Rate Limiting in Verilog

Rate limiting is an important mechanism for controlling resource utilization and maintaining quality of service. Verilog can simulate rate limiting using system tasks and behavioral modeling.

module rate_limiter;
  reg [31:0] requests[$];
  reg [31:0] bursty_requests[$];
  integer i;
  
  initial begin
    // First we'll look at basic rate limiting. Suppose
    // we want to limit our handling of incoming requests.
    // We'll serve these requests from a queue.
    for (i = 1; i <= 5; i = i + 1) begin
      requests.push_back(i);
    end
    
    // This timer will trigger every 200 milliseconds.
    // This is the regulator in our rate limiting scheme.
    forever begin
      #200;
      if (requests.size() > 0) begin
        $display("request %0d %0t", requests.pop_front(), $time);
      end
    end
  end
  
  // We may want to allow short bursts of requests in
  // our rate limiting scheme while preserving the
  // overall rate limit. We can accomplish this by
  // using a separate process for bursty requests.
  initial begin
    // Simulate 5 more incoming requests. The first
    // 3 of these will benefit from the burst capability.
    for (i = 1; i <= 5; i = i + 1) begin
      bursty_requests.push_back(i);
    end
    
    // Process bursty requests
    fork
      // Allow initial burst of 3 requests
      for (i = 0; i < 3; i = i + 1) begin
        if (bursty_requests.size() > 0) begin
          $display("bursty request %0d %0t", bursty_requests.pop_front(), $time);
        end
      end
      
      // Then process remaining requests with rate limiting
      forever begin
        #200;
        if (bursty_requests.size() > 0) begin
          $display("bursty request %0d %0t", bursty_requests.pop_front(), $time);
        end
      end
    join
  end
  
  // Run simulation for 2000 time units
  initial #2000 $finish;
endmodule

To run this Verilog simulation, you would typically use a Verilog simulator like ModelSim, VCS, or Icarus Verilog. The exact command might vary depending on your simulator, but it generally looks like this:

$ iverilog rate_limiter.v
$ ./a.out

Running our simulation, we’d see the first batch of requests handled once every ~200 time units as desired:

request 1 200
request 2 400
request 3 600
request 4 800
request 5 1000

For the second batch of requests (bursty requests), we’d see the first 3 handled immediately due to the burst capability, then the remaining 2 with ~200 time unit delays each:

bursty request 1 0
bursty request 2 0
bursty request 3 0
bursty request 4 200
bursty request 5 400

Note that Verilog, being a hardware description language, doesn’t have built-in concepts like goroutines or channels. Instead, we use parallel initial blocks and queues to simulate similar behavior. The timing is controlled using delay statements (#) and the $time system task.

Waitgroups in Verilog

Atomic Counters in Verilog

Learn X By Example

Title here

Rate Limiting in Verilog