Re: When route is changing....?

[Polly Huang <huang@catarina.usc.edu>]

Sun,

Simulation is essentially an abstraction of the real world.  It is 
difficult, if not impossible, to duplicate what exactly happen in
the real world.  Usually, depending on what levels of details you care,
you make sure the modules you're using capture the essence.  

Neither session routing nor DV work exactly like the real world in a sense
that session routing ignores the dynamics that could happen in the
transition of route changes and DV ignores processing time (and some other
timer details furhter specified in PIM DM or DVMRP docs).  Nevertheless, 
session routing and DV are already of great help to a significant set of
research problems.  

And, the causality problem I thought (not sure if Lloyd agrees) was that
session routing further abstracts away the transition details so that
the timing/ordering of packets might be different from the results of
DV.

hope this helps to clarify things a bit.
-Polly

On Tue, 12 Oct 1999, K Sun wrote:

> Hello,
>     Thanks for your discussion!
> However, I am still not understanding it well.
> 
> 
> > > > Sun,
> > > > 
> > > > That means (I think) if new routes happen to be much better than the old
> > > > ones (e.g. bringing up a direct link in between src and dst), packets
> > > > fired at a later time can arrive before earlier packets, which
> > > > might cause events to be triggered in a differenr order (thus called the
> > > > causality problem).
> > > 
> > > I'm not sure that's it; if routing changes as you describe on e.g. TCP
> > > flows you will get out of order packets and odd effects anyway (I
> > > think Partridge is working on a paper detailing observed traces and
> > > effects on TCP for Transactions on Networking). That is entirely
> > > normal (packet network! feature!), and not a causality variation per
> > > se, since time is still flowing forward. In fact, you'd want to study
> > > it...
> > 
> > understood. but session routing may exaggerate the degree of disordering.
> 
>      So, there is no causality variabtion in such a sense. It exactly 
> accords what happens in the real network. May I have such an 
> understanding? 
>      Actually, not only in session routing but also in DV rouing (or 
> other dynamic routing method), there should be the same effect after 
> route are changing. So, the lost packets should be able to be traced in 
> ns, should not them? I hope a definite answer for above things. 

> 
>  
> > > What I think it is (and I'm sure Kannan will correct me if I'm wrong)
> > > is that events scheduled at or near the same time as the routing
> > > change may not experience the routing in the network that you would
> > > expect _for that instant in time_ and the neat forwards flow of
> > > simulated time is effectively violated. Effects probably vary
> > > depending on the scheduler...
> 
>      It means there is difference between the effect influenced by 
> route changeing in real network is DIFFERENCE from the simulation 
> effect in ns? If so, might we correct it so that the simulation effect 
> can be in accordance with the real effect? Needs accurate control on 
> scheduler?
> 
>      I do not know, in real network, whether it also needs some time to 
> put new routes in effect?
>  
> > We might meant the same thing.  What I was trying to say is that
> > packets may come in an order that's different from what we would expect
> > with a more detailed routing implementation, which usually takes some time
> > to converge on new routes. During the transition, data packet might be
> > lost somewhere (and perhaps retransmitted later) as oppose to that
> > session routing allows new routes to be computed right at the moment
> > and succeeding packets will have no problem (maybe just a little) taking
> > the new routes to the receiving ends. 
> 
> 
>     Anyway, thanks a lot for your discussion and hope further 
> clarification!
> 
>      Best Regards,
>              Sun Kai
> 
> 
>