A very interesting response to a question posted on another site:
The bearings of choice for the large end of a two stroke connecting rod are caged needle bearings -- generally, the crankshaft rod journal serves as the inner race, so must be surface hardened to RC60, and the connecting rod large end serves as the outer race and must be similarly hardened. One of the weak points of these bearings is the cages -- as the rod moves through 360 degrees of rotation the surface speed of the bearing continuously changes speed, requiring the rollers to accelerate and decelerate.
This would not be much of a problem were it not for the mass of the cage holding the rollers in alignment: the normal copper coated steel cages weigh enough to push against the rollers and skid them across the races at high RPM, scarring the races and flattening the rollers. The highest end large end bearings use silver plated titanium cages which weigh as much as one third of their steel counterparts, and are therefore much less prone to skidding. Interestingly, the most common crank train failure is not the bearing itself, but the thrust washers centering the rod on the crank -- being copper or silver plated steel, they do not at all like high speed rubbing loads with the scanty lubrication in the crankcase, and they scrub and get hot enough to flash the oil, excluding lubrication from the bearing.
The best cure for this, done by Walter Kaaden first for MZ, and then Suzuki, was to eliminate the thrust washers altogether and center the rod with washers on the small end inside the piston, improving RPM potential by as much as 500 mean FPM, or 1,500 RPM for the 50 and 125 multi cylinder engines.
On the small end, the cage part of the needle bearings used is again the weak link, but for a different reason. The bearing load is almost exclusively linear (in line with the cylinder) and the cage doesn't really rotate very much at all, so the decelerative / accelerative forces at TDC and BDC flex the cage, eventually causing fatigue cracks quickly followed by outright failure which sends the cage pieces and rollers showering down into the crankcase, with disastrous consequences to all those expensive parts. The most reliable needle bearings are simply the needles crowded together with no cage at all. This, of course, is a real bear to assemble, but almost absolutely insures the small end bearing is quite reliable