| Internet-Draft | DCTCP | October 2020 |
| Bensley, et al. | Expires 4 April 2021 | [Page] |
This memo describes Datacenter TCP (DCTCP), an improvement to TCP congestion control for datacenter traffic. DCTCP uses improved Explicit Congestion Notification (ECN) processing to estimate the fraction of bytes that encounter congestion, rather than simply detecting that some congestion has occurred. DCTCP then scales the TCP congestion window based on this estimate. This method achieves high burst tolerance, low latency, and high throughput with shallow-buffered switches.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 4 April 2021.¶
Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
Large datacenters necessarily need a large number of network switches to interconnect the servers in the datacenter. Therefore, a datacenter can greatly reduce its capital expenditure by leveraging low cost switches. However, low cost switches tend to have limited queue capacities and thus are more susceptible to packet loss due to congestion.¶
Network traffic in the datacenter is often a mix of short and long flows, where the short flows require low latency and the long flows require high throughput. Datacenters also experience incast bursts, where many endpoints send traffic to a single server at the same time. For example, this is a natural consequence of MapReduce algorithms. The worker nodes complete at approximately the same time, and all reply to the master node concurrently.¶
These factors place some conflicting demands on the queue occupancy of a switch:¶
Standard TCP congestion control [RFC5681] relies on segment loss to detect congestion. This does not meet the demands described above. First, the short flows will start to experience unacceptable latencies before packet loss occurs. Second, by the time TCP congestion control kicks in on the sender, most of the incast burst has already been dropped.¶
[RFC3168] describes a mechanism for using Explicit Congestion Notification (ECN) from the switch for early detection of congestion, rather than waiting for segment loss to occur. However, this method only detects the presence of congestion, not the extent. In the presence of mild congestion, the TCP congestion window is reduced too aggressively and unnecessarily affects the throughput of long flows.¶
Datacenter TCP (DCTCP) improvises upon traditional ECN processing by estimating the fraction of bytes that encounter congestion, rather than simply detecting that some congestion has occurred. DCTCP then scales the TCP congestion window based on this estimate. This method achieves high burst tolerance, low latency, and high throughput with shallow-buffered switches.¶
It is recommended that DCTCP be deployed in a datacenter environment where the endpoints and the switching fabric are under a single administrative domain. Deployment issues around coexistence of DCTCP and conventional TCP, and lack of a negotiating mechanism between sender and receiver, and possible mitigations are also discussed.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].¶
There are three components involved in the DCTCP algorithm:¶
The switch indicates congestion to the end nodes by setting the CE codepoint in the IP header as specified in Section 5 of [RFC3168]. For example, the switch may be configured with a congestion threshold. When a packet arrives at the switch and its queue length is greater than the congestion threshold, the switch sets the CE codepoint in the packet. For example, Section 3.4 of [DCTCP10] suggests threshold marking with a threshold K > (RTT * C)/7, where C is the sending rate in packets per second. However, the actual algorithm for marking congestion is an implementation detail of the switch and will generally not be known to the sender and receiver. Therefore, sender and receiver MUST NOT assume that a particular marking algorithm is implemented by the switching fabric.¶
According to Section 6.1.3 of [RFC3168], the receiver sets the ECE flag if any of the packets being acknowledged had the CE code point set. The receiver then continues to set the ECE flag until it receives a packet with the Congestion Window Reduced (CWR) flag set. However, the DCTCP algorithm requires more detailed congestion information. In particular, the sender must be able to determine the number of sent bytes that encountered congestion. Thus, the scheme described in [RFC3168] does not suffice.¶
One possible solution is to ACK every packet and set the ECE flag in the ACK if and only if the CE code point was set in the packet being acknowledged. However, this prevents the use of delayed ACKs, which are an important performance optimization in datacenters.¶
Instead, DCTCP introduces a new Boolean TCP state variable, DCTCP Congestion Encountered (DCTCP.CE), which is initialized to false and stored in the Transmission Control Block (TCB). When sending an ACK, the ECE flag MUST be set if and only if DCTCP.CE is true. When receiving packets, the CE codepoint MUST be processed as follows:¶
The sender estimates the fraction of sent bytes that encountered congestion. The current estimate is stored in a new TCP state variable, DCTCP.Alpha, which is initialized to 1 and MUST be updated as follows:¶
where¶
In order to update DCTCP.Alpha, the TCP state variables defined in [RFC0793] are used, and three additional TCP state variables are introduced:¶
The congestion estimator on the sender MUST process acceptable ACKs as follows:¶
Compute the bytes acknowledged (TCP SACK options [RFC2018] are ignored):¶
Update the bytes sent:¶
If the ECE flag is set, update the bytes marked:¶
Compute the congestion level for the current observation window:¶
Update the congestion estimate:¶
Determine the end of the next observation window:¶
Reset the byte counters:¶
Rather than always halving the congestion window as described in [RFC3168], when the sender receives an indication of congestion (ECE), the sender MUST update cwnd as follows:¶
Thus, when no sent byte experienced congestion, DCTCP.Alpha equals zero, and cwnd is left unchanged. When all sent bytes experienced congestion, DCTCP.Alpha equals one, and cwnd is reduced by half. Lower levels of congestion will result in correspondingly smaller reductions to cwnd.¶
Just as specified in [RFC3168], TCP should not react to congestion indications more than once every window of data. The setting of the "Congestion Window Reduced" (CWR) bit is also exactly as per [RFC3168].¶
[RFC3168] requires that compliant TCP MUST NOT set ECT on SYN or SYN-ACK packets. [RFC5562] proposes setting ECT on SYN-ACK packets, but maintains the restriction of no ECT on SYN packets. Both these RFCs prohibit ECT in SYN packets due to security concerns regarding malicious SYN packets with ECT set. These RFCs, however, are intended for general Internet use, and do not directly apply to a controlled datacenter deployment. The switching fabric can drop TCP packets that do not have the ECT set in the IP header. If SYN and SYN-ACK packets for DCTCP connections are non-ECT they will be dropped with high probability. For DCTCP connections the sender SHOULD set ECT for SYN, SYN-ACK and RST packets.¶
As noted in Section 3.3, the implementation must choose a suitable estimation gain. [DCTCP10] provides a theoretical basis for selecting the gain. However, it may be more practical to use experimentation to select a suitable gain for a particular network and workload. The Microsoft implementation of DCTCP in Windows Server 2012 uses a fixed estimation gain of 1/16.¶
The implementation must also decide when to use DCTCP. Datacenter servers may need to communicate with endpoints outside the datacenter, where DCTCP is unsuitable or unsupported. Thus, a global configuration setting to enable DCTCP will generally not suffice. DCTCP may be configured based on the IP address of the remote endpoint. Microsoft Windows Server 2012 also supports automatic selection of DCTCP if the estimated RTT is less than 10 msec and ECN is successfully negotiated, under the assumption that if the RTT is low, then the two endpoints are likely on the same datacenter network.¶
To prevent incast throughput collapse the minimum RTO (MinRTO) used by TCP should be lowered significantly. The default value of MinRTO in Windows is 300 msec which is much greater than the maximum latencies inside a datacenter. Server 2012 onwards the MinRTO value is configurable allowing values as low as 10 msec on a per subnet or per TCP port basis or even globally. A lower MinRTO value requires corresponding a lower delayed ACK timeout on the receiver. It is recommended that the implementation allow configuration of lower timeouts for DCTCP connections.¶
In the same vein, it is also recommended that the implementation allow configuration of restarting the cwnd of idle DCTCP connections as described in [RFC5681] since network conditions change rapidly in the datacenter. The implementation can also allow configuration for discarding the value of DCTCP.Alpha after cwnd restart and timeouts.¶
[RFC3168] forbids the ECN-marking of pure ACK packets because of the inability of TCP to mitigate ACK-path congestion and protocol-wise preferential treatment by routers. However dropping pure ACKs rather than ECN marking them is disadvantageous in traffic scenarios typical in the datacenter. Because of the prevalence of bursty traffic patterns which involve transient congestion, the dropping of ACKS causes subsequent retransmission. It is recommended that the implementation a configuration knob that forces ECT on TCP pure ACK packets.¶
DCTCP and conventional TCP congestion control does not coexist well. In DCTCP, the marking threshold is set very low value to reduce queueing delay, thus a relatively small amount of congestion will exceed the marking threshold. During such periods of congestion, conventional TCP will suffer packet losses and quickly scale back cwnd. DCTCP, on the other hand, will use the fraction of marked packets to scale back cwnd. Thus rate reduction in DCTCP will be much lower than that of conventional TCP, and DCTCP traffic will dominate conventional TCP traffic traversing the same link. Hence if the traffic in the datacenter is a mix of conventional TCP and DCTCP, it is recommended that DCTCP traffic be segregated from conventional TCP traffic. [MORGANSTANLEY] describes a deployment that uses IP DSCP bits where AQM is applied to DCTCP traffic, while TCP traffic is managed via drop-tail queueing.¶
Today's commodity switches allow configuration of a different marking/drop profile for non-TCP and non-IP packets. Non-TCP and non-IP packets should be able to pass through the switch unless the switch is really out of buffers. If the traffic in the datacenter consists of such traffic (including UDP), one possible mitigation would be to mark IP packets as ECT even when there is no transport that is reacting to the marking.¶
Since DCTCP relies on congestion marking by the switch, DCTCP can only be deployed in datacenters where the network infrastructure supports ECN. The switches may also support configuration of the congestion threshold used for marking. The proposed parameterization can be configured with switches that implement RED. [DCTCP10] provides a theoretical basis for selecting the congestion threshold, but as with estimation gain, it may be more practical to rely on experimentation or simply to use the default configuration of the device. DCTCP will degrade to loss-based congestion control when transiting a congested drop-tail link.¶
DCTCP requires changes on both the sender and the receiver, so both endpoints must support DCTCP. Furthermore, DCTCP provides no mechanism for negotiating its use, so both endpoints must be configured through some out-of-band mechanism to use DCTCP. A variant of DCTCP that can be deployed unilaterally and only requires standard ECN behavior has been described in [ODCTCP][BSDCAN], but requires additional experimental evaluation.¶
DCTCP relies on the sender's ability to reconstruct the stream of CE codepoints received by the remote endpoint. To accomplish this, DCTCP avoids using a single ACK packet to acknowledge segments received both with and without the CE codepoint set. However, if one or more ACK packets are dropped, it is possible that a subsequent ACK will cumulatively acknowledge a mix of CE and non-CE segments. This will, of course, result in a less accurate congestion estimate. There are some potential mitigations:¶
However, the affect of packet drops on DCTCP under real world conditions has not been analyzed.¶
DCTCP provides no mechanism for negotiating its use. Thus, there is additional management and configuration overhead required to ensure that DCTCP is not used with non-DCTCP endpoints. The affect of using DCTCP with a standard ECN endpoint has been analyzed in [ODCTCP][BSDCAN]. Furthermore, it is possible that other implementations may also modify [RFC3168] behavior without negotiation, causing further interoperability issues.¶
Much like standard TCP, DCTCP is biased against flows with longer RTTs. A method for improving the fairness of DCTCP has been proposed in [ADCTCP], but requires additional experimental evaluation.¶
This section documents the implementation status of the specification in this document, as recommended by [RFC6982].¶
This document describes DCTCP as implemented in Microsoft Windows Server 2012. Since publication of the first versions of this document, the Linux [LINUX] and FreeBSD [FREEBSD] operating systems have also implemented support for DCTCP in a way that is believed to follow this document.¶
DCTCP enhances ECN and thus inherits the security considerations discussed in [RFC3168]. The processing changes introduced by DCTCP do not exacerbate these considerations or introduce new ones. In particular, with either algorithm, the network infrastructure or the remote endpoint can falsely report congestion and thus cause the sender to reduce cwnd. However, this is no worse than what can be achieved by simply dropping packets.¶
This document has no actions for IANA.¶
The DCTCP algorithm was originally proposed and analyzed in [DCTCP10] by Mohammad Alizadeh, Albert Greenberg, Dave Maltz, Jitu Padhye, Parveen Patel, Balaji Prabhakar, Sudipta Sengupta, and Murari Sridharan.¶
Lars Eggert has received funding from the European Union's Horizon 2020 research and innovation program 2014-2018 under grant agreement No. 644866 ("SSICLOPS"). This document reflects only the authors' views and the European Commission is not responsible for any use that may be made of the information it contains.¶