Consider the algorithmic task of compressing a video file for a cartoon. Initially, the video is a stream of separate images. The images are large, and in many places, differ only slightly from frame to frame. For instance, when an object is moving, the part of the image not involving the object does not change.

Say we want to design a compression mechanism based on determining the difference from the previous frame to the next frame, and saving only the difference into a compressed result. Of the five issues in multicore programming, which is the most problematic for multithreading this system? (Dont worry about audio.)