The quick download:
CWR and ECE flags let TCP connections signal congestion explicitly, so routers warn senders to slow down instead of dropping packets.
-
ECN can reduce reliance on packet drops by creating a four-field signaling chain across IP and TCP headers. By warning senders about congestion before queues overflow, it can reduce retransmissions and latency for streaming, browsing, and real-time traffic when the path supports ECN
-
The handshake negotiates ECN support upfront; routers mark congestion via CE, receivers echo it via ECE, and senders reduce their window via CWR
-
OS support is broad (Linux, macOS, iOS, Windows), but real-world adoption depends on whether your network infrastructure handles CE-marked packets correctly
-
If you see unexplained latency or retransmissions, verify that ECN is negotiated and handled correctly across the path, then correlate packet-level congestion signals with infrastructure metrics to identify the source
CWR (Congestion Window Reduced) and ECE (ECN-Echo) are two TCP header flags that work together to signal network congestion between sender and receiver without dropping packets.
CWR and ECE, combined with two IP header flags (ECT and CE), replace the traditional drop-based congestion signal with an explicit warning that lets senders slow down before packets are lost.
Why ECN Replaced Packet Drops as a Congestion Signal
Without ECN, the only congestion feedback mechanism was packet drops. Drops triggered retransmit timeouts (RTO) at the transport layer, adding latency that hit hardest on latency-sensitive applications such as streaming, browsing, and real-time communication.
How CWR and ECE Work Together
ECN required changes to both the TCP and IP headers (RFC 3168). Two one-bit flags were added to the TCP header:
- CWR (bit 8)
- ECE (bit 9)
Two flags were repurposed in the IP header’s differentiated services field:
- ECT (bit 14)
- CE (bit 15)

Step 1: ECN Negotiation During the Handshake
During the synchronization phase of a connection between client and server, the TCP CWR and ECE flags work together to establish whether the connection supports congestion notification.
The sender transmits a SYN packet with the ECE and CWR flags set, and the receiver responds with a SYN-ACK that has only the ECE flag set. Any other configuration indicates a non-ECN setup.
Step 2: Router Marks Congestion Instead of Dropping
In an ECN-aware network, the process works like this (simplified): when a router detects congestion, instead of dropping packets destined for a receiver, it marks them with the CE flag in the IP header and delivers the packet.
Step 3: Receiver Signals Congestion Back to Sender
Before acknowledging the receipt of the packet, the receiver sets the ECE flag in the TCP header of the ACK and sends it back to the sender.
After receiving an ECE-marked ACK, the sender treats the signal as congestion feedback and reduces its congestion window according to its congestion-control algorithm. In classic TCP Reno-style behavior, this commonly means reducing the congestion window and slow-start threshold by about half.
Together, these three steps provide an alternative congestion signal: the router marks congestion with CE, the receiver echoes it via ECE, and the sender responds by reducing its congestion window and setting CWR.
This can prevent some congestion-related drops and retransmissions, but it does not guarantee that no packets will be lost.
Is ECN Widely Supported Today?
ECN support is now broadly available across modern operating systems and increasingly enabled by default:
- Linux: ECN has been supported for many years. Most Linux distributions default to passive negotiation, accepting ECN when requested by the peer rather than initiating it on outbound connections.
- macOS and iOS: Apple introduced ECN beginning with iOS 9 and OS X El Capitan, then gradually enabled it by default across supported Apple platforms over subsequent releases.
- Windows: ECN support arrived with Windows Vista/Server 2008. Windows Server 2012 introduced Data Center TCP (DCTCP), which uses ECN for congestion signaling in data-center workloads. Consumer Windows generally leaves ECN disabled for standard TCP traffic by default, although administrators can enable or configure it with netsh or PowerShell.
Despite broad OS support, real-world adoption depends on network infrastructure. Some older or misconfigured routers mishandle ECN-marked packets, which is why passive negotiation (ECN enabled only when both sides agree) remains the safer default for public internet traffic.
Where ECN Is Heading: L4S and AccECN
ECN is also a building block for newer low-latency networking approaches. Low Latency, Low Loss, and Scalable throughput (L4S) uses ECN marking to provide more frequent and precise congestion feedback, helping compatible senders and networks keep queues short while maintaining high throughput.
AccECN extends the original ECN feedback model so endpoints can communicate more detailed information about congestion markings and support newer use cases.
These technologies are not universal replacements for classic ECN, but they show why ECN remains relevant in modern transport design.
Should You Enable ECN?
For modern, controlled environments, ECN can be worth evaluating, particularly where you can test endpoint and network-path compatibility.
The main caveat is network infrastructure: if your traffic passes through older routers that mishandle CE-marked packets, ECN can cause more problems than it solves.
If you’re seeing unexplained latency spikes or retransmission patterns in your network traffic, check whether ECN is being negotiated, preserved, and handled correctly across the relevant path.
LM Envision helps teams monitor infrastructure and network performance across hybrid environments. Use it alongside packet captures and host-level TCP data to investigate whether latency, packet loss, or resource constraints coincide with application-performance issues.
For traffic that crosses external networks and internet paths, LogicMonitor Digital Experience extends that visibility beyond your internal environment into the provider paths, ISPs, and routing behavior that ECN-aware traffic actually depends on.
Turn TCP congestion signals into faster troubleshooting
Correlate packet-level findings with infrastructure, network, application, and external-path performance data to determine whether latency and retransmissions originate inside your environment or beyond it.
FAQs
What happens if a router does not support ECN?
If a router does not support ECN marking, the connection can still use traditional drop-based congestion signaling, provided the endpoints and path successfully negotiate a compatible mode. When queues fill, packets may be dropped, and the sender detects congestion through duplicate acknowledgments or retransmission timeouts. ECN negotiation helps endpoints avoid using ECN when they do not mutually support it, but it cannot guarantee that every middlebox will handle ECN correctly.
Can ECN cause connectivity problems?
Yes, some older or misconfigured middleboxes may mishandle packets with ECN-related markings or negotiation flags, which can cause connectivity or performance problems. For that reason, many systems use conservative or passive defaults. Test ECN behavior on representative paths before enabling it network-wide.
How do you enable ECN on Windows?
ECN can be enabled on Windows via PowerShell using the Set-NetTCPSetting cmdlet. Windows Server 2012 and later also support Data Center TCP (DCTCP), which uses ECN for congestion signaling in data center environments. Standard Windows configurations keep ECN disabled by default.




