I was actually brainstorming a bit on a scenario like this today.
I think one of the most important things is to make the mechanics of the game relevant to the war being fought - i.e. a short and modern one. Assigning New York to build an aircraft carrier and launching it in a dozen turns sounds downright silly. This also applies to research in a way too.
One idea I had was to replace the tech-tree with a sort of barometer that would bring special events into play (because really, how much research and development would go on in the course of a war that lasts mere months?). I picture it almost like the experience bar that shows your progress towards a Great General.
An example could be the use of nuclear weapons.
Various events could contribute points towards nuclear weapons becoming an option. Perhaps every unit you lose will contribute a little bit, losing cities (or having cities bombed even by conventional air units) even more, or having the U.S. mainland attacked by Soviet forces (or vice-versa). I would imagine that the other side using nuclear weapons would automatically fill out your bar. When this baromter gets maxed out, weapons can be deployed.
Even more interesting would be that at some point short of maximum you are given tactical and theatre weapons to use if you wish - but when you reach the tipping point (have lost too much ground in Europe, have lost too many battles, have exchanged weapons in a low-level nuclear confrontation, etc.) the result is an automatic full-scale nuclear exchange and, of course, game over for all involved.
I picture similar measurements for a draft/mobilisation that brings you more units (that goes up as you take losses), war weariness, etc. Think of it as replacing research with a sort of running measure of public opinion/political events and decision making.
Another idea would be fuel and supplies. Like research, I don't picture the player building a lot of units due to the sophistication of the weapons platforms involved and the relative shortness of the war. Instead, cities - and their production/commerce value - will contribe to your country's overall ability to supply and reinforce your units. At the start West German units might be able to repair themselves at, say, +10% the normal rate. As cities are lost, this bonus could also be lost, and instead they will suffer a penalty to repair. A city's productivity could even be represented with a simple percentage, just like defense, which can be knocked down through bombing. In fact, the domestic/economic management of the game could simply be represented as a function of a city's population and the percentage of intact industry/commerce.
Personally, I'd rather see a degree of simplicity when it comes to the number of units, resources and the size of the map. Rather than making the scenario big (a dozen different aircraft per side, a huge map, 30 different resources etc.), make it deep (use promotions to differentiate between units, make each tile and city interesting and of strategic value, etc.).
Anyway, some ideas I had.