Does it make sense to separate the Publisher (QDS) from the Server (QVS) from one physical server to two? I am aware of clustering but am looking for a "license neutral" incremental ability to scale up and/or increase performance. Specifically, if QVS remains on machine one and QDS is separated onto machine two, which machine is doing the heavy lifting when a reload task is run? It is easy and would be ideal to assume that the answer is QDS, but if some or all of the RAM and CPU cycles are actually consumed by the QVS on server one for a reload task, then it would not make sense to split them from a performance perspective. Does anyone have some specific insight into this?