Scalable IO in Java

Scalable IO in Java

http://gee.cs.oswego.edu/dl/cpjslides/nio.pdf

大部分IO都是下面這個步驟,

Most have same basic structure:
Read request
Decode request
Process service
Encode reply
Send reply

關鍵是如何處理並發, 最原始就是單純的用多線程

class Server implements Runnable {
    public void run() {
        try {
            ServerSocket ss = new ServerSocket(PORT);
            while (!Thread.interrupted())
            new Thread(new Handler(ss.accept())).start(); //創建新線程來handle
            // or, single-threaded, or a thread pool
        } catch (IOException ex) { /* ... */ }
    }
    
    static class Handler implements Runnable {
        final Socket socket;
        Handler(Socket s) { socket = s; }
        public void run() {
            try {
                byte[] input = new byte[MAX_INPUT];
                socket.getInputStream().read(input);
                byte[] output = process(input);
                socket.getOutputStream().write(output);
            } catch (IOException ex) { /* ... */ }
        }       
        private byte[] process(byte[] cmd) { /* ... */ }
    }
}

顯然簡單的多線程會帶來擴展性問題, 當client數量變的很多的時候, 還其他的可用性, 性能的問題
解決方法就是Divide-and-conquer, 分開后, 就需要Event-driven Designs來串聯起來...

 

單線程版本的Reactor, 所有事情read, process, write都由單個線程完成, 完成一步重新設置下一步的event, 然后干其他的事
問題當然就是, 其中任何步驟不能消耗太多時間, 因為只有一個線程, 你占住了就會block其他任務
ps, 不明白為什么要畫那么大個acceptor, 只是作為第一步的callback對象...

看代碼會更清楚,

class Reactor implements Runnable { 
    final Selector selector;
    final ServerSocketChannel serverSocket;
    Reactor(int port) throws IOException { //Reactor初始化
        selector = Selector.open();
        serverSocket = ServerSocketChannel.open();
        serverSocket.socket().bind(new InetSocketAddress(port));
        serverSocket.configureBlocking(false); //非阻塞
        SelectionKey sk = serverSocket.register(selector, SelectionKey.OP_ACCEPT); //分步處理,第一步,接收accept事件
        sk.attach(new Acceptor()); //attach callback object, Acceptor
    }
    
    public void run() { 
        try {
            while (!Thread.interrupted()) {
                selector.select();
                Set selected = selector.selectedKeys();
                Iterator it = selected.iterator();
                while (it.hasNext())
                    dispatch((SelectionKey)(it.next()); //Reactor負責dispatch收到的事件
                selected.clear();
            }
        } catch (IOException ex) { /* ... */ }
    }
    
    void dispatch(SelectionKey k) {
    	Runnable r = (Runnable)(k.attachment()); //調用之前注冊的callback對象
    	if (r != null)
    	    r.run();
    }
    
    class Acceptor implements Runnable { // inner
        public void run() {
            try {
                SocketChannel c = serverSocket.accept();
                if (c != null)
                new Handler(selector, c);
            }
            catch(IOException ex) { /* ... */ }
        }
    }
}

final class Handler implements Runnable {
    final SocketChannel socket;
    final SelectionKey sk;
    ByteBuffer input = ByteBuffer.allocate(MAXIN);
    ByteBuffer output = ByteBuffer.allocate(MAXOUT);
    static final int READING = 0, SENDING = 1;
    int state = READING;
    
    Handler(Selector sel, SocketChannel c) throws IOException {
        socket = c; c.configureBlocking(false);
        // Optionally try first read now
        sk = socket.register(sel, 0);
        sk.attach(this); //將Handler作為callback對象
        sk.interestOps(SelectionKey.OP_READ); //第二步,接收Read事件
        sel.wakeup();
    }
    boolean inputIsComplete() { /* ... */ }
    boolean outputIsComplete() { /* ... */ }
    void process() { /* ... */ }
    
    public void run() {
        try {
            if (state == READING) read();
            else if (state == SENDING) send();
        } catch (IOException ex) { /* ... */ }
    }
    
    void read() throws IOException {
        socket.read(input);
        if (inputIsComplete()) {
            process();
            state = SENDING;
            // Normally also do first write now
            sk.interestOps(SelectionKey.OP_WRITE); //第三步,接收write事件
        }
    }
    void send() throws IOException {
        socket.write(output);
        if (outputIsComplete()) sk.cancel(); //write完就結束了, 關閉select key
    }
}

//上面 的實現用Handler來同時處理Read和Write事件, 所以里面出現狀態判斷
//我們可以用State-Object pattern來更優雅的實現
class Handler { // ...
    public void run() { // initial state is reader
        socket.read(input);
        if (inputIsComplete()) {
            process();
            sk.attach(new Sender());  //狀態遷移, Read后變成write, 用Sender作為新的callback對象
              sk.interest(SelectionKey.OP_WRITE);
            sk.selector().wakeup();
        }
    }
    class Sender implements Runnable {
        public void run(){ // ...
            socket.write(output);
            if (outputIsComplete()) sk.cancel();
        }
    }
}

 

單線程模式的局限還是比較明顯的
所以改進是, 將比較耗時的部分, 從reactor線程中分離出去, 讓reactor專門負責IO
而另外創建Thread Pool和queue來緩存和處理任務
所以其實已經進化成Proactor模式, 異步模式

 

class Handler implements Runnable {
    // uses util.concurrent thread pool
    static PooledExecutor pool = new PooledExecutor(...);
    static final int PROCESSING = 3;
    // ...
    synchronized void read() { // ...
        socket.read(input);
        if (inputIsComplete()) {
            state = PROCESSING;
            pool.execute(new Processer()); //使用線程pool異步執行
        }
    }
    
    synchronized void processAndHandOff() {
        process();
        state = SENDING; // or rebind attachment
        sk.interest(SelectionKey.OP_WRITE); //process完,開始等待write事件
    }
    
    class Processer implements Runnable {
        public void run() { processAndHandOff(); }
    }
}

使用多個reactor進程, 主reactor只負責accept, 然后將接收到的socketchannel交給subReactor去listen和處理
當然也可以在subReactor下加上線程池進行異步處理
坦白的說, 沒看出用多個reactor有啥大的提升, 降低mainReactor listen的負擔?

Selector[] selectors; //subReactors集合, 一個selector代表一個subReactor
int next = 0;
class Acceptor { // ...
    public synchronized void run() { ...
        Socket connection = serverSocket.accept(); //主selector負責accept
        if (connection != null)
            new Handler(selectors[next], connection); //選個subReactor去負責接收到的connection
        if (++next == selectors.length) next = 0;
    }
}