Sucking flames and smoke into the cabins on Velocity is usually not a
  problem.  In most cases, the fire just burns until the engine falls off
  (only semi-joking).  This is why many, including myself, have a fire
  detection system on the pusher that signals or alarms up front when the
  fire in back is not contained inside the six cylinders.  You tractor
  guys are lucky, you get a face full of smoke and flames, so you know
  what's going on (only semi-joking).
 
  Different smoke, for different folks.

Chuck,

 

Verrrry Interrrrresting!

 

Ya, we just wait 'till our feet get hot!  

 

The manufacturer of one of the new VLJs has developed a very compact fire supression system.  The cartridge is the size of a 35mm film container and is equivalent to a quite large Halon system.  I can't remember which manufacturer (the source was either AvWeb, Aero News, or AOPA Pilot) but I hope they will make the system available to the experimental aircraft community.

 

All joking aside, the lower cowl area of down draft cooled engines typically seems to be 2-3 inches of H20 pressure higher than ambient (static) and the cockpit runs 2-5 inches of H20 pressure lower than ambient.  The difference is 4-8 inches between the cockpit and the lower cowl area - independent of whether the engine is in the front or the back.  In rear engined canard airplanes, the cockpit exit air travels down the spar in the strake or exits via some sort of vent behind the pilot area.  That does not mean the flames or CO can't mix with cockpit air - maybe it will get the rear seat occupants first.  Again, there is no high performance composite airplane that I know that has a higher pressure in the cockpit than ambient. 

 

Up draft cooled rear-mounted engines typically have fresh cooling air circulating around the firewall - good for reducing the CO problem, but even more dangerous with a fire from, say an oil line, since the pressure between the cooling plenum and the cockpit is even higher!