Reactive programming has a important concept of stream which allow application scale vertically and apply good multi-thread dispatch for concurrent event.Network programming is designed as many layer for different data types. It's nature and useful if use Reactive Style to implement network data stream.At upper of TCP, the library complete a basic function of Session Layer and Json Presentation Layer.

From v0.12.0, RxSocket only support Scala 2.12.x and 2.11.x

"com.scalachan" %% "rxsocket" % "0.13.1"

• do logic with reactive style with Observable and Future
• with RPC stream can get a pipe-like socket communicate.
• Asynchronous & Non-blocking
• Json based Server mode make it easy to build TCP service.
• consume data with back-pressure
• custom protocol parser which could be active mode or passive mode

• simple example in test/demo directory
• simple benchmark in test/benchmark directory

• support heartbeat
  • timing wheel
• priority scheduler thread pool
  • avoid request callback re-enqueue to thread pool(forkjoinpool now) caused timeout.
  • allow handler massive request and give a slight performance affect
    • give a space to monitor network request status
• better benchmark