编解码含义
- 编码(Encode)
编码它将对象序列化为字节数组,用于网络传输、数据持久化或者其他用途。 - 解码(Decode)
解码把从网络、磁盘等读取的字节数组还原成原始对象(通常是原始对象的拷贝),以方便后续的业务逻辑操作。
粘包 拆包
tcp 为什么会出现粘包 拆包的问题?
TCP报文有个比较大的特点,就是它传输的时候,会先把应用层的数据项拆开成字节,然后按照自己的传输需要,选择合适数量的字节进行传输。什么叫"自己的传输需要"?首先TCP包有最大长度限制,那么太大的数据项肯定是要拆开的。其次因为TCP以及下层协议会附加一些协议头信息,如果数据项太小,那么可能报文大部分都是没有价值的头信息,这样传输是很不划算的。
TCP/IP协议与buffer
解决方案
- 消息的定长,例如定1000个字节
- 就是在包尾增加回车或空格等特殊字符作为切割,典型的FTP协议
- 将消息分为消息头消息体,并记录数据长度。例如 dubbo
- netty自带的编解码器
public ChannelPipeline getPipeline() throws Exception {
ChannelPipeline channelPipeline = Channels.pipeline();
channelPipeline.addLast("decoder", new StringDecoder());
channelPipeline.addLast("encoder", new StringEncoder());
channelPipeline.addLast("handler", new ServerLogicHandler());
return channelPipeline;
}
-
dubbo自己的编解码协议
出现粘包拆包的核心是需要将TCP层的字节数据包切分业务层的Request、Response数据包。所以dubbo使用了魔法数和数据长度作为分割符,将Buffer中的数据隔离出来了生成Request、Response这样的一个个业务层的对象。
编解码核心
-
编解码流程
注册编解码器
#com.alibaba.dubbo.remoting.transport.netty.NettyServer#doOpen
bootstrap.setPipelineFactory(new ChannelPipelineFactory() {
public ChannelPipeline getPipeline() {
NettyCodecAdapter adapter = new NettyCodecAdapter(getCodec() ,getUrl(), NettyServer.this);
ChannelPipeline pipeline = Channels.pipeline();
/*int idleTimeout = getIdleTimeout();
if (idleTimeout > 10000) {
pipeline.addLast("timer", new IdleStateHandler(timer, idleTimeout / 1000, 0, 0));
}*/
//解码
pipeline.addLast("decoder", adapter.getDecoder());
//编码
pipeline.addLast("encoder", adapter.getEncoder());
//逻辑处理类
pipeline.addLast("handler", nettyHandler);
return pipeline;
}
});
#com.alibaba.dubbo.remoting.exchange.codec.ExchangeCodec#encode
public void encode(Channel channel, ChannelBuffer buffer, Object msg) throws IOException {
if (msg instanceof Request) {
encodeRequest(channel, buffer, (Request) msg);
} else if (msg instanceof Response) {
encodeResponse(channel, buffer, (Response) msg);
} else {
super.encode(channel, buffer, msg);
}
}
public Object decode(Channel channel, ChannelBuffer buffer) throws IOException {
int readable = buffer.readableBytes();
//读取header,可能可读数据没有完全传输过来
byte[] header = new byte[Math.min(readable, HEADER_LENGTH)];
buffer.readBytes(header);
return decode(channel, buffer, readable, header);
}
Request编码
----------1------consumer请求编码----------------------
-->NettyCodecAdapter.InternalEncoder.encode
-->DubboCountCodec.encode
-->ExchangeCodec.encode
-->ExchangeCodec.encodeRequest
-->DubboCodec.encodeRequestData
consumer发起Request是在调用代理对象的方法后会触发DubboInvoker.doInvoke
使用netty发送Invocktion数据。编码器将一些Request标志位按照编码规则处理成字节数组,将Invocktion序列化并计算出数据长度存放在data部分,将处理好的字节数组都写入可扩容的Buffer完成编码
#com.alibaba.dubbo.rpc.protocol.dubbo.DubboInvoker#doInvoke
RpcInvocation inv = (RpcInvocation) invocation;
ResponseFuture future = currentClient.request(inv, timeout) ;
//取得结果的future放在上下文中
RpcContext.getContext().setFuture(new FutureAdapter
- 消息头
dubbo的消息头是一个定长的 16个字节。
第1-2个字节:是一个魔数数字:就是一个固定的数字
第3个字节:序列化器的id、是否是事件、是双向(有去有回) 或单向(有去无回)的标记
第4个字节:status??? (request 没有第四个字节)
第5-12个字节:请求id:long型8个字节。异步变同步的全局唯一ID,用来做consumer和provider的来回通信标记。
第13-16个字节:消息体的长度,也就是消息头+请求数据的长度。 - 消息体
data实际上是一个Invocation,标识了需要调用的方法以及参数。而且使用的buffer是可以动态扩展的
#com.alibaba.dubbo.rpc.protocol.dubbo.DubboInvoker#doInvoke
RpcInvocation inv = (RpcInvocation) invocation;
ResponseFuture future = currentClient.request(inv, timeout) ;
- 编码流程
#com.alibaba.dubbo.remoting.exchange.codec.ExchangeCodec#encodeRequest
protected void encodeRequest(Channel channel, ChannelBuffer buffer, Request req) throws IOException {
Serialization serialization = getSerialization(channel);
// header.16字节,128位
byte[] header = new byte[HEADER_LENGTH];
// set magic number.16位魔法数0xdabb
//short2bytes 将16位数存到两个byte中
Bytes.short2bytes(MAGIC, header);
// set request and serialization flag.
header[2] = (byte) (FLAG_REQUEST | serialization.getContentTypeId());
if (req.isTwoWay()) header[2] |= FLAG_TWOWAY;
if (req.isEvent()) header[2] |= FLAG_EVENT;
// set request id.
//long 64位8字节,从第5个字节
Bytes.long2bytes(req.getId(), header, 4);
//buffer这里写入buffer并不是直接写入到netty,是一个新建的buffer
//写完需要netty发送
// encode request data. 保存当前写入位置
int savedWriteIndex = buffer.writerIndex();
buffer.writerIndex(savedWriteIndex + HEADER_LENGTH);
ChannelBufferOutputStream bos = new ChannelBufferOutputStream(buffer);
ObjectOutput out = serialization.serialize(channel.getUrl(), bos);
if (req.isEvent()) {
encodeEventData(channel, out, req.getData());
} else {
//编码写入请求数据DubboCodec
encodeRequestData(channel, out, req.getData());
}
out.flushBuffer();
bos.flush();
bos.close();
//获取写入数据的长度
int len = bos.writtenBytes();
checkPayload(channel, len);
Bytes.int2bytes(len, header, 12);
// write
buffer.writerIndex(savedWriteIndex);
//写header
buffer.writeBytes(header); // write header.
//设置当前写入位置
buffer.writerIndex(savedWriteIndex + HEADER_LENGTH + len);
}
Request解码
provider接收到consumer的Request数据之后会进入Request解码流程
----------2------provider 请求解码----------------------
--NettyCodecAdapter.InternalDecoder.messageReceived
-->DubboCountCodec.decode
-->ExchangeCodec.decode
-->ExchangeCodec.decodeBody
解码流程就是接收到数据之后,先读取header长度的字节。判断开头两字节是不是magicNumber,如果不是则遍历数据的每一位来判断魔法数,交给父类来解码;是则校验header长度,数据长度,根据header的属性构造Response、Request等对象,然后读取序列化器来解析data默认使用Hessian进行序列化和反序列化
#com.alibaba.dubbo.remoting.exchange.codec.ExchangeCodec#decode(com.alibaba.dubbo.remoting.Channel, com.alibaba.dubbo.remoting.buffer.ChannelBuffer)
public Object decode(Channel channel, ChannelBuffer buffer) throws IOException {
int readable = buffer.readableBytes();
//读取header
byte[] header = new byte[Math.min(readable, HEADER_LENGTH)];
buffer.readBytes(header);
return decode(channel, buffer, readable, header);
}
protected Object decode(Channel channel, ChannelBuffer buffer, int readable, byte[] header) throws IOException {
// check magic number.
//每次解码先判断是否是magic——number
if (readable > 0 && header[0] != MAGIC_HIGH
|| readable > 1 && header[1] != MAGIC_LOW) {
//数据的起始位置不是header
int length = header.length;
if (header.length < readable) {
header = Bytes.copyOf(header, readable);
//读取完整数据
buffer.readBytes(header, length, readable - length);
}
//遍历每一位检查
for (int i = 1; i < header.length - 1; i ++) {
if (header[i] == MAGIC_HIGH && header[i + 1] == MAGIC_LOW) {
//检查到magic
buffer.readerIndex(buffer.readerIndex() - header.length + i);
header = Bytes.copyOf(header, i);
break;
}
}
return super.decode(channel, buffer, readable, header);
}
// check length.数据长度不够继续等待读取数据
if (readable < HEADER_LENGTH) {
return DecodeResult.NEED_MORE_INPUT;
}
// get data length.消息体长度
int len = Bytes.bytes2int(header, 12);
checkPayload(channel, len);
int tt = len + HEADER_LENGTH;
if( readable < tt ) {
return DecodeResult.NEED_MORE_INPUT;
}
// limit input stream.
ChannelBufferInputStream is = new ChannelBufferInputStream(buffer, len);
try {
//body解析
return decodeBody(channel, is, header);
} finally {
if (is.available() > 0) {
try {
if (logger.isWarnEnabled()) {
logger.warn("Skip input stream " + is.available());
}
StreamUtils.skipUnusedStream(is);
} catch (IOException e) {
logger.warn(e.getMessage(), e);
}
}
}
provider响应结果编码
provider调用完本地的invoker后需要返回Response,返回数据也是需要编码
#com.alibaba.dubbo.remoting.exchange.support.header.HeaderExchangeHandler#received
public void received(Channel channel, Object message) throws RemotingException {
if (message instanceof Request) {
// handle request.
Request request = (Request) message;
if (request.isEvent()) {
handlerEvent(channel, request);
} else {
if (request.isTwoWay()) {
//接收数据网络通信接收处理
Response response = handleRequest(exchangeChannel, request);
channel.send(response);
} else {
handler.received(exchangeChannel, request.getData());
}
}
} else if (message instanceof Response) {
handleResponse(channel, (Response) message);
}
}
----------3------provider响应结果编码----------------------
-->NettyCodecAdapter.InternalEncoder.encode
-->DubboCountCodec.encode
-->ExchangeCodec.encode
-->ExchangeCodec.encodeResponse
-->DubboCodec.encodeResponseData//先写入一个字节 这个字节可能是RESPONSE_NULL_VALUE RESPONSE_VALUE RESPONSE_WITH_EXCEPTION
编码协议都是一致的,和consumer请求编码有一小点不同
第3个字节:序列号组件类型,它用于和客户端约定的序列号编码号
第4个字节:它是response的结果响应码 例如 OK=20
consumer请求结果解码
----------4------consumer响应结果解码----------------------
--NettyCodecAdapter.InternalDecoder.messageReceived
-->DubboCountCodec.decode
-->ExchangeCodec.decode
-->DubboCodec.decodeBody
-->DecodeableRpcResult.decode//根据RESPONSE_NULL_VALUE RESPONSE_VALUE RESPONSE_WITH_EXCEPTION进行响应的处理
解码完成后是一个Response
解码之后
nettyServer的pipeline中注册三个处理器,前两个用来数据的编解码。在收到的数据已经完成解码工作成为了Request、Response之后就该进入nettyHandler的处理了
pipeline.addLast("decoder", adapter.getDecoder());
//编码
pipeline.addLast("encoder", adapter.getEncoder());
//逻辑处理类
pipeline.addLast("handler", nettyHandler);
#com.alibaba.dubbo.remoting.transport.netty.NettyHandler#messageReceived
@Override
public void messageReceived(ChannelHandlerContext ctx, MessageEvent e) throws Exception {
//数据接收
NettyChannel channel = NettyChannel.getOrAddChannel(ctx.getChannel(), url, handler);
try {
handler.received(channel, e.getMessage());
} finally {
NettyChannel.removeChannelIfDisconnected(ctx.getChannel());
}
}
DecodeHandler区分数据是Request还是Response,Request中的数据解析为DecodeableRpcInvocation,Response中的数据解析为DecodeableRpcResult
#com.alibaba.dubbo.remoting.transport.DecodeHandler#received
public void received(Channel channel, Object message) throws RemotingException {
if (message instanceof Decodeable) {
decode(message);
}
if (message instanceof Request) {
//如果是consumer请求则解析data为invocation
decode(((Request)message).getData());
}
if (message instanceof Response) {
//解析请求返回的response数据
decode( ((Response)message).getResult());
}
handler.received(channel, message);
}
如果是Request则寻找本地的invoker并调用并发回返回值Response,如果是Response则处理处理之前保存的future唤醒等待的线程
#com.alibaba.dubbo.remoting.exchange.support.header.HeaderExchangeHandler#received
public void received(Channel channel, Object message) throws RemotingException {
channel.setAttribute(KEY_READ_TIMESTAMP, System.currentTimeMillis());
ExchangeChannel exchangeChannel = HeaderExchangeChannel.getOrAddChannel(channel);
try {
if (message instanceof Request) {
// handle request.
Request request = (Request) message;
if (request.isEvent()) {
handlerEvent(channel, request);
} else {
if (request.isTwoWay()) {
//有返回值的请求
Response response = handleRequest(exchangeChannel, request);
channel.send(response);
} else {
handler.received(exchangeChannel, request.getData());
}
}
} else if (message instanceof Response) {
handleResponse(channel, (Response) message);
} else if (message instanceof String) {
if (isClientSide(channel)) {
Exception e = new Exception("Dubbo client can not supported string message: " + message + " in channel: " + channel + ", url: " + channel.getUrl());
logger.error(e.getMessage(), e);
} else {
String echo = handler.telnet(channel, (String) message);
if (echo != null && echo.length() > 0) {
channel.send(echo);
}
}
} else {
handler.received(exchangeChannel, message);
}
} finally {
HeaderExchangeChannel.removeChannelIfDisconnected(channel);
}
}
static void handleResponse(Channel channel, Response response) throws RemotingException {
if (response != null && !response.isHeartbeat()) {
DefaultFuture.received(channel, response);
}
}
解决粘包拆包问题
出现拆包可能没有魔法数或者数据长度不够,那就跳出循环等待下一次数据;如果出现粘包会有多个魔法数每次解码一个循环多次解码。
#com.alibaba.dubbo.rpc.protocol.dubbo.DubboCountCodec
public Object decode(Channel channel, ChannelBuffer buffer) throws IOException {
int save = buffer.readerIndex();
MultiMessage result = MultiMessage.create();
do {
Object obj = codec.decode(channel, buffer);
//出现拆包跳出循环等待下一次数据
if (Codec2.DecodeResult.NEED_MORE_INPUT == obj) {
buffer.readerIndex(save);
break;
} else {
//出现粘包会多次解析
result.addMessage(obj);
logMessageLength(obj, buffer.readerIndex() - save);
save = buffer.readerIndex();
}
} while (true);
if (result.isEmpty()) {
return Codec2.DecodeResult.NEED_MORE_INPUT;
}
if (result.size() == 1) {
return result.get(0);
}
return result;
}
总结
出现粘包拆包的核心是需要将TCP层的字节数据包切分业务层的Request、Response数据包。所以dubbo使用了魔法数和数据长度作为分割符,将Buffer中的数据隔离出来了生成Request、Response这样的一个个业务层的对象。