Figure 3-14. An example of false contention. If process P100 wants access to memory structure MS1 and process P200 wants access to memory structure MS2, they will both contend for the same control structure CS999, resulting in false contention.

Another consequence associated with false contention is control structure contention. For example, suppose you want to build a shed to store farm equipment, but you must first obtain a permit. Many other people also want to build sheds, so the governmental office is overwhelmed with permit requests. To ensure there is no illegal builder collaboration, each permit is serially reviewed. As you can imagine, during peak permit submission times, the permit review process can experience long delays. In this situation, acquiring the permit is the bottleneck, not shed construction.

Relating this to Oracle, the bottleneck would be acquiring the control structure, not accessing the underlying memory structure. When complex memory structures are protected by a limited number of latches, solutions focused on the control structure are to either add control structures (for example, specialized governmental permit reviewers) or decrease the control structure access time (for example, streamline the permit review process). In Oracle terms, this means either add more latches and improve latch acquisition performance or optimize the kernel code accessing the memory structure. A creative option is to not use latches at all, but instead use a completely different structure that inherently provides both of these benefits-like a mutex! Reducing the likelihood of control structure contention is another reason mutexes are an attractive memory control option.