Grizzly RoundRobinConnectionDistributor 的變化

grizzly1.9:

private SelectorRunner getSelectorRunner(int interestOps) {    
       SelectorRunner[] runners = getTransportSelectorRunners();    
       int index;    
       if (interestOps == SelectionKey.OP_ACCEPT || runners.length == 1) {    
           index = 0;    
       } else {    
           index = (counter.incrementAndGet() % (runners.length - 1)) + 1;    
       }    
          
       return runners[index];    
   }   

一個SelectorRunner 負責accept,其他負責read/write

grizzly2後

private SelectorRunner getSelectorRunner() {  
        final SelectorRunner[] runners = getTransportSelectorRunners();  
        final int index = counter.getAndIncrement() % runners.length;  
          
        return runners[index];  
    }   

不再區分accept/read/write 的SelectRunner.

諮詢了作者,主要的架構權衡的問題，如果是系統的連接不多，如果區分accept和read/write,accept的 SelectRunner就會空閒，無事可做，而read/write 的SelectRunner很忙。如果連接很多，可以考慮類似1.9的實現，區分accept和read/write的 SelectRunnerSelectRunner.