生產者-消費者模式是經典的多線程設計模式,它為多線程間的協作提供了良好的解決方案。在生產者-消費者模式中,有兩類線程:若干個生產者線程和若干個消費者線程。生產者負責提交用戶請求,消費者用於具體的處理生產者提交的任務。生產者和消費者通過共享內存緩沖區進行數據通信。
生產者-消費者模式的基本結構如下圖:
通過上圖可以看出,生產者和消費者通過共享內存緩沖區進行通信,他們之間並不進行直接的通信,從而減少了他們之間的耦合,生產者不需要直到消費者的存在,消費者也不需要知道生產者的存在。內存緩沖區主要的功能是實現數據在多線程間的共享,此外,通過該緩沖區,還可以緩解生產者和消費者之間的性能差異。
下面以生產者-消費者模式的簡單實現,介紹該模式的優點:
生產者代碼:
1 public class Producer implements Runnable { 2 3 private volatile boolean isRunnig = true; 4 5 private BlockingQueue<PCData> queue;//緩沖隊列 6 7 private static AtomicInteger count = new AtomicInteger(); 8 9 private static final int SLEEPTIME = 1000; 10 11 12 public Producer(BlockingQueue<PCData> queue) { 13 this.queue = queue; 14 } 15 16 17 @Override 18 public void run() { 19 PCData pcData = null; 20 Random r = new Random(); 21 22 System.out.println("start producer id:"+Thread.currentThread().getId()); 23 24 try { 25 while(true){ 26 Thread.sleep(r.nextInt(SLEEPTIME)); 27 pcData = new PCData(count.incrementAndGet()); 28 System.out.println(pcData +"is put into queue"); 29 if(!queue.offer(pcData, 2, TimeUnit.SECONDS)){ 30 System.err.println("fail to put data:"+pcData); 31 } 32 } 33 } catch (Exception e) { 34 // TODO: handle exception 35 e.printStackTrace(); 36 Thread.currentThread().interrupt(); 37 } 38 39 } 40 41 public void stop(){ 42 isRunnig = false; 43 } 44 45 46 }
消費者代碼:
1 public class Consumer implements Runnable { 2 3 private BlockingQueue<PCData> queue;//緩沖隊列 4 5 private static final int SLEEPTIME=1000; 6 7 8 9 public Consumer(BlockingQueue<PCData> queue) { 10 this.queue = queue; 11 } 12 13 14 15 @Override 16 public void run() { 17 System.out.println("start constomer id:"+Thread.currentThread().getId()); 18 Random r = new Random(); 19 try { 20 while(true){ 21 PCData data = queue.take(); 22 int re = data.getIntData()*data.getIntData(); 23 System.out.println(MessageFormat.format("{0}*{1}={2}", data.getIntData(),data.getIntData(),re)); 24 25 Thread.sleep(r.nextInt(SLEEPTIME)); 26 } 27 } catch (Exception e) { 28 // TODO: handle exception 29 e.printStackTrace(); 30 Thread.currentThread().interrupt(); 31 } 32 33 } 34 35 }
消費者、生產者之間的共享數據模型:
1 public final class PCData { 2 private final int intData; 3 4 public PCData(int intData) { 5 this.intData = intData; 6 } 7 8 public PCData(String strData){ 9 this.intData = Integer.valueOf(strData); 10 } 11 12 public synchronized int getIntData() { 13 return intData; 14 } 15 16 @Override 17 public String toString() { 18 return "data:" + intData; 19 } 20 21 22 23 }
在客戶端中,啟動3個消費者和3個生產者,並讓他們協作運行:
1 public class Client { 2 public static void main(String[] args) throws InterruptedException { 3 BlockingQueue<PCData> queue = new LinkedBlockingQueue<PCData>(10); 4 5 Producer p1 = new Producer(queue); 6 Producer p2 = new Producer(queue); 7 Producer p3 = new Producer(queue); 8 9 Consumer c1 = new Consumer(queue); 10 Consumer c2 = new Consumer(queue); 11 Consumer c3 = new Consumer(queue); 12 13 ExecutorService exe = Executors.newCachedThreadPool(); 14 exe.execute(p1); 15 exe.execute(p2); 16 exe.execute(p3); 17 18 exe.execute(c1); 19 exe.execute(c2); 20 exe.execute(c3); 21 22 Thread.sleep(10*1000); 23 24 p1.stop(); 25 p2.stop(); 26 p3.stop(); 27 28 Thread.sleep(3000); 29 exe.shutdown(); 30 } 31 }
優點:生產者-消費者模式能很好的對生產者線程和消費者線程進行解耦,優化系統的整體結構。同時,由於緩沖區的存在,運行生產者和消費者在性能上存在一定的差異,從而一定程度上緩解了性能瓶頸對系統性能的影響。