Given the amount of debris, and recent trends in launch costs, launch rates, space project completion times and budgets... I'd say decades sounds entirely reasonable. If we had a more serious launch infrastructure in place, it might be done in a few years. Right now, we might get a useful pilot result in a few years if we started a high-priority project today. It would be more years before the project was operating at a useful rate, and yet more to finish cleanup.

Part of the problem is that very small objects moving at high velocity are problematic, and there simply is no easy way to catch them without creating more such objects. At present, many serious proposals actually involve handling every piece of debris individually. That's expensive, to say the least. Other proposals involve complicated stuff with deploying large objects like thin films, threads, etc., which is an activity with a very poor success rate thus far.