注:请转载后注明出处
Reduced latency in the face of packet loss (goal 3), as compared with TCP, may potentially
be achieved by predominantly using error correction, rather than retransmission. Here
again, such an extension of TCP did not appear plausible, without major (slow to deploy,
perhaps over 10 years or more) modifications of standard TCP. We see wide deployment
of this protocol as being plausible within 1-2 years, and would hope that TCP might evolve
similarly over a much longer time period.
【针对丢失数据包降低时延(目标3),相比于TCP,要达到这个方法可可能潜在地通过使用错误纠正而不是重传,再次重申,这种对TCP的扩展在不修改标准的TCP的前提下显得并不乐观(因为要缓慢地不熟可能需要10年或者更久),在我们看来,这个协议的部署乐观地在1~2年内可以完成,希望TCP的进化可能需要类似更长一段时间】
With the rise of mobile clients, and their tendency to turn off a radio communications,
minimal roundtrip latency costs (goal 4) for session continuations (resumptions?) will be
progressively critical. Already today, desktop users benefit greatly from more rapid
session continuations, and we’re hoping that QUIC will take this to the ultimate goal of zero
RTT for session resumption, including cryptographic negotiations (most of the time).
【随着手机客户端的逐渐增多,并且他们趋向于关闭无线通信,因此减少(往返的延时消耗(目标4)对该会话的延续或恢复将是重要的。今天桌面客户端受益于更快的会话速度,因此我们也希望QUIC可以带来极限的目标,0RTT的会话恢复,包括加密协商(大多数如此))】
Congestion avoidance algorithms are critical to protecting the Internet. To the greatest
extent possible we’ll provide a TCP friendly and compatible protocol, to coexist fairly with
TCP’s highly evolved and universally deployed congestion avoidance approaches. We
anticipate plausible improvements in agility and notification, courtesy of having a multitude
of streams adjusted (re: congestion window?) in unison based on overall losses in all
multiplexed streams.
【对于保护英特网,拥塞避免算法是非常关键的,为了达到更好的效果,我们将提供一个TCP兼容、有好的协议,用于与TCP的广泛部署的拥塞避免的处理兼容。我们预期到在敏捷性,通知上,以及基于在所有多路复用的码流中所有的丢包和谐地在大多数码流的调整上, 都会有乐观的改进】
Experience with SPDY development has taught us that the only way to prevent
middleboxes from maligning a new protocol built atop UDP or TCP (e.g., misconstruing it
for a “known” protocol, and making “less than helpful” changes), is to encrypt as much of the
payload and control structure as feasible. As a result we plan to employ security elements
including tamper resistance, privacy, replay protection, and authentication, that is similar to
what is provided via TLS (perhaps DTLS like).
【在经历过SPDY的开发后,告诉我们,对于阻止中间设备丢掉、破环基于UDP和TCP的新协议上(例如,误解成“未知的”协议和贯彻“毫无帮助”的改变),对数据加密和灵活地控制结构是唯一的方法,因此我们计划部署些安全的元素,包括:稳定性,因此和中继保护,鉴权。这个跟TLS很像】