Re: [quicwg/base-drafts] Simplify TLP and RTO into Probe Timeout (#2114)

janaiyengar <> Thu, 13 December 2018 21:05 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 101A8130E9A for <>; Thu, 13 Dec 2018 13:05:59 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -4.46
X-Spam-Status: No, score=-4.46 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIMWL_WL_HIGH=-1.46, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=0.001, MAILING_LIST_MULTI=-1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_PASS=-0.001] autolearn=ham autolearn_force=no
Authentication-Results: (amavisd-new); dkim=pass (1024-bit key)
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id D0WgntViYM6J for <>; Thu, 13 Dec 2018 13:05:57 -0800 (PST)
Received: from ( []) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by (Postfix) with ESMTPS id D7E91130E95 for <>; Thu, 13 Dec 2018 13:05:56 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha1; c=relaxed/relaxed;; h=from:reply-to:to:cc:in-reply-to:references:subject:mime-version:content-type:content-transfer-encoding:list-id:list-archive:list-post:list-unsubscribe; s=s20150108; bh=MOzXDYfj7bFXitlz4oHblmFdWsU=; b=tz1GHJNDfqSa6ExW Qe7Vv+NAUmKPHiUM8pNoLIaaLi4AmLQgQg5eWDEUIPH7KXqr4RMUjkRyboJLyLcQ Xaf+JQzPR4Jj5OjBIZLSHg//ftZt1Rj9vyWHdTA7mYJkjOpmyR2PjZdKXYxVzc11 RDjhBopw+mB7pDZOgpKqGDI7jFA=
Received: by with SMTP id filter1573p1mdw1-6838-5C12C9B3-1D 2018-12-13 21:05:55.904270698 +0000 UTC m=+544.571465091
Received: from (unknown []) by (SG) with ESMTP id -HbO2U0bSUmz_EXID1ZbkA for <>; Thu, 13 Dec 2018 21:05:55.884 +0000 (UTC)
Received: from (localhost []) by (Postfix) with ESMTP id CBE1EC00DF for <>; Thu, 13 Dec 2018 13:05:55 -0800 (PST)
Date: Thu, 13 Dec 2018 21:05:55 +0000
From: janaiyengar <>
Reply-To: quicwg/base-drafts <>
To: quicwg/base-drafts <>
Cc: Subscribed <>
Message-ID: <quicwg/base-drafts/pull/2114/review/>
In-Reply-To: <quicwg/base-drafts/pull/>
References: <quicwg/base-drafts/pull/>
Subject: Re: [quicwg/base-drafts] Simplify TLP and RTO into Probe Timeout (#2114)
Mime-Version: 1.0
Content-Type: multipart/alternative; boundary="--==_mimepart_5c12c9b3c8f49_56233f7f410d45bc83961"; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Precedence: list
X-GitHub-Sender: janaiyengar
X-GitHub-Recipient: quic-issues
X-GitHub-Reason: subscribed
X-Auto-Response-Suppress: All
X-SG-EID: l64QuQ2uJCcEyUykJbxN122A6QRmEpucztpreh3Pak35R/3HVlcabgRUxyPDk4rWwUQSxq+Txin9Os G8r4MqHDCZMIvZX0WhInygfYR77O8oYRDTEZeHGlZ7GY89Ok9Gn9n4pCGls6fNjKV5a8zqmEO0XbNU f4qOIaBeW+VFUClTB6Atod2Beskz3UwTsiAZtmGryFXjzDIAnUrp3aby10NK7WnfoPt1fhhe0O6FNR c=
Archived-At: <>
X-Mailman-Version: 2.1.29
List-Id: Notification list for GitHub issues related to the QUIC WG <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 13 Dec 2018 21:05:59 -0000

janaiyengar commented on this pull request.

-To reduce latency, it is RECOMMENDED that the sender set and allow the TLP timer
-to fire twice before setting an RTO timer. In other words, when the TLP timer
-expires the first time, a TLP packet is sent, and it is RECOMMENDED that the TLP
-timer be scheduled for a second time. When the TLP timer expires the second
-time, a second TLP packet is sent, and an RTO timer SHOULD be scheduled {{rto}}.
+There is no requirement on the clock granularity. The PTO value MUST be set to

I've changed this, but also added some text at the definition of granularity.

-* When the final TLP packet is sent, the RTO period is set to max(SRTT +
-  4*RTTVAR + MaxAckDelay, kMinRTOTimeout)
+Consecutive PTO periods increase exponentially, and as a result, connection
+recovery latency increases exponentially as packets continue to be dropped in
+the network.  Sending two packets on PTO expiration increases resilience to
+packet drop in both directions, thus reducing the probability of consecutive PTO

I'll remove "in both directions" -- more trouble than it's worth.

-QUIC also diverges from TCP by including MaxAckDelay in the RTO period. Since
-QUIC corrects for this delay in its SRTT and RTTVAR computations, it is
-necessary to add this delay explicitly in the TLP and RTO computation.
+Delivery or loss of packets in flight is established when an acknowledgement is
+received that newly acknowledges one or more packets.  An endpoint might receive
+an acknowledgement after multiple consecutive PTO periods.

I think this is not required. There's mention of consecutive PTOs later on where it's more relevant.

> @@ -722,14 +695,13 @@ Pseudocode for OnAckReceived and UpdateRtt follow:
       // Find the smallest newly acknowledged packet
       smallest_newly_acked =
-      // If any packets sent prior to RTO were acked, then the
-      // RTO was spurious. Otherwise, inform congestion control.
-      if (rto_count > 0 &&
-            smallest_newly_acked > largest_sent_before_rto):
-        OnRetransmissionTimeoutVerified(smallest_newly_acked)
+      // If any packets sent prior to PTO were acked, then the
+      // PTO was spurious. Otherwise, inform congestion control.
+      if (pto_count > 0 &&
+            smallest_newly_acked > largest_sent_before_pto):

@martinthomson : @ianswett is trying to capture the difference in congestion response that the current loss recovery spec has for TLP vs. RTO. The idea is to say that for a couple of PTOs, there will be a smaller reduction in cwnd, and after that, the cwnd will be collapsed. I can write that up.

-## Retransmission Timeout
+A PTO expiration is classified as spurious or valid when an ACK frame is
+received that newly acknowledges packets in flight, see {{pto-loss}}.  On a
+valid PTO, the congestion window MUST be reduced to the minimum congestion
+window and slow start is re-entered.

I'm going to simplify this further. I think that having a verification is unnecessary; the behavior ends up being the same whether a PTO is verified or not. I can retain this notion of reducing the cwnd differently for the first 2 PTOs and then collapsing it to min_cwnd after that.

You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub: