwidth, it only works well for BWs of less than 100Mbs.
On Sun, Aug 29, 2021 at 01:48:41AM +0300, Andria Trigeorgi wrote:
> > On 29 Aug 2021, at 12:54 AM, Eric Eide <ee...@cs.utah.edu
> > ???Andria Trigeorgi <an.tri...@gmail.com
> >> I am referring to the default traffic shaping parameters between nodes.
> > I think you might have a minunderstanding about how Emulab works.
> > When you request resources from Emulab --- in other words, when you
> > "instantiate a profile" --- Emulab gives you isolated, *physical* compute nodes
> > and network links. Emulab does not give you virtual nodes and links unless you
> > specifically ask for virtual resources.
> Thank you for your response!! Yes I have requested a Lan with RawPC nodes of d400 type. I have seen that I can specify the traffic shaping parameters for each of the node in the Lan (e.g. by setting value to request.LAN(???lan???).bandwidth etc). So I am wondering about the default value for bandwidth, latency, packet loss.
> > This is why Mike Hibler pointed you to the documentation on the hardware.
> > http://docs.emulab.net/hardware.html
> > The different kinds of hardware nods have different kinds and speeds of network
> > interfaces. They vary from 100Mbps to 10Gpbs.
> > Emulab doesn't do traffic shaping on the network links that it allocates to you
> > for your experiment. Again, this is something you can *ask* for, but by
> > default, Emulab does not impose "traffic shaping parameters between nodes."
> > I hope that this is helpful!
> > Eric.
> > --
> > -------------------------------------------------------------------------------
> > Eric Eide <ee...@cs.utah.edu
> . University of Utah School of Computing
> > https://www.cs.utah.edu/~eeide/
. +1 801-585-5512
. Salt Lake City, Utah, USA
> > --
> > You received this message because you are subscribed to the Google Groups "emulab-users" group.
> > To unsubscribe from this group and stop receiving emails from it, send an email to emulab-users...@googlegroups.com
> > To view this discussion on the web visit https://groups.google.com/d/msgid/emulab-users/m1y28lfd71.fsf%40cs.utah.edu
> You received this message because you are subscribed to the Google Groups "emulab-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to emulab-users...@googlegroups.com
> To view this discussion on the web visit https://groups.google.com/d/msgid/emulab-users/7CDD48A4-1069-4862-8E0C-318E80FB0608%40gmail.com