This library intents to provide a **smooth, resilient, ordered, error-checked and anonymous** delivery of streams over **UDP** packets, it has been battle-tested with opensource project [kcptun](https://github.com/xtaci/kcptun). Millions of devices(from low-end MIPS routers to high-end servers) have deployed **kcp-go** powered program in a variety of forms like **online games, live broadcasting, file synchronization and network acceleration**.
1. Handles **>5K concurrent connections** on a single commodity server.
1. Compatible with [net.Conn](https://golang.org/pkg/net/#Conn) and [net.Listener](https://golang.org/pkg/net/#Listener), a drop-in replacement for [net.TCPConn](https://golang.org/pkg/net/#TCPConn).
1. [FEC(Forward Error Correction)](https://en.wikipedia.org/wiki/Forward_error_correction) Support with [Reed-Solomon Codes](https://en.wikipedia.org/wiki/Reed%E2%80%93Solomon_error_correction)
1. Packet level encryption support with [AES](https://en.wikipedia.org/wiki/Advanced_Encryption_Standard), [TEA](https://en.wikipedia.org/wiki/Tiny_Encryption_Algorithm), [3DES](https://en.wikipedia.org/wiki/Triple_DES), [Blowfish](https://en.wikipedia.org/wiki/Blowfish_(cipher)), [Cast5](https://en.wikipedia.org/wiki/CAST-128), [Salsa20]( https://en.wikipedia.org/wiki/Salsa20), etc. in [CFB](https://en.wikipedia.org/wiki/Block_cipher_mode_of_operation#Cipher_Feedback_.28CFB.29) mode, which generates completely anonymous packet.
1. Only **A fixed number of goroutines** will be created for the entire server application, costs in **context switch** between goroutines have been taken into consideration.
1. Compatible with [skywind3000's](https://github.com/skywind3000) C version with various improvements.
List structure introduces **heavy cache misses** compared to slice which owns better **locality**, 5000 connections with 32 window size and 20ms interval will cost 6us/0.03%(cpu) using slice, and 8.7ms/43.5%(cpu) for list for each `kcp.flush()`.
Timing is **critical** to **RTT estimator**, inaccurate timing leads to false retransmissions in KCP, but calling `time.Now()` costs 42 cycles(10.5ns on 4GHz CPU, 15.6ns on my MacBook Pro 2.7GHz).
In kcp-go, after each `kcp.output()` function call, current clock time will be updated upon return, and for a single `kcp.flush()` operation, current time will be queried from system once. For most of the time, 5000 connections costs 5000 * 15.6ns = 78us(a fixed cost while no packet needs to be sent), as for 10MB/s data transfering with 1400 MTU, `kcp.output()` will be called around 7500 times and costs 117us for `time.Now()` in **every second**.
Control messages like **SYN/FIN/RST** in TCP **are not defined** in KCP, you need some **keepalive/heartbeat mechanism** in the application-level. A real world example is to use some **multiplexing** protocol over session, such as [smux](https://github.com/xtaci/smux)(with embedded keepalive mechanism), see [kcptun](https://github.com/xtaci/kcptun) for example.
A: A standalone `agent` or `gate` server for running kcp-go is suggested, not only for CPU utilization, but also important to the **precision** of RTT measurements(timing) which indirectly affects retransmission. By increasing update `interval` with `SetNoDelay` like `conn.SetNoDelay(1, 40, 1, 1)` will dramatically reduce system load, but lower the performance.