Not sure I fully understand the issue, but I believe that convergence relates to equipment swaps.  Nokia and Alcatel both had Wireless Network equipment, when the companies merged, they chose between the two sets of equipment, i.e. one was chosen and the other one probably was soon to be "end-of-lifed".  So it is my guess (and that is all it is, a guess) that some customers may have required Nokia to swap their equipment towards the one that was chosen to continue being supported. A costly problem, but one that must be carried out to preserve customers.

I can also understand why Suri states that the workload is heavy and costly.  I am not sure if he has given guidance on when this equipment swaps costs will come to an end. 

The guidance for Network's Equipment swap outs is $1.4B going into 2019.  So this item will weigh on liquid cash

lt cap,

Thanks for all the details posted.  I'm uncertain of my amateur analysis. But the question of to what extent one currently stays in NOK or not appears to me as -

Even if there are no visible drivers of growth until 2019 at the earliest, one might stay in the stock for the dividend, which would make it worth the wait.

$1.4 b for the necessary support of legacy-like equipment has negative downstream consequences, and impedes free cash, as you cite. Might that not be a reason

to cut the dividend, at least until the expected pick-up in business in 2019-20?

Yes the $1.4B price tag for equipment swaps is a steep amount, gives me pause and I retract my assesment that the dividend is safe.  I am thinking of biting the bullet and reducing my position and waiting to put back money into Nokia at a more opportune time.  Hate to do it with these steep loses, but it may be best.