Two flavors each of UDP inbound and outbound channel adapters are provided UnicastSendingMessageHandler sends a datagram packet to a single destination. UnicastReceivingChannelAdapter receives incoming datagram packets. MulticastSendingMessageHandler sends (broadcasts) datagram packets to a multicast address. MulticastReceivingChannelAdapter receives incoming datagram packets by joining to a multicast address.

TCP inbound and outbound channel adapters are provided TcpSendingMessageHandler sends messages over TCP. TcpReceivingChannelAdapter receives messages over TCP.

An inbound TCP gateway is provided; this allows for simple request/response processing. While the gateway can support any number of connections, each connection can only process serially. The thread that reads from the socket waits for, and sends, the response before reading again. If the connection factory is configured for single use connections, the connection is closed after the socket times out.

An outbound TCP gateway is provided; this allows for simple request/response processing. If the associated connection factory is configured for single use connections, a new connection is immediately created for each new request. Otherwise, if the connection is in use, the calling thread blocks on the connection until either a response is received or a timeout or I/O error occurs.

The TCP and UDP inbound channel adapters, and the TCP inbound gateway, support the "error-channel" attribute. This provides the same basic functionality as described in Section 7.2.1, “Enter the GatewayProxyFactoryBean”.

 <int-ip:udp-outbound-channel-adapter id="udpOut"
    host="somehost"
    port="11111"
    multicast="false"
    channel="exampleChannel" />

A simple UDP outbound channel adapter.

UDP is an efficient, but unreliable protocol. Two attributes are added to improve reliability. When check-length is set to true, the adapter precedes the message data with a length field (4 bytes in network byte order). This enables the receiving side to verify the length of the packet received. If a receiving system uses a buffer that is too short the contain the packet, the packet can be truncated. The length header provides a mechanism to detect this.

  <int-ip:udp-outbound-channel-adapter id="udpOut"
    host="somehost"
    port="11111"
    multicast="false"
    check-length="true"
    channel="exampleChannel" />

An outbound channel adapter that adds length checking to the datagram packets.

The second reliability improvement allows an application-level acknowledgment protocol to be used. The receiver must send an acknowledgment to the sender within a specified time.

  <int-ip:udp-outbound-channel-adapter id="udpOut"
    host="somehost"
    port="11111"
    multicast="false"
    check-length="true"
    acknowledge="true"
    ack-host="thishost"
    ack-port="22222"
    ack-timeout="10000"
    channel="exampleChannel" />

An outbound channel adapter that adds length checking to the datagram packets and waits for an acknowledgment.

For even more reliable networking, TCP can be used.

  <int-ip:udp-inbound-channel-adapter id="udpReceiver"
    channel="udpOutChannel"
    port="11111"
    receive-buffer-size="500"
    multicast="false"
    check-length="true" />

A basic unicast inbound udp channel adapter.

  <int-ip:udp-inbound-channel-adapter id="udpReceiver"
    channel="udpOutChannel"
    port="11111"
    receive-buffer-size="500"
    multicast="true"
    multicast-address="225.6.7.8"
    check-length="true" />

A basic multicast inbound udp channel adapter.

By default, reverse DNS lookups are done on inbound packets to convert IP addresses to hostnames for use in message headers. In environments where DNS is not configured, this can cause delays. This default behavior can be overridden by setting the lookup-host attribute to "false".

For TCP, the configuration of the underlying connection is provided using a Connection Factory. Two types of connection factory are provided; a client connection factory and a server connection factory. Client connection factories are used to establish outgoing connections; Server connection factories listen for incoming connections.

A client connection factory is used by an outbound channel adapter but a reference to a client connection factory can also be provided to an inbound channel adapter and that adapter will receive any incoming messages received on connections created by the outbound adapter.

A server connection factory is used by an inbound channel adapter or gateway (in fact the connection factory will not function without one). A reference to a server connection factory can also be provided to an outbound adapter; that adapter can then be used to send replies to incoming messages to the same connection.

A maximum of one adapter of each type may be given a reference to a connection factory.

Connection factories using java.net.Socket and java.nio.channel.SocketChannel are provided.

  <int-ip:tcp-connection-factory id="server"
    type="server"
    port="1234"
  />

A simple server connection factory that uses java.net.Socket connections.

  <int-ip:tcp-connection-factory id="server"
    type="server"
    port="1234"
    using-nio="true"
  />

A simple server connection factory that uses java.nio.channel.SocketChannel connections.

  <int-ip:tcp-connection-factory id="client"
    type="client"
    host="localhost"
    port="1234"
    single-use="true"
    so-timeout="10000"
  />

A client connection factory that uses java.net.Socket connections and creates a new connection for each message.

  <int-ip:tcp-connection-factory id="client"
    type="client"
    host="localhost"
    port="1234"
    single-use="true"
    so-timeout="10000"
    using-nio=true
  />

A client connection factory that uses java.nio.channel.Socket connections and creates a new connection for each message.

TCP is a streaming protocol; this means that some structure has to be provided to data transported over TCP, so the receiver can demarcate the data into discrete messages. Connection factories are configured to use (de)serializers to convert between the message payload and the bits that are sent over TCP. This is accomplished by providing a deserializer and serializer for inbound and outbound messages respectively. A number of standard (de)serializers are provided.

The ByteArrayCrlfSerializer, converts a byte array to a stream of bytes followed by carriage return and linefeed characters (\r\n). This is the default (de)serializer and can be used with telnet as a client, for example.

The ByteArrayStxEtxSerializer, converts a byte array to a stream of bytes preceded by an STX (0x02) and followed by an ETX (0x03).

The ByteArrayLengthHeaderSerializer, converts a byte array to a stream of bytes preceded by a binary length in network byte order (big endian). This a very efficient deserializer because it does not have to parse every byte looking for a termination character sequence. It can also be used for payloads containing binary data; the above serializers only support text in the payload. The default size of the length header is 4 bytes (Integer), allowing for messages up to 2**31-1 bytes. However, the length header can be a single byte (unsigned) for messages up to 255 bytes, or an unsigned short (2 bytes) for messages up to 2**16 bytes. If you need any other format for the header, you can subclass this class and provide implementations for the readHeader and writeHeader methods. The absolute maximum data size supported is 2**31-1 bytes.

The ByteArrayRawSerializer, converts a byte array to a stream of bytes and adds no additional message demarcation data; with this (de)serializer, the end of a message is indicated by the client closing the socket in an orderly fashion. When using this serializer, message reception will hang until the client closes the socket, or a timeout occurs; a timeout will NOT result in a message. When this serializer is being used, and the client is a Spring Integration application, the client must use a connection factory that is configured with single-use=true - this causes the adapter to close the socket after sending the message; the serializer will not, itself, close the connection. This serializer should only be used with connection factories used by channel adapters (not gateways), and the connection factories should be used by either an inbound or outbound adapter, and not both.

Each of these is a subclass of AbstractByteArraySerializer which implements both org.springframework.core.serializer.Serializer and org.springframework.core.serializer.Deserializer. For backwards compatibility, connections using any subclass of AbstractByteArraySerializer for serialization will also accept a String which will be converted to a byte array first. Each of these (de)serializers converts an input stream containing the corresponding format to a byte array payload.

To avoid memory exhaustion due to a badly behaved client (one that does not adhere to the protocol of the configured serializer), these serializers impose a maximum message size. If the size is exceeded by an incoming message, an exception will be thrown. The default maximum message size is 2048 bytes, and can be increased by setting the maxMessageSize property. If you are using the default (de)serializer and wish to increase the maximum message size, you must declare it as an explicit bean with the property set and configure the connection factory to use that bean.

The final standard serializer is org.springframework.core.serializer.DefaultSerializer which can be used to convert Serializable objects using java serialization. org.springframework.core.serializer.DefaultDeserializer is provided for inbound deserialization of streams containing Serializable objects.

To implement a custom (de)serializer pair, implement the org.springframework.core.serializer.Deserializer and org.springframework.core.serializer.Serializer interfaces.

If you do not wish to use the default (de)serializer (ByteArrayCrLfSerializer), you must supply serializer and deserializer attributes on the connection factory (example below).

  <bean id="javaSerializer"
        class="org.springframework.core.serializer.DefaultSerializer" />
  <bean id="javaDeserializer"
        class="org.springframework.core.serializer.DefaultDeserializer" />

  <int-ip:tcp-connection-factory id="server"
    type="server"
    port="1234"
    deserializer="javaDeserializer"
    serializer="javaSerializer"
  />

A server connection factory that uses java.net.Socket connections and uses Java serialization on the wire.

For full details of the attributes available on connection factories, see the reference at the end of this section.

By default, reverse DNS lookups are done on inbound packets to convert IP addresses to hostnames for use in message headers. In environments where DNS is not configured, this can cause connection delays. This default behavior can be overridden by setting the lookup-host attribute to "false".

Connection factories can be configured with a reference to a TcpConnectionInterceptorFactoryChain. Interceptors can be used to add behavior to connections, such as negotiation, security, and other setup. No interceptors are currently provided by the framework but, for an example, see the InterceptedSharedConnectionTests in the source repository.

The HelloWorldInterceptor used in the test case works as follows:

When configured with a client connection factory, when the first message is sent over a connection that is intercepted, the interceptor sends 'Hello' over the connection, and expects to receive 'world!'. When that occurs, the negotiation is complete and the original message is sent; further messages that use the same connection are sent without any additional negotiation.

When configured with a server connection factory, the interceptor requires the first message to be 'Hello' and, if it is, returns 'world!'. Otherwise it throws an exception causing the connection to be closed.

All TcpConnection methods are intercepted. Interceptor instances are created for each connection by an interceptor factory. If an interceptor is stateful, the factory should create a new instance for each connection. Interceptor factories are added to the configuration of an interceptor factory chain, which is provided to a connection factory using the interceptor-factory attribute. Interceptors must implement the TcpConnectionInterceptor interface; factories must implement the TcpConnectionInterceptorFactory interface. A convenience class AbstractTcpConnectionInterceptor is provided with passthrough methods; by extending this class, you only need to implement those methods you wish to intercept.

<bean id="helloWorldInterceptorFactory"
   class="org.springframework.integration.ip.tcp.connection.TcpConnectionInterceptorFactoryChain">
  <property name="interceptors">
    <array>
      <bean class="org.springframework.integration.ip.tcp.connection.HelloWorldInterceptorFactory"/>
    </array>
  </property>
</bean>

<int-ip:tcp-connection-factory id="server"
  type="server"
  port="12345"
  using-nio="true"
  single-use="true"
  interceptor-factory-chain="helloWorldInterceptorFactory"
/>

<int-ip:tcp-connection-factory id="client"
  type="client"
  host="localhost"
  port="12345"
  single-use="true"
  so-timeout="100000"
  using-nio="true"
  interceptor-factory-chain="helloWorldInterceptorFactory"
/>

Configuring a connection interceptor factory chain.

TCP inbound and outbound channel adapters that utilize the above connection factories are provided. These adapters have attributes connection-factory and channel. The channel attribute specifies the channel on which messages arrive at an outbound adapter and on which messages are placed by an inbound adapter. The connection-factory attribute indicates which connection factory is to be used to manage connections for the adapter. While both inbound and outbound adapters can share a connection factory, server connection factories are always 'owned' by an inbound adapter; client connection factories are always 'owned' by an outbound adapter. One, and only one, adapter of each type may get a reference to a connection factory.

  <bean id="javaSerializer"
              class="org.springframework.core.serializer.DefaultSerializer" />
  <bean id="javaDeserializer"
              class="org.springframework.core.serializer.DefaultDeserializer" />

  <int-ip:tcp-connection-factory id="server"
    type="server"
    port="1234"
    deserializer="javaDeserializer"
    serializer="javaSerializer"
    using-nio="true"
    single-use="true"
  />

  <int-ip:tcp-connection-factory id="client"
    type="client"
    host="localhost"
    port="#{server.port}"
    single-use="true"
    so-timeout="10000"
    deserializer="javaDeserializer"
    serializer="javaSerializer"
  />

  <int:channel id="input" />

  <int:channel id="replies">
    <int:queue/>
  </int:channel>

  <int-ip:tcp-outbound-channel-adapter id="outboundClient"
    channel="input"
    connection-factory="client"/>

  <int-ip:tcp-inbound-channel-adapter id="inboundClient"
    channel="replies"
    connection-factory="client"/>

  <int-ip:tcp-inbound-channel-adapter id="inboundServer"
    channel="loop"
    connection-factory="server"/>

  <int-ip:tcp-outbound-channel-adapter id="outboundServer"
    channel="loop"
    connection-factory="server"/>

  <int:channel id="loop" />

In this configuration, messages arriving in channel 'input' are serialized over connections created by 'client' received at the server and placed on channel 'loop'. Since 'loop' is the input channel for 'outboundServer' the message is simply looped back over the same connection and received by 'inboundClient' and deposited in channel 'replies'. Java serialization is used on the wire.

Normally, inbound adapters use a type="server" connection factory, which listens for incoming connection requests. In some cases, it is desireable to establish the connection in reverse, whereby the inbound adapter connects to an external server and then waits for inbound messages on that connection.

This topology is supported by using client-mode="true" on the inbound adapter. In this case, the connection factory must be of type 'client' and must have single-use set to false.

Two additional attributes are used to support this mechanism: retry-interval specifies (in milliseconds) how often the framework will attempt to reconnect after a connection failure. scheduler is used to supply a TaskScheduler used to schedule the connection attempts, and to test that the connection is still active.

For an outbound adapter, the connection is normally established when the first message is sent. client-mode="true" on an outbound adapter will cause the connection to be established when the adapter is started. Adapters are automatically started by default. Again, the connection factory must be of type client and have single-use set to false and retry-interval and scheduler are also supported. If a connection fails, it will be re-established either by the scheduler or when the next message is sent.

For both inbound and outbound, if the adapter is started, you may force the adapter to establish a connection by sending a <control-bus /> command: @adapter_id.retryConnection() and examine the current state with @adapter_id.isConnected().

The inbound TCP gateway TcpInboundGateway and outbound TCP gateway TcpOutboundGateway use a server and client connection factory respectively. Each connection can process a single request/response at a time.

The inbound gateway, after constructing a message with the incoming payload and sending it to the requestChannel, waits for a response and sends the payload from the response message by writing it to the connection.

As with inbound adapters, inbound gateways normally use a type="server" connection factory, which listens for incoming connection requests. In some cases, it is desireable to establish the connection in reverse, whereby the inbound gateway connects to an external server and then waits for, and replies to, inbound messages on that connection.

This topology is supported by using client-mode="true" on the inbound gateway. In this case, the connection factory must be of type 'client' and must have single-use set to false.

Two additional attributes are used to support this mechanism: retry-interval specifies (in milliseconds) how often the framework will attempt to reconnect after a connection failure. scheduler is used to supply a TaskScheduler used to schedule the connection attempts, and to test that the connection is still active.

If the gateway is started, you may force the gateway to establish a connection by sending a <control-bus /> command: @adapter_id.retryConnection() and examine the current state with @adapter_id.isConnected().

The outbound gateway, after sending a message over the connection, waits for a response and constructs a response message and puts in on the reply channel. Communications over the connections are single-threaded. Users should be aware that only one message can be handled at a time and, if another thread attempts to send a message before the current response has been received, it will block until any previous requests are complete (or time out). If, however, the client connection factory is configured for single-use connections each new request gets its own connection and is processed immediately.

  <int-ip:tcp-inbound-gateway id="inGateway"
    request-channel="tcpChannel"
    reply-channel="replyChannel"
    connection-factory="cfServer"
    reply-timeout="10000"
        />

A simple inbound TCP gateway; if a connection factory configured with the default (de)serializer is used, messages will be \r\n delimited data and the gateway can be used by a simple client such as telnet.

  <int-ip:tcp-outbound-gateway id="outGateway"
    request-channel="tcpChannel"
    reply-channel="replyChannel"
    connection-factory="cfClient"
    request-timeout="10000"
    remote-timeout="10000"
    />

A simple outbound TCP gateway.

Using NIO (see using-nio in Section 16.10, “IP Configuration Attributes”) avoids dedicating a thread to read from each socket. For a small number of sockets, you will likely find that not using NIO, together with an async handoff (e.g. to a QueueChannel), will perform as well as, or better than, using NIO.

Consider using NIO when handling a large number of connections. However, the use of NIO has some other ramifications. A pool of threads (in the task executor) is shared across all the sockets; each incoming message is assembled and sent to the configured channel as a separate unit of work on a thread selected from that pool. Two sequential messages arriving on the same socket might be processed by different threads. This means that the order in which the messages are sent to the channel is indeterminate; the strict ordering of the messages arriving on the socket is not maintained.

For some applications, this is not an issue; for others it is. If strict ordering is required, consider setting using-nio to false and using async handoff.

Alternatively, you may choose to insert a resequencer downstream of the inbound endpoint to return the messages to their proper sequence. Set apply-sequence to true on the connection factory, and messages arriving on a TCP connection will have sequenceNumber and correlationId headers set. The resequencer uses these headers to return the messages to their proper sequence.

Pool Size

The pool size attribute is no longer used; previously, it specified the size of the default thread pool when a task-executor was not specified. It was also used to set the connection backlog on server sockets. The first function is no longer needed (see below); the second function is replaced by the backlog attribute.

Previously, when using a fixed thread pool task executor (which was the default), with NIO, it was possible to get a deadlock and processing would stop. The problem occurred when a buffer was full, a thread reading from the socket was trying to add more data to the buffer, and there were no threads available to make space in the buffer. This only occurred with a very small pool size, but it could be possible under extreme conditions. Since 2.2, two changes have eliminated this problem. First, the default task executor is a cached thread pool executor. Second, deadlock detection logic has been added such that if thread starvation occurs, instead of deadlocking, an exception is thrown, thus releasing the deadlocked resources.