Stroke length dictates mechanical limits on RPM for a couple of reasons:
-
long stroke + high RPM = high piston speeds, which generates a lot of heat where the piston and its rings slide on the bore. Pick any car, check the redline on the tachometer, and divide by the engine’s stroke length; you’ll come up with a number comparable to most other cars. High-performance sports cars push this limit a bit, but not by much.
-
long stroke + high RPM = high piston accelerations. at TDC, as the piston stops and begins moving downward, the piston ring can lift up off of the land if combustion chamber pressure isn’t enough to hold it down. The result is more hot combustion gases blowing by the piston ring, which shortens the life of parts.
Check any high-revving engine, and you’ll find a short stroke length. Diesels, which typically aren’t very high-revving (because more time is required for diesel combustion) tend to be designed with longer strokes. Extremely large marine diesel engines, with stroke lengths measured in feet, usually don’t rev to more than a few hundred RPM. OTOH, the Yamaha R6 motorcycle, which has a stroke of just 42.5 mm, redlines at somewhere between 16,000 and 17,000 RPM.
If you’ve got short stroke length, and you want any kind of displacement, you’re going to need large bores, or a lot of pistons, or both. This is why Formula One race cars have V-8 (and until just a few years ago, V-10) engines, despite having displacements of 2.4 - 3.0 liters; compare that with my Nissan Maxima, which spreads 3.5 liters of displacement over just six cylinders. The downside is that my Maxima will not rev to 19,000 RPM. 
The power curve for an engine depends almost entirely on how it breathes. Exhaust system characteristics matter for this, but not nearly as much as intake system characteristics. Most performance-oriented gasoline engines these days use four valves per cylinder - two for exhaust, two for intake. That provides the opportunity to maximize airflow, even at elevated RPM. To take advantage of that opportunity, designers use the inertia of the air in the intake system, and the fact that air is compressible. Want an engine to make maximum torque at some arbitrary RPM? Simple: set up the intake system so that at that RPM, it stuffs as much air into the cylinder as possible before the intake valve closes. It’s a bit like tuning the pipes in a pipe organ.
Suppose you want an engine to produce lots of torque at low RPM. you make the air duct leading from the airbox to that cylinder skinny and long, and you set the intake valve timing so that it closes relatively close to BDC. At mid-stroke, the air is flowing very fast through the duct; it’s got lots of inertia. When the piston reaches BDC, that momentum continues cramming air into the cylinder. If you’ve designed things correctly, that air flow comes to a stop just as the intake valve closes and traps the maximum amount possible in the cylinder. Presto, your engine is making peak torque at this RPM. What happens if the RPM increases away from this ideal? Now the piston passes through mid-stroke too quickly, so the air in the duct doesn’t have a chance to accelerate to the desired maximum speed; moreover, the intake valves close too soon, before the air has come to a dead stop, so you haven’t crammed as much air into the cylinder as you were at lower RPM. Presto, the engine is now making less-than-peak torque.
Want an engine to make peak torque at higher RPM? shorten the intake tract a bit so those pressure spikes arrive at the intake valve sooner, and leave the intake valve open a little longer, even after BDC. Seems paradoxical, but even after BDC when you would expect the piston to be pumping air backward into the intake system through that open valve, the momentum of air in the intake system is still shoving even more air into the cylinder.
Want an engine to make peak torque over a wider range of RPM? Use variable valve timing. In the ideal, at higher RPM you would open the intake valve earlier (before TDC) and close it later (after BDC). A lot of production passenger cars simply shift the phasing of the camshaft so that the intake valve opens and closes later; this is less than ideal, but it’s economical. If you’ve got more money, you can also change the geometry of the intake system. The cheap way is to keep one of the two intake ports closed off at low RPM, and open it at higher RPM; there are a number of production passenger cars that do this now. The expensive way is to actually make those intake ducts adjustable in length: some late-model high-end sportscars (Ferrari, etc.) utilize computer-controlled intake ducts that extend or retract into an airbox according to RPM so that they provide very optimum breathing over a huge RPM range.
Different vehicles have different needs, and the engine’s characteristics are designed into them so as to best meet those needs. An 80,000-pound tractor-trailer needs an engine that can deliver lots of torque at very low RPM; without that, the driver will fry the clutch just trying to get started from a dead stop. They don’t need a shitload of power, though; most big-rigs have a few hundred peak horsepower, not much considering their weight. In other words, they don’t make much torque at high RPM. These are the characteristics you’re likely to find on any vehicle designed for towing applications.
Drag-racing cars have similar needs: they want a lot of torque to launch quickly off of the line.
Track racing cars don’t need low-end torque for zero-speed launches; instead, they need lots of power to accelerate quickly and maintain high speeds even in the presence of heavy aerodynamic drag. So they deliver their peak torque at high RPM, or the spend money on a variable-geometry intake system as described above.
Run-of-the-mill production passenger vehicles tend to be tuned with a fairly flat torque curve - not extremely peaky at either high or low RPM. They’re designed to be easy to drive, rather than being high-performance.
