Comment by HelloNurse

Is there any similar algorithm that trades degraded performance for a guaranteed maximum node size? Using fixed size buffers can be more important than average-case performance.