Server Name Indication Through a Proxy

ColdProxy Team7 min read

Server Name Indication sends a TLS ClientHello through a proxy tunnel so a shared destination edge can select the matching certificate

Server Name Indication (SNI) produced a clear result in our direct ColdProxy TLS test: 3 of 3 handshakes worked with SNI; 3 of 3 failed without it.

A client can receive 200 Connection Established and then meet a TLS alert before the first web request. That sequence places the fault at the destination handshake.

SNI carries the requested hostname in TLS ClientHello, helping the destination choose the certificate for that host. HTTP CONNECT is the concrete proxy model below.

Key takeaways

  • SNI is a TLS ClientHello extension. On shared infrastructure, the server can use its hostname to choose the intended virtual host and certificate.
  • An HTTP proxy processes CONNECT before destination TLS begins. Validate tunnel establishment and TLS identity as distinct checks.
  • Check CONNECT and ClientHello as two visibility points. ECH applies to the ClientHello field.
  • A useful HTTPS proxy test covers three layers: transport, TLS identity, and the application request.

What Server Name Indication does

Many websites share an underlying network address at a hosting platform, content delivery network, or load balancer. The destination edge can use the requested hostname to choose the matching site identity before it returns a certificate.

"Allow TLS clients to provide to the TLS server the name of the server they are contacting." - RFC 6066

SNI guides virtual-host and certificate selection. Authentication still happens when the client validates the certificate chain and checks that the certificate covers the intended hostname. HTTP host or authority information arrives after TLS has been established.

The IANA TLS ExtensionType Values registry assigns server_name value 0 and encrypted_client_hello value 65037. Use these protocol assignments when reading a TLS trace.

The HTTP proxy overview explains the forwarding model and request flow used here.

Why SNI matters on today's web

HTTP Archive's 2025 Web Almanac found that more than 98.8% of measured mobile requests used HTTPS. It also reported HTTPS on 97.5% of measured desktop homepages and 97.3% of measured mobile homepages, while TLS 1.3 powered roughly 76% of measured web pages. These figures describe HTTPS and TLS adoption. Their practical implication here is that socket reachability leaves certificate selection and identity untested.

When the intended name is absent or wrong, a shared edge can respond in several valid ways:

  • return a default certificate for another virtual host;
  • stop the handshake with an alert; or
  • continue until the client detects a certificate-name mismatch.

The server chooses the behavior. A missing SNI value therefore has no universal error message.

The HTTP proxy and TLS sequence

HTTP CONNECT provides the concrete model. Other ordinary byte-forwarding tunnels carry the same destination TLS session once the transport path exists.

  • Tunnel target. The HTTP CONNECT definition uses a request target containing the destination host and port. The explicit proxy processes that target so it can establish the route.
  • Tunnel response. A successful response opens the bidirectional byte path. It confirms transport establishment only.
  • Destination ClientHello. The client sends ClientHello through the tunnel. When the client sends SNI without ECH, the hostname appears in that message.
  • Certificate and identity. The destination edge selects a virtual host and certificate. The client then validates the certificate against the hostname it intended to reach.

A CONNECT success and a TLS failure can occur in sequence. Record both results; a single "proxy worked" flag hides which layer needs attention.

The HTTP CONNECT tunnel guide covers target syntax, the response, and byte forwarding in detail.

What an explicit proxy can see

A proxy path exposes different signals to different participants. Keep the tunnel request and destination handshake distinct.

  • At CONNECT: the CONNECT request target supplies the explicit proxy with a destination host and port. It contains a hostname when that target names one; another host form can also be used.
  • At ClientHello: when a client sends conventional SNI without ECH, the hostname is present in the TLS message. An observer with access to that part of the path can inspect it.
  • After the handshake: HTTP path, authority, headers, cookies, and response content travel inside destination TLS in an ordinary transport tunnel.

Record CONNECT and ClientHello visibility as two fields in the trace. TLS interception is a different architecture and should be named when it exists.

How ECH changes SNI visibility

RFC 9849, published in March 2026, standardizes Encrypted Client Hello. ECH encrypts a ClientHello under a server public key and protects SNI plus other sensitive fields such as the ALPN list.

ECH uses an outer and inner ClientHello. The outer message carries information needed to reach an ECH-capable service. The encrypted inner message carries the actual SNI and other protected values for the destination handshake.

The protection is scoped to ClientHello. DNS activity, destination addresses, an explicit proxy request, and traffic patterns are separate signals that may still reveal related information.

As a dated deployment example, Cloudflare's September 2023 ECH announcement described ECH as the successor to ESNI and reported availability on all Cloudflare plans. That report documents one operator deployment. Actual negotiation still depends on client, DNS, network, and server support for the connection being tested.

What we observed on ColdProxy's public endpoint

On July 23, 2026, we repeated a direct TLS comparison against the public coldproxy.com endpoint. The purpose was to isolate destination handshake behavior from customer proxy routing.

The saved method used Node.js tls.connect to coldproxy.com on port 443. Both variants enabled certificate validation, offered h2 and http/1.1, used a 15-second timeout, and ran in one process and network environment. No application data was sent. The server-name setting was the only changed connection option: coldproxy.com in one group and no value in the other.

  • With SNI: all 3 of 3 attempts connected, passed certificate authorization, negotiated TLS 1.3, and selected HTTP/2 through ALPN. The certificate covered coldproxy.com and *.coldproxy.com.
  • Without SNI: all 3 of 3 attempts ended with a TLS handshake-failure alert.

Every successful attempt returned the same certificate identity and HTTP/2 ALPN result. The no-SNI group stopped earlier, so no certificate or ALPN result was available. Recording those fields together makes the two traces easy to compare without sending an HTTP request or exposing application data.

This one-endpoint snapshot supports a bounded result. Other servers and proxy-to-destination paths can behave differently, and the artifact reveals no private server configuration.

The six results point to the next checks: validate certificate identity, then send an application request.

A three-layer check for SNI failures

Record transport, TLS, and HTTP as separate result fields. Each field gives the next investigation a smaller search area.

  • CONNECT succeeds, then a TLS alert appears -> TLS layer. Compare the intended hostname, CONNECT target, SNI value, and destination trace.
  • A certificate arrives with the wrong name -> identity layer. Keep hostname validation enabled, then compare the certificate names with the hostname the client requested.
  • The request times out -> stage still unknown. Record time to CONNECT, time to TLS, and time to the first HTTP response before labeling the cause.

The proxy testing checklist extends this sequence. Use Proxy Checker for IPv4 or IPv6 reachability, exit address, and location. Use a TLS-aware client or trace for SNI, certificate selection, and ECH.

Where SNI fits in a proxy decision

Treat a proxy path as ready only after all three layers pass: the tunnel opens, the certificate matches the intended hostname, and the application request succeeds.

After these technical checks pass, use ColdProxy's pricing overview to compare proxy type, coverage, plan families, buying models, and your measured workflow results.

Frequently Asked Questions

How does SNI work in TLS?

Server Name Indication places the requested hostname in ClientHello. On a server that handles several HTTPS hostnames, that value guides certificate selection before HTTP begins. The client still validates the returned certificate, so SNI guides selection while certificate validation establishes server identity.

Does an HTTP proxy replace or change SNI?

An ordinary transport tunnel forwards the destination TLS bytes. The proxy handles CONNECT, and the client creates ClientHello with its SNI value. A later TLS failure points to the client hostname, certificate exchange, or destination policy. Inspect the handshake trace and confirm any TLS interception explicitly.

When can an explicit proxy see a hostname?

The proxy processes the CONNECT target and sees a hostname when that target names one. Another host form can be used, so the exact visible field depends on the request target. Encrypting the client-to-proxy leg protects the request from unrelated observers while the proxy endpoint still receives it.

What happens when a client omits SNI?

The outcome depends on the destination edge. It may return a default certificate, continue under another virtual host, or stop the handshake. In this run, all 3 of 3 no-SNI attempts failed at one public endpoint. That observation is scoped to the endpoint and date recorded in the test.

Does ECH hide SNI from a proxy?

ECH protects the actual SNI inside the encrypted ClientHello. An explicit HTTP proxy still processes the separate CONNECT target supplied earlier, and signals such as DNS or destination addresses can remain available elsewhere in the path. Check the actual client, DNS, and server negotiation before assuming ECH was used.

ColdProxy Team

ColdProxy Team

Content Team

The ColdProxy Content Team consists of proxy-service experts, developers, and technical writers dedicated to providing clear, accurate insights on web scraping, online privacy, and advanced proxy technologies.