Java NIO?看這一篇就夠了!

現在使用NIO的場景越來越多,很多網上的技術框架或多或少的使用NIO技術,譬如Tomcat,Jetty。學習和掌握NIO技術已經不是一個JAVA攻城獅的加分技能,而是一個必備技能。在前面2篇文章《什麼是Zero-Copy?》和《NIO相關基礎篇》中我們學習了NIO的相關理論知識,而在本篇中我們一起來學習一下Java NIO的實戰知識。全文較長,建議先馬後看(記得關注不迷路)。

一、概述

NIO主要有三大核心部分:Channel(通道),Buffer(緩衝區), Selector。傳統IO基於字節流和字符流進行操作,而NIO基於Channel和Buffer(緩衝區)進行操作,數據總是從通道讀取到緩衝區中,或者從緩衝區寫入到通道中。Selector(選擇區)用於監聽多個通道的事件(比如:連接打開,數據到達)。因此,單個線程可以監聽多個數據通道。

NIO和傳統IO(一下簡稱IO)之間第一個最大的區別是,IO是面向流的,NIO是面向緩衝區的。 Java IO面向流意味着每次從流中讀一個或多個字節,直至讀取所有字節,它們沒有被緩存在任何地方。此外,它不能前後移動流中的數據。如果需要前後移動從流中讀取的數據,需要先將它緩存到一個緩衝區。NIO的緩衝導向方法略有不同。數據讀取到一個它稍後處理的緩衝區,需要時可在緩衝區中前後移動。這就增加了處理過程中的靈活性。但是,還需要檢查是否該緩衝區中包含所有您需要處理的數據。而且,需確保當更多的數據讀入緩衝區時,不要覆蓋緩衝區裏尚未處理的數據。

IO的各種流是阻塞的。這意味着,當一個線程調用read() 或 write()時,該線程被阻塞,直到有一些數據被讀取,或數據完全寫入。該線程在此期間不能再幹任何事情了。 NIO的非阻塞模式,使一個線程從某通道發送請求讀取數據,但是它僅能得到目前可用的數據,如果目前沒有數據可用時,就什麼都不會獲取。而不是保持線程阻塞,所以直至數據變得可以讀取之前,該線程可以繼續做其他的事情。 非阻塞寫也是如此。一個線程請求寫入一些數據到某通道,但不需要等待它完全寫入,這個線程同時可以去做別的事情。 線程通常將非阻塞IO的空閒時間用於在其它通道上執行IO操作,所以一個單獨的線程現在可以管理多個輸入和輸出通道(channel)。

Channel

首先說一下Channel,國內大多翻譯成“通道”。Channel和IO中的Stream(流)是差不多一個等級的。只不過Stream是單向的,譬如:InputStream, OutputStream.而Channel是雙向的,既可以用來進行讀操作,又可以用來進行寫操作。
NIO中的Channel的主要實現有:

FileChannel

DatagramChannel

SocketChannel

ServerSocketChannel

這裏看名字就可以猜出個所以然來:分別可以對應文件IO、UDP和TCP(Server和Client)。下面演示的案例基本上就是圍繞這4個類型的Channel進行陳述的。

Buffer

NIO中的關鍵Buffer實現有:ByteBuffer, CharBuffer, DoubleBuffer, FloatBuffer, IntBuffer, LongBuffer, ShortBuffer,分別對應基本數據類型: byte, char, double, float, int, long, short。當然NIO中還有MappedByteBuffer, HeapByteBuffer, DirectByteBuffer等這裏先不進行陳述。

Selector

Selector運行單線程處理多個Channel,如果你的應用打開了多個通道,但每個連接的流量都很低,使用Selector就會很方便。例如在一個聊天服務器中。要使用Selector, 得向Selector註冊Channel,然後調用它的select()方法。這個方法會一直阻塞到某個註冊的通道有事件就緒。一旦這個方法返回,線程就可以處理這些事件,事件的例子有如新的連接進來、數據接收等。

二、FileChannel

看完上面的陳述,對於第一次接觸NIO的同學來說雲裏霧裏,只說了一些概念,也沒記住什麼,更別說怎麼用了。這裏開始通過傳統IO以及更改後的NIO來做對比,以更形象的突出NIO的用法,進而使你對NIO有一點點的瞭解。

傳統IO vs NIO

首先,案例1是採用FileInputStream讀取文件內容的:

    public static void method2(){
        InputStream in = null;
        try{
            in = new BufferedInputStream(new FileInputStream("src/nomal_io.txt"));
            byte [] buf = new byte[1024];
            int bytesRead = in.read(buf);
            while(bytesRead != -1)
            {
                for(int i=0;i<bytesRead;i++)
                    System.out.print((char)buf[i]);
                bytesRead = in.read(buf);
            }
        }catch (IOException e)
        {
            e.printStackTrace();
        }finally{
            try{
                if(in != null){
                    in.close();
                }
            }catch (IOException e){
                e.printStackTrace();
            }
        }
    }

輸出結果:(略)

案例是對應的NIO(這裏通過RandomAccessFile進行操作,當然也可以通過FileInputStream.getChannel()進行操作):

    public static void method1(){
        RandomAccessFile aFile = null;
        try{
            aFile = new RandomAccessFile("src/nio.txt","rw");
            FileChannel fileChannel = aFile.getChannel();
            ByteBuffer buf = ByteBuffer.allocate(1024);
            int bytesRead = fileChannel.read(buf);
            System.out.println(bytesRead);
            while(bytesRead != -1)
            {
                buf.flip();
                while(buf.hasRemaining())
                {
                    System.out.print((char)buf.get());
                }
                buf.compact();
                bytesRead = fileChannel.read(buf);
            }
        }catch (IOException e){
            e.printStackTrace();
        }finally{
            try{
                if(aFile != null){
                    aFile.close();
                }
            }catch (IOException e){
                e.printStackTrace();
            }
        }
    }

輸出結果:(略)
通過仔細對比案例1和案例2,應該能看出個大概,最起碼能發現NIO的實現方式比叫複雜。有了一個大概的印象可以進入下一步了。

Buffer的使用

從案例2中可以總結出使用Buffer一般遵循下面幾個步驟:

  • 分配空間(ByteBuffer buf = ByteBuffer.allocate(1024); 還有一種allocateDirector後面再陳述)

  • 寫入數據到Buffer(int bytesRead = fileChannel.read(buf);)

  • 調用filp()方法( buf.flip();)

  • 從Buffer中讀取數據(System.out.print((char)buf.get());)

  • 調用clear()方法或者compact()方法

Buffer顧名思義:緩衝區,實際上是一個容器,一個連續數組。Channel提供從文件、網絡讀取數據的渠道,但是讀寫的數據都必須經過Buffer。如下圖:

 

向Buffer中寫數據:

  • 從Channel寫到Buffer (fileChannel.read(buf))

  • 通過Buffer的put()方法 (buf.put(…))

從Buffer中讀取數據:

  • 從Buffer讀取到Channel (channel.write(buf))

  • 使用get()方法從Buffer中讀取數據 (buf.get())

可以把Buffer簡單地理解爲一組基本數據類型的元素列表,它通過幾個變量來保存這個數據的當前位置狀態:capacity, position, limit, mark:

索引

說明

capacity

緩衝區數組的總長度

position

下一個要操作的數據元素的位置

limit

緩衝區數組中不可操作的下一個元素的位置:limit<=capacity

mark

用於記錄當前position的前一個位置或者默認是-1

 


無圖無真相,舉例:我們通過ByteBuffer.allocate(11)方法創建了一個11個byte的數組的緩衝區,初始狀態如上圖,position的位置爲0,capacity和limit默認都是數組長度。當我們寫入5個字節時,變化如下圖:

這時我們需要將緩衝區中的5個字節數據寫入Channel的通信信道,所以我們調用ByteBuffer.flip()方法,變化如下圖所示(position設回0,並將limit設成之前的position的值):

 

這時底層操作系統就可以從緩衝區中正確讀取這個5個字節數據併發送出去了。在下一次寫數據之前我們再調用clear()方法,緩衝區的索引位置又回到了初始位置。

調用clear()方法:position將被設回0,limit設置成capacity,換句話說,Buffer被清空了,其實Buffer中的數據並未被清除,只是這些標記告訴我們可以從哪裏開始往Buffer裏寫數據。如果Buffer中有一些未讀的數據,調用clear()方法,數據將“被遺忘”,意味着不再有任何標記會告訴你哪些數據被讀過,哪些還沒有。如果Buffer中仍有未讀的數據,且後續還需要這些數據,但是此時想要先寫些數據,那麼使用compact()方法。compact()方法將所有未讀的數據拷貝到Buffer起始處。然後將position設到最後一個未讀元素正後面。limit屬性依然像clear()方法一樣,設置成capacity。現在Buffer準備好寫數據了,但是不會覆蓋未讀的數據。

通過調用Buffer.mark()方法,可以標記Buffer中的一個特定的position,之後可以通過調用Buffer.reset()方法恢復到這個position。Buffer.rewind()方法將position設回0,所以你可以重讀Buffer中的所有數據。limit保持不變,仍然表示能從Buffer中讀取多少個元素。

三、SocketChannel

說完了FileChannel和Buffer, 大家應該對Buffer的用法比較瞭解了,這裏使用SocketChannel來繼續探討NIO。NIO的強大功能部分來自於Channel的非阻塞特性,套接字的某些操作可能會無限期地阻塞。例如,對accept()方法的調用可能會因爲等待一個客戶端連接而阻塞;對read()方法的調用可能會因爲沒有數據可讀而阻塞,直到連接的另一端傳來新的數據。總的來說,創建/接收連接或讀寫數據等I/O調用,都可能無限期地阻塞等待,直到底層的網絡實現發生了什麼。慢速的,有損耗的網絡,或僅僅是簡單的網絡故障都可能導致任意時間的延遲。然而不幸的是,在調用一個方法之前無法知道其是否阻塞。NIO的channel抽象的一個重要特徵就是可以通過配置它的阻塞行爲,以實現非阻塞式的信道。

            channel.configureBlocking(false)

在非阻塞式信道上調用一個方法總是會立即返回。這種調用的返回值指示了所請求的操作完成的程度。例如,在一個非阻塞式ServerSocketChannel上調用accept()方法,如果有連接請求來了,則返回客戶端SocketChannel,否則返回null。

這裏先舉一個TCP應用案例,客戶端採用NIO實現,而服務端依舊使用BIO實現。
客戶端代碼(案例3):

    public static void client(){
        ByteBuffer buffer = ByteBuffer.allocate(1024);
        SocketChannel socketChannel = null;
        try
        {
            socketChannel = SocketChannel.open();
            socketChannel.configureBlocking(false);
            socketChannel.connect(new InetSocketAddress("10.10.195.115",8080));
            if(socketChannel.finishConnect())
            {
                int i=0;
                while(true)
                {
                    TimeUnit.SECONDS.sleep(1);
                    String info = "I'm "+i+++"-th information from client";
                    buffer.clear();
                    buffer.put(info.getBytes());
                    buffer.flip();
                    while(buffer.hasRemaining()){
                        System.out.println(buffer);
                        socketChannel.write(buffer);
                    }
                }
            }
        }
        catch (IOException | InterruptedException e)
        {
            e.printStackTrace();
        }
        finally{
            try{
                if(socketChannel!=null){
                    socketChannel.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }

服務端代碼(案例4):

    public static void server(){
        ServerSocket serverSocket = null;
        InputStream in = null;
        try
        {
            serverSocket = new ServerSocket(8080);
            int recvMsgSize = 0;
            byte[] recvBuf = new byte[1024];
            while(true){
                Socket clntSocket = serverSocket.accept();
                SocketAddress clientAddress = clntSocket.getRemoteSocketAddress();
                System.out.println("Handling client at "+clientAddress);
                in = clntSocket.getInputStream();
                while((recvMsgSize=in.read(recvBuf))!=-1){
                    byte[] temp = new byte[recvMsgSize];
                    System.arraycopy(recvBuf, 0, temp, 0, recvMsgSize);
                    System.out.println(new String(temp));
                }
            }
        }
        catch (IOException e)
        {
            e.printStackTrace();
        }
        finally{
            try{
                if(serverSocket!=null){
                    serverSocket.close();
                }
                if(in!=null){
                    in.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }

輸出結果:(略)

根據案例分析,總結一下SocketChannel的用法。
打開SocketChannel:

            socketChannel = SocketChannel.open();
            socketChannel.connect(new InetSocketAddress("10.10.195.115",8080));

關閉:

            socketChannel.close();

讀取數據:

                    String info = "I'm "+i+++"-th information from client";
                    buffer.clear();
                    buffer.put(info.getBytes());
                    buffer.flip();
                    while(buffer.hasRemaining()){
                        System.out.println(buffer);
                        socketChannel.write(buffer);
                    }

注意SocketChannel.write()方法的調用是在一個while循環中的。write()方法無法保證能寫多少字節到SocketChannel。所以,我們重複調用write()直到Buffer沒有要寫的字節爲止。
非阻塞模式下,read()方法在尚未讀取到任何數據時可能就返回了。所以需要關注它的int返回值,它會告訴你讀取了多少字節。

四、TCP服務端的NIO寫法

到目前爲止,所舉的案例中都沒有涉及Selector。不要急,好東西要慢慢來。Selector類可以用於避免使用阻塞式客戶端中很浪費資源的“忙等”方法。例如,考慮一個IM服務器。像QQ或者旺旺這樣的,可能有幾萬甚至幾千萬個客戶端同時連接到了服務器,但在任何時刻都只是非常少量的消息。

需要讀取和分發。這就需要一種方法阻塞等待,直到至少有一個信道可以進行I/O操作,並指出是哪個信道。NIO的選擇器就實現了這樣的功能。一個Selector實例可以同時檢查一組信道的I/O狀態。用專業術語來說,選擇器就是一個多路開關選擇器,因爲一個選擇器能夠管理多個信道上的I/O操作。然而如果用傳統的方式來處理這麼多客戶端,使用的方法是循環地一個一個地去檢查所有的客戶端是否有I/O操作,如果當前客戶端有I/O操作,則可能把當前客戶端扔給一個線程池去處理,如果沒有I/O操作則進行下一個輪詢,當所有的客戶端都輪詢過了又接着從頭開始輪詢;這種方法是非常笨而且也非常浪費資源,因爲大部分客戶端是沒有I/O操作,我們也要去檢查;而Selector就不一樣了,它在內部可以同時管理多個I/O,當一個信道有I/O操作的時候,他會通知Selector,Selector就是記住這個信道有I/O操作,並且知道是何種I/O操作,是讀呢?是寫呢?還是接受新的連接;所以如果使用Selector,它返回的結果只有兩種結果,一種是0,即在你調用的時刻沒有任何客戶端需要I/O操作,另一種結果是一組需要I/O操作的客戶端,這時你就根本不需要再檢查了,因爲它返回給你的肯定是你想要的。這樣一種通知的方式比那種主動輪詢的方式要高效得多!

要使用選擇器(Selector),需要創建一個Selector實例(使用靜態工廠方法open())並將其註冊(register)到想要監控的信道上(注意,這要通過channel的方法實現,而不是使用selector的方法)。最後,調用選擇器的select()方法。該方法會阻塞等待,直到有一個或更多的信道準備好了I/O操作或等待超時。select()方法將返回可進行I/O操作的信道數量。現在,在一個單獨的線程中,通過調用select()方法就能檢查多個信道是否準備好進行I/O操作。如果經過一段時間後仍然沒有信道準備好,select()方法就會返回0,並允許程序繼續執行其他任務。

下面將上面的TCP服務端代碼改寫成NIO的方式(案例5):

public class ServerConnect
{
    private static final int BUF_SIZE=1024;
    private static final int PORT = 8080;
    private static final int TIMEOUT = 3000;
    public static void main(String[] args)
    {
        selector();
    }
    public static void handleAccept(SelectionKey key) throws IOException{
        ServerSocketChannel ssChannel = (ServerSocketChannel)key.channel();
        SocketChannel sc = ssChannel.accept();
        sc.configureBlocking(false);
        sc.register(key.selector(), SelectionKey.OP_READ,ByteBuffer.allocateDirect(BUF_SIZE));
    }
    public static void handleRead(SelectionKey key) throws IOException{
        SocketChannel sc = (SocketChannel)key.channel();
        ByteBuffer buf = (ByteBuffer)key.attachment();
        long bytesRead = sc.read(buf);
        while(bytesRead>0){
            buf.flip();
            while(buf.hasRemaining()){
                System.out.print((char)buf.get());
            }
            System.out.println();
            buf.clear();
            bytesRead = sc.read(buf);
        }
        if(bytesRead == -1){
            sc.close();
        }
    }
    public static void handleWrite(SelectionKey key) throws IOException{
        ByteBuffer buf = (ByteBuffer)key.attachment();
        buf.flip();
        SocketChannel sc = (SocketChannel) key.channel();
        while(buf.hasRemaining()){
            sc.write(buf);
        }
        buf.compact();
    }
    public static void selector() {
        Selector selector = null;
        ServerSocketChannel ssc = null;
        try{
            selector = Selector.open();
            ssc= ServerSocketChannel.open();
            ssc.socket().bind(new InetSocketAddress(PORT));
            ssc.configureBlocking(false);
            ssc.register(selector, SelectionKey.OP_ACCEPT);
            while(true){
                if(selector.select(TIMEOUT) == 0){
                    System.out.println("==");
                    continue;
                }
                Iterator<SelectionKey> iter = selector.selectedKeys().iterator();
                while(iter.hasNext()){
                    SelectionKey key = iter.next();
                    if(key.isAcceptable()){
                        handleAccept(key);
                    }
                    if(key.isReadable()){
                        handleRead(key);
                    }
                    if(key.isWritable() && key.isValid()){
                        handleWrite(key);
                    }
                    if(key.isConnectable()){
                        System.out.println("isConnectable = true");
                    }
                    iter.remove();
                }
            }
        }catch(IOException e){
            e.printStackTrace();
        }finally{
            try{
                if(selector!=null){
                    selector.close();
                }
                if(ssc!=null){
                    ssc.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }
}

下面來慢慢講解這段代碼。

ServerSocketChannel

打開ServerSocketChannel:

ServerSocketChannel serverSocketChannel = ServerSocketChannel.open();

關閉ServerSocketChannel:

serverSocketChannel.close();

監聽新進來的連接:

while(true){
    SocketChannel socketChannel = serverSocketChannel.accept();
}

ServerSocketChannel可以設置成非阻塞模式。在非阻塞模式下,accept() 方法會立刻返回,如果還沒有新進來的連接,返回的將是null。 因此,需要檢查返回的SocketChannel是否是null.如:

        ServerSocketChannel serverSocketChannel = ServerSocketChannel.open();
        serverSocketChannel.socket().bind(new InetSocketAddress(9999));
        serverSocketChannel.configureBlocking(false);
        while (true)
        {
            SocketChannel socketChannel = serverSocketChannel.accept();
            if (socketChannel != null)
            {
                // do something with socketChannel...
            }
        }

Selector

Selector的創建:Selector selector = Selector.open();

爲了將Channel和Selector配合使用,必須將Channel註冊到Selector上,通過SelectableChannel.register()方法來實現,沿用案例5中的部分代碼:

            ssc= ServerSocketChannel.open();
            ssc.socket().bind(new InetSocketAddress(PORT));
            ssc.configureBlocking(false);
            ssc.register(selector, SelectionKey.OP_ACCEPT);

與Selector一起使用時,Channel必須處於非阻塞模式下。這意味着不能將FileChannel與Selector一起使用,因爲FileChannel不能切換到非阻塞模式。而套接字通道都可以。

注意register()方法的第二個參數。這是一個“interest集合”,意思是在通過Selector監聽Channel時對什麼事件感興趣。可以監聽四種不同類型的事件:

1. Connect
2. Accept
3. Read
4. Write

通道觸發了一個事件意思是該事件已經就緒。所以,某個channel成功連接到另一個服務器稱爲“連接就緒”。一個server socket channel準備好接收新進入的連接稱爲“接收就緒”。一個有數據可讀的通道可以說是“讀就緒”。等待寫數據的通道可以說是“寫就緒”。

這四種事件用SelectionKey的四個常量來表示:

1. SelectionKey.OP_CONNECT
2. SelectionKey.OP_ACCEPT
3. SelectionKey.OP_READ
4. SelectionKey.OP_WRITE

SelectionKey

當向Selector註冊Channel時,register()方法會返回一個SelectionKey對象。這個對象包含了一些你感興趣的屬性:

  • interest集合

  • ready集合

  • Channel

  • Selector

  • 附加的對象(可選)

interest集合:就像向Selector註冊通道一節中所描述的,interest集合是你所選擇的感興趣的事件集合。可以通過SelectionKey讀寫interest集合。

ready 集合是通道已經準備就緒的操作的集合。在一次選擇(Selection)之後,你會首先訪問這個ready set。Selection將在下一小節進行解釋。可以這樣訪問ready集合:

int readySet = selectionKey.readyOps();

可以用像檢測interest集合那樣的方法,來檢測channel中什麼事件或操作已經就緒。但是,也可以使用以下四個方法,它們都會返回一個布爾類型:

selectionKey.isAcceptable();
selectionKey.isConnectable();
selectionKey.isReadable();
selectionKey.isWritable();

從SelectionKey訪問Channel和Selector很簡單。如下:

Channel  channel  = selectionKey.channel();
Selector selector = selectionKey.selector();

可以將一個對象或者更多信息附着到SelectionKey上,這樣就能方便的識別某個給定的通道。例如,可以附加 與通道一起使用的Buffer,或是包含聚集數據的某個對象。使用方法如下:

selectionKey.attach(theObject);
Object attachedObj = selectionKey.attachment();

還可以在用register()方法向Selector註冊Channel的時候附加對象。如:

SelectionKey key = channel.register(selector, SelectionKey.OP_READ, theObject);

通過Selector選擇通道

一旦向Selector註冊了一或多個通道,就可以調用幾個重載的select()方法。這些方法返回你所感興趣的事件(如連接、接受、讀或寫)已經準備就緒的那些通道。換句話說,如果你對“讀就緒”的通道感興趣,select()方法會返回讀事件已經就緒的那些通道。

下面是select()方法:

  • int select()

  • int select(long timeout)

  • int selectNow()

select()阻塞到至少有一個通道在你註冊的事件上就緒了。
select(long timeout)和select()一樣,除了最長會阻塞timeout毫秒(參數)。
selectNow()不會阻塞,不管什麼通道就緒都立刻返回(譯者注:此方法執行非阻塞的選擇操作。如果自從前一次選擇操作後,沒有通道變成可選擇的,則此方法直接返回零。)。

select()方法返回的int值表示有多少通道已經就緒。亦即,自上次調用select()方法後有多少通道變成就緒狀態。如果調用select()方法,因爲有一個通道變成就緒狀態,返回了1,若再次調用select()方法,如果另一個通道就緒了,它會再次返回1。如果對第一個就緒的channel沒有做任何操作,現在就有兩個就緒的通道,但在每次select()方法調用之間,只有一個通道就緒了。

一旦調用了select()方法,並且返回值表明有一個或更多個通道就緒了,然後可以通過調用selector的selectedKeys()方法,訪問“已選擇鍵集(selected key set)”中的就緒通道。如下所示:

Set selectedKeys = selector.selectedKeys();

當向Selector註冊Channel時,Channel.register()方法會返回一個SelectionKey 對象。這個對象代表了註冊到該Selector的通道。

注意每次迭代末尾的keyIterator.remove()調用。Selector不會自己從已選擇鍵集中移除SelectionKey實例。必須在處理完通道時自己移除。下次該通道變成就緒時,Selector會再次將其放入已選擇鍵集中。

SelectionKey.channel()方法返回的通道需要轉型成你要處理的類型,如ServerSocketChannel或SocketChannel等。

一個完整的使用Selector和ServerSocketChannel的案例可以參考案例5的selector()方法。


五、內存映射文件

JAVA處理大文件,一般用BufferedReader,BufferedInputStream這類帶緩衝的IO類,不過如果文件超大的話,更快的方式是採用MappedByteBuffer。

MappedByteBuffer是NIO引入的文件內存映射方案,讀寫性能極高。NIO最主要的就是實現了對異步操作的支持。其中一種通過把一個套接字通道(SocketChannel)註冊到一個選擇器(Selector)中,不時調用後者的選擇(select)方法就能返回滿足的選擇鍵(SelectionKey),鍵中包含了SOCKET事件信息。這就是select模型。

SocketChannel的讀寫是通過一個類叫ByteBuffer來操作的.這個類本身的設計是不錯的,比直接操作byte[]方便多了. ByteBuffer有兩種模式:直接/間接.間接模式最典型(也只有這麼一種)的就是HeapByteBuffer,即操作堆內存 (byte[]).但是內存畢竟有限,如果我要發送一個1G的文件怎麼辦?不可能真的去分配1G的內存.這時就必須使用"直接"模式,即 MappedByteBuffer,文件映射.

先中斷一下,談談操作系統的內存管理.一般操作系統的內存分兩部分:物理內存;虛擬內存.虛擬內存一般使用的是頁面映像文件,即硬盤中的某個(某些)特殊的文件.操作系統負責頁面文件內容的讀寫,這個過程叫"頁面中斷/切換". MappedByteBuffer也是類似的,你可以把整個文件(不管文件有多大)看成是一個ByteBuffer.MappedByteBuffer 只是一種特殊的ByteBuffer,即是ByteBuffer的子類。 MappedByteBuffer 將文件直接映射到內存(這裏的內存指的是虛擬內存,並不是物理內存)。通常,可以映射整個文件,如果文件比較大的話可以分段進行映射,只要指定文件的那個部分就可以。

概念

FileChannel提供了map方法來把文件影射爲內存映像文件: MappedByteBuffer map(int mode,long position,long size); 可以把文件的從position開始的size大小的區域映射爲內存映像文件,mode指出了 可訪問該內存映像文件的方式:

  • READ_ONLY,(只讀): 試圖修改得到的緩衝區將導致拋出 ReadOnlyBufferException.(MapMode.READ_ONLY)

  • READ_WRITE(讀/寫): 對得到的緩衝區的更改最終將傳播到文件;該更改對映射到同一文件的其他程序不一定是可見的。 (MapMode.READ_WRITE)

  • PRIVATE(專用): 對得到的緩衝區的更改不會傳播到文件,並且該更改對映射到同一文件的其他程序也不是可見的;相反,會創建緩衝區已修改部分的專用副本。 (MapMode.PRIVATE)

MappedByteBuffer是ByteBuffer的子類,其擴充了三個方法:

  • force():緩衝區是READ_WRITE模式下,此方法對緩衝區內容的修改強行寫入文件;

  • load():將緩衝區的內容載入內存,並返回該緩衝區的引用;

  • isLoaded():如果緩衝區的內容在物理內存中,則返回真,否則返回假;

案例對比

這裏通過採用ByteBuffer和MappedByteBuffer分別讀取大小約爲5M的文件"src/1.ppt"來比較兩者之間的區別,method3()是採用MappedByteBuffer讀取的,method4()對應的是ByteBuffer。

    public static void method4(){
        RandomAccessFile aFile = null;
        FileChannel fc = null;
        try{
            aFile = new RandomAccessFile("src/1.ppt","rw");
            fc = aFile.getChannel();
            long timeBegin = System.currentTimeMillis();
            ByteBuffer buff = ByteBuffer.allocate((int) aFile.length());
            buff.clear();
            fc.read(buff);
            //System.out.println((char)buff.get((int)(aFile.length()/2-1)));
            //System.out.println((char)buff.get((int)(aFile.length()/2)));
            //System.out.println((char)buff.get((int)(aFile.length()/2)+1));
            long timeEnd = System.currentTimeMillis();
            System.out.println("Read time: "+(timeEnd-timeBegin)+"ms");
        }catch(IOException e){
            e.printStackTrace();
        }finally{
            try{
                if(aFile!=null){
                    aFile.close();
                }
                if(fc!=null){
                    fc.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }
    public static void method3(){
        RandomAccessFile aFile = null;
        FileChannel fc = null;
        try{
            aFile = new RandomAccessFile("src/1.ppt","rw");
            fc = aFile.getChannel();
            long timeBegin = System.currentTimeMillis();
            MappedByteBuffer mbb = fc.map(FileChannel.MapMode.READ_ONLY, 0, aFile.length());
            // System.out.println((char)mbb.get((int)(aFile.length()/2-1)));
            // System.out.println((char)mbb.get((int)(aFile.length()/2)));
            //System.out.println((char)mbb.get((int)(aFile.length()/2)+1));
            long timeEnd = System.currentTimeMillis();
            System.out.println("Read time: "+(timeEnd-timeBegin)+"ms");
        }catch(IOException e){
            e.printStackTrace();
        }finally{
            try{
                if(aFile!=null){
                    aFile.close();
                }
                if(fc!=null){
                    fc.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }

通過在入口函數main()中運行:

        method3();
        System.out.println("=============");
        method4();

輸出結果(運行在普通PC機上):

Read time: 2ms
=============
Read time: 12ms

通過輸出結果可以看出彼此的差別,一個例子也許是偶然,那麼下面把5M大小的文件替換爲200M的文件,輸出結果:

Read time: 1ms
=============
Read time: 407ms

可以看到差距拉大。

注:MappedByteBuffer有資源釋放的問題:被MappedByteBuffer打開的文件只有在垃圾收集時纔會被關閉,而這個點是不確定的。在Javadoc中這裏描述:A mapped byte buffer and the file mapping that it represents remian valid until the buffer itself is garbage-collected。詳細可以翻閱參考資料5和6.


六、其餘功能介紹

看完以上陳述,詳細大家對NIO有了一定的瞭解,下面主要通過幾個案例,來說明NIO的其餘功能,下面代碼量偏多,功能性講述偏少。

Scatter/Gatter

分散(scatter)從Channel中讀取是指在讀操作時將讀取的數據寫入多個buffer中。因此,Channel將從Channel中讀取的數據“分散(scatter)”到多個Buffer中。

聚集(gather)寫入Channel是指在寫操作時將多個buffer的數據寫入同一個Channel,因此,Channel 將多個Buffer中的數據“聚集(gather)”後發送到Channel。

scatter / gather經常用於需要將傳輸的數據分開處理的場合,例如傳輸一個由消息頭和消息體組成的消息,你可能會將消息體和消息頭分散到不同的buffer中,這樣你可以方便的處理消息頭和消息體。

案例:

import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.OutputStream;
import java.nio.ByteBuffer;
import java.nio.channels.Channel;
import java.nio.channels.FileChannel;
public class ScattingAndGather
{
    public static void main(String args[]){
        gather();
    }
    public static void gather()
    {
        ByteBuffer header = ByteBuffer.allocate(10);
        ByteBuffer body = ByteBuffer.allocate(10);
        byte [] b1 = {'0', '1'};
        byte [] b2 = {'2', '3'};
        header.put(b1);
        body.put(b2);
        ByteBuffer [] buffs = {header, body};
        try
        {
            FileOutputStream os = new FileOutputStream("src/scattingAndGather.txt");
            FileChannel channel = os.getChannel();
            channel.write(buffs);
        }
        catch (IOException e)
        {
            e.printStackTrace();
        }
    }
}

transferFrom & transferTo

FileChannel的transferFrom()方法可以將數據從源通道傳輸到FileChannel中。

    public static void method1(){
        RandomAccessFile fromFile = null;
        RandomAccessFile toFile = null;
        try
        {
            fromFile = new RandomAccessFile("src/fromFile.xml","rw");
            FileChannel fromChannel = fromFile.getChannel();
            toFile = new RandomAccessFile("src/toFile.txt","rw");
            FileChannel toChannel = toFile.getChannel();
            long position = 0;
            long count = fromChannel.size();
            System.out.println(count);
            toChannel.transferFrom(fromChannel, position, count);
        }
        catch (IOException e)
        {
            e.printStackTrace();
        }
        finally{
            try{
                if(fromFile != null){
                    fromFile.close();
                }
                if(toFile != null){
                    toFile.close();
                }
            }
            catch(IOException e){
                e.printStackTrace();
            }
        }
    }

方法的輸入參數position表示從position處開始向目標文件寫入數據,count表示最多傳輸的字節數。如果源通道的剩餘空間小於 count 個字節,則所傳輸的字節數要小於請求的字節數。此外要注意,在SoketChannel的實現中,SocketChannel只會傳輸此刻準備好的數據(可能不足count字節)。因此,SocketChannel可能不會將請求的所有數據(count個字節)全部傳輸到FileChannel中。

transferTo()方法將數據從FileChannel傳輸到其他的channel中。

    public static void method2()
    {
        RandomAccessFile fromFile = null;
        RandomAccessFile toFile = null;
        try
        {
            fromFile = new RandomAccessFile("src/fromFile.txt","rw");
            FileChannel fromChannel = fromFile.getChannel();
            toFile = new RandomAccessFile("src/toFile.txt","rw");
            FileChannel toChannel = toFile.getChannel();
            long position = 0;
            long count = fromChannel.size();
            System.out.println(count);
            fromChannel.transferTo(position, count,toChannel);
        }
        catch (IOException e)
        {
            e.printStackTrace();
        }
        finally{
            try{
                if(fromFile != null){
                    fromFile.close();
                }
                if(toFile != null){
                    toFile.close();
                }
            }
            catch(IOException e){
                e.printStackTrace();
            }
        }
    }

上面所說的關於SocketChannel的問題在transferTo()方法中同樣存在。SocketChannel會一直傳輸數據直到目標buffer被填滿。

Pipe

Java NIO 管道是2個線程之間的單向數據連接。Pipe有一個source通道和一個sink通道。數據會被寫到sink通道,從source通道讀取。

    public static void method1(){
        Pipe pipe = null;
        ExecutorService exec = Executors.newFixedThreadPool(2);
        try{
            pipe = Pipe.open();
            final Pipe pipeTemp = pipe;
            exec.submit(new Callable<Object>(){
                @Override
                public Object call() throws Exception
                {
                    Pipe.SinkChannel sinkChannel = pipeTemp.sink();//向通道中寫數據
                    while(true){
                        TimeUnit.SECONDS.sleep(1);
                        String newData = "Pipe Test At Time "+System.currentTimeMillis();
                        ByteBuffer buf = ByteBuffer.allocate(1024);
                        buf.clear();
                        buf.put(newData.getBytes());
                        buf.flip();
                        while(buf.hasRemaining()){
                            System.out.println(buf);
                            sinkChannel.write(buf);
                        }
                    }
                }
            });
            exec.submit(new Callable<Object>(){
                @Override
                public Object call() throws Exception
                {
                    Pipe.SourceChannel sourceChannel = pipeTemp.source();//向通道中讀數據
                    while(true){
                        TimeUnit.SECONDS.sleep(1);
                        ByteBuffer buf = ByteBuffer.allocate(1024);
                        buf.clear();
                        int bytesRead = sourceChannel.read(buf);
                        System.out.println("bytesRead="+bytesRead);
                        while(bytesRead >0 ){
                            buf.flip();
                            byte b[] = new byte[bytesRead];
                            int i=0;
                            while(buf.hasRemaining()){
                                b[i]=buf.get();
                                System.out.printf("%X",b[i]);
                                i++;
                            }
                            String s = new String(b);
                            System.out.println("=================||"+s);
                            bytesRead = sourceChannel.read(buf);
                        }
                    }
                }
            });
        }catch(IOException e){
            e.printStackTrace();
        }finally{
            exec.shutdown();
        }
    }

DatagramChannel

Java NIO中的DatagramChannel是一個能收發UDP包的通道。因爲UDP是無連接的網絡協議,所以不能像其它通道那樣讀取和寫入。它發送和接收的是數據包。

    public static void  reveive(){
        DatagramChannel channel = null;
        try{
            channel = DatagramChannel.open();
            channel.socket().bind(new InetSocketAddress(8888));
            ByteBuffer buf = ByteBuffer.allocate(1024);
            buf.clear();
            channel.receive(buf);
            buf.flip();
            while(buf.hasRemaining()){
                System.out.print((char)buf.get());
            }
            System.out.println();
        }catch(IOException e){
            e.printStackTrace();
        }finally{
            try{
                if(channel!=null){
                    channel.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }
    public static void send(){
        DatagramChannel channel = null;
        try{
            channel = DatagramChannel.open();
            String info = "I'm the Sender!";
            ByteBuffer buf = ByteBuffer.allocate(1024);
            buf.clear();
            buf.put(info.getBytes());
            buf.flip();
            int bytesSent = channel.send(buf, new InetSocketAddress("10.10.195.115",8888));
            System.out.println(bytesSent);
        }catch(IOException e){
            e.printStackTrace();
        }finally{
            try{
                if(channel!=null){
                    channel.close();
                }
            }catch(IOException e){
                e.printStackTrace();
            }
        }
    }

文章出處:https://mp.weixin.qq.com/s/c9tkrokcDQR375kiwCeV9w? 

發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章