The goal of this paper is to assess the value of simple features that are widely available in off-the-shelf CORBA and Java platforms for the implementation of faulttolerance mechanisms in industry-grade systems. This work builds on knowledge gained at LAAS from previous work on the prototyping of reflective fault tolerant frameworks. We describe how we used the interception and state capture mechanisms that are available in CORBA and Java to implement a simple replication strategy on a small middleware-based system built upon GNU/Linux and JOrbacus. We discuss the benefits and the limits of the resulting system from a practical point of view.