CS144-Lab3-TCPSender

lab 地址 :lab3-doc
代码实现:lab3-code
完整目录:
0. ByteStream
1. StreamReassembler
2. TCPReceiver
3. TCPSender
4. TCPConnection
5. ARP
6. IP-Router

1. 目标

lab2 实现完 TCPReceiver 后,需要实现 TCPSender,相对于前者,Sender 要实现的功能相对更复杂一些,需要支持:

  • 根据 Sender 当前的状态对可发送窗口进行填充,发包
  • Sender 需要根据对方通知的窗口大小和 ackno 来确认对方当前收到的字节流进度
  • 需要支持超时重传机制,根据时间变化(RTO),定时重传那些还没有 ack 的报文

2. 实现

2.1 窗口填充

这里主要根据 lab 中提供的状态说明图进行实现:

image

理论上只有 3 种情况需要处理:

  • 初始化,需要发送 SYN 包,即 CLOSED 状态,此刻需要填充 Syn 包
  • ByteStream 中还有数据可写入发送,且对方窗口大小足够,此时正常按照 payload 大小限制填充数据包
  • ByteStream 已经 eof,但是 Fin 包还未发送,此刻需要填充 Fin 包
    为了配合重传机制的实现,当填充成功一个数据包的同时,也需要对其进行缓存备份,代码如下:
void TCPSender::fill_window() {
	// CLOSED
	if (next_seqno_absolute() == 0) {
		// need to send the syn
		TCPSegment seg;
		seg.header().syn = true;
		seg.header().seqno = _isn;
		_segments_out.push(seg);
		_next_seqno += seg.length_in_sequence_space();
		_bytes_in_flight += seg.length_in_sequence_space();
		// cache the segments
		_cache_segment(next_seqno_absolute() , seg);
	}
	
	// SYN_ACKED
	if (not stream_in().eof() && next_seqno_absolute() > bytes_in_flight()) {
		// if windows size equal zero , the fill window method should act like the window size is one
		size_t max_seg_size = TCPConfig::MAX_PAYLOAD_SIZE;
		size_t remaining_wsz = _peer_windows_size ? _peer_windows_size : 1;
		size_t flight_size = bytes_in_flight();
		if (remaining_wsz < flight_size) {
			return;
		}
		remaining_wsz -= flight_size;
		string read_stream = _stream.read(min(remaining_wsz, max_seg_size));
		while (!read_stream.empty()) {
			TCPSegment seg;
			seg.header().seqno = next_seqno();
			seg.payload() = Buffer(std::move(read_stream));
			remaining_wsz -= seg.length_in_sequence_space();
			if (stream_in().eof() && next_seqno_absolute() < stream_in().bytes_written() + 2 && remaining_wsz >= 1) {
				seg.header().fin = true;
				remaining_wsz -= 1;
			}
			_next_seqno += seg.length_in_sequence_space();
			_bytes_in_flight += seg.length_in_sequence_space();
			_segments_out.push(seg);
			// cache the seg
			_cache_segment(next_seqno_absolute(), seg);
			read_stream = _stream.read(min(remaining_wsz, max_seg_size));
		}
	}
	
	// SYN_ACKED
	if (stream_in().eof() && next_seqno_absolute() < stream_in().bytes_written() + 2) {
		// stream has reached EOF, but FIN flag hasn't been sent yet. so send the fin
		size_t remaining_wsz = _peer_windows_size ? _peer_windows_size : 1;
		size_t flight_size = bytes_in_flight();
		if (remaining_wsz <= flight_size) {
			return;
		}
		
		TCPSegment seg;
		seg.header().seqno = next_seqno();
		seg.header().fin = true;
		_next_seqno += seg.length_in_sequence_space();
		_bytes_in_flight += seg.length_in_sequence_space();
		_segments_out.push(seg);
		// cache the segments
		_cache_segment(next_seqno_absolute(), seg);
	}
}

2.2 定时重传

每次发送数据包的会对其进行缓存,当超过 RTO 时间没有收到这个包的确认时,就需要执行重传,更新等待时间为 当前 RTO * 2,假如有多个数据包超时,每次只会重传 seqno 最小的数据包。

//! \param[in] ms_since_last_tick the number of milliseconds since the last call to this method
void TCPSender::tick(const size_t ms_since_last_tick) {
	if (_segments_cache.empty()) {
		return;
	}
	_time += ms_since_last_tick;
	if (_time >= _initial_retransmission_timeout && _retx_times <= TCPConfig::MAX_RETX_ATTEMPTS) {
		_segments_out.push(_segments_cache.begin()->second);
		_time = 0;
		if (_peer_windows_size != 0) {
			_initial_retransmission_timeout *= 2;
			_retx_times += 1;
		}
	}
}

需要注意的是,只有当对方通知的窗口大小不为 0 的时候,才更新等待时间为 RTO * 2。

2.3 Ack 确认

Ack 确认的逻辑主要有几点:

  • 如果有有效确认(即缓存的未确认数据包有被成功确认的情况),则需要:
    • 更新重传机制相关变量(重传时间,重传次数)
    • 删除确认成功数据包的缓存
    • 填充窗口,因为对端成功确认了数据,对端可用窗口变大了
  • 更新对端窗口大小
//! \param ackno The remote receiver's ackno (acknowledgment number)
//! \param window_size The remote receiver's advertised window size
void TCPSender::ack_received(const WrappingInt32 ackno, const uint16_t window_size) {
	int left_need_ack_bytes = next_seqno() - ackno;
	if (left_need_ack_bytes < 0) {
		return;
	}
	// pop the cache segments;
	vector<uint64_t> remove_acknos;
	for (auto & seg : _segments_cache) {
		uint64_t abs_ackno = unwrap(ackno, _isn, seg.first);
		if (seg.first <= abs_ackno) {
			remove_acknos.push_back(seg.first);
		}
	}
	for (auto & remove_ackno : remove_acknos) {
		_segments_cache.erase(remove_ackno);
	}
	_peer_windows_size = window_size;
	// valid ackno
	if (!remove_acknos.empty()) {
		_bytes_in_flight = left_need_ack_bytes;
		_time = 0;
		_retx_times = 0;
		_initial_retransmission_timeout = _default_initial_retransmission_timeout;
		fill_window();
	}
}

3. 测试

image

posted @ 2023-01-25 10:24  lawliet9  阅读(207)  评论(0编辑  收藏  举报