I would expect to see the complete elimination of the lines that currently exist between communications types (static text, IM, voice, scheduling, etc...) into single dynamic communications platform. The beginnings of this are already starting to happen but until standardization between the way these various items talk to each other progress a true merger will be a ways off.
Server virtualization will be mainstream in 2016 and Exchange will be taking full advantage of that, running nearly as an OS in it's own right in single session virtual session. Cluster awareness for Exchange will be no longer needed or neccessary in the sense that this will be handled by the underlying OS layer, which will handle the start/stoppage of virtual servers that hang or need to be taken offline.
Mail/IM/Voice/Scheduling stores will be both dynamic and replicated; with access to terabytes of storage becoming common-place an Exchange administrator will be able to designate multiple copies of database items replicated over multiple storage volumes over geographically disperse locations. Load balancing will be acheived by best available/least cost methods, where the response to a data request will be handled by the system with the lowest load and latency times for that request. Should a virtual Exchange server/data store drop offline, the other desginated servers that have replicates of the date will automatically step in to fill the request.
The client (Outlook) will continue to become thinner, very possibly becoming a framework that will load the neccessary code to the device in question (PDA/Phone/Laptop/Brain Implant?) to display the information that the device is capable of displaying, offering a more rich environment on more capable systems and a simpler interface on system of lean requirements. Code updates will be loaded centrally to the Exchange environments and then pushed out to clients.
Hows that for start?
- MEK