Arbiter Using Fixed Priority Logic for Fair Resource Access

The arbitration system is segmented into two independent modules: a fixed priority arbiter for fast grant generation and a priority update module for fairness maintenance. This segmentation allows the FPA to operate independently with low latency while the priority updates occur in parallel, achieving both speed and fairness without the complexity of full LRU implementation.

Priority values are pre-calculated and stored in priority storage elements before arbitration occurs. The priority update module proactively updates these stored priority values based on grant outcomes, so that when the next arbitration cycle begins, the correct priorities are already in place. This preliminary action eliminates the need for complex real-time priority calculation during arbitration, reducing latency.

The system separates the arbitration function (performed by the fixed priority arbiter) from the priority management function (performed by the priority update module). This segmentation allows each module to be optimized independently: the FPA uses simple combinational logic with minimal area, while the priority update module uses lightweight sequential logic that only stores and updates priority values, avoiding the large state machines required by traditional LRU arbiters.

Instead of implementing full LRU state tracking which requires storing complete arbitration histories, the system changes the parameter representation to simple priority values (e.g., 0-7 for 8 clients). These priority values are updated incrementally based on grant outcomes, transforming the complex LRU state into compact numerical parameters that require minimal storage area while still enabling fair arbitration.

The system combines the static structure of a fixed priority arbiter with dynamic priority values. While the FPA itself remains a simple static circuit for low latency, the priority inputs to the FPA are dynamically updated by the priority update module based on previous grant outcomes. This dynamics allows low-priority clients to eventually receive grants when higher-priority clients do not continuously contend, preventing starvation while maintaining the speed of the FPA.

The priority update module implements feedback by monitoring grant signals and using them to update priority values. When a client receives a grant, its priority is adjusted (e.g., set to lowest or incremented), and this updated priority feeds back into the arbitration process for the next cycle. This feedback mechanism ensures that continuously winning clients cannot monopolize the resource, providing fairness while the FPA maintains low latency through its simple structure.