Server Name Indication

When making a TLS connection, the client requests a digital certificate from the web server. Once the server sends the certificate, the client examines it and compares the name it was trying to connect to with the name(s) included in the certificate. If a match occurs, the connection proceeds as normal. If a match is not found, the user may be warned of the discrepancy and the connection may abort as the mismatch may indicate an attempted man-in-the-middle attack. However, some applications allow the user to bypass the warning to proceed with the connection, with the user taking on the responsibility of trusting the certificate and, by extension, the connection.

However it may be difficult — or even impossible, due to lack of a full list of all names in advance — to obtain a single certificate that covers all names a server will be responsible for. A server that is responsible for multiple hostnames is likely to need to present a different certificate for each name (or small group of names). Since 2005, CAcert has run experiments on different methods of using TLS on virtual servers.[2] Most of the experiments are unsatisfactory and impractical. For example, it is possible to use subjectAltName to contain multiple domains controlled by one person[3] in a single certificate. Such "unified communications certificates" must be reissued every time the list of domains changes.

Name-based virtual hosting allows multiple DNS hostnames to be hosted by a single server (usually a web server) on the same IP address. To achieve this, the server uses a hostname presented by the client as part of the protocol (for HTTP the name is presented in the host header). However, when using HTTPS, the TLS handshake happens before the server sees any HTTP headers. Therefore, it is not possible for the server to use the information in the HTTP host header to decide which certificate to present and as such only names covered by the same certificate can be served from the same IP address.

In practice, this means that an HTTPS server can only serve one domain (or small group of domains) per IP address for secured and efficient browsing. Assigning a separate IP address for each site increases the cost of hosting, since requests for IP addresses must be justified to the regional Internet registry and IPv4 addresses are now exhausted. The result is that many websites are effectively constrained from using secure communications over IPv4. IPv6 address space is not exhausted, so websites served using IPv6 are unaffected by this issue.

SNI addresses this issue by having the client send the name of the virtual domain as part of the TLS negotiation's ClientHello message.[4] This enables the server to select the correct virtual domain early and present the browser with the certificate containing the correct name. Therefore, with clients and servers that implement SNI, a server with a single IP address can serve a group of domain names for which it is impractical to get a common certificate.

SNI was added to the IETF's Internet RFCs in June 2003 through RFC 3546, Transport Layer Security (TLS) Extensions. The latest version of the standard is RFC 6066.

The desired hostname is not encrypted, so an eavesdropper can see which site is being requested. This helps security companies provide a filtering feature[5][6][7] and governments implement censorship.[8] While domain fronting was used as a workaround,[9] now both Google and AWS have taken action to disallow this - and it's therefore becoming less of an alternative.[10]

As of mid 2018, an extension to TLS 1.3 called Encrypted SNI (ESNI)[11] is being rolled out in an "experimental phase" to address this risk of domain eavesdropping.[12][13][14]

Support for ESNI was incorporated into Firefox Nightly in October 2018[15]. While the mainline Firefox browser supports ESNI, it requires using DNS-over-HTTPS to be enabled[16].

In March 2020, ESNI specification was renamed Encrypted Client Hello (ECHO for short) and reworked to encrypt the entire Client Hello (not just Server Name Indication).[17] In May 2020, the short name changed from ECHO to ECH.[18]

In 2004, a patch for adding TLS/SNI into OpenSSL was created by the EdelKey project.[19] In 2006, this patch was then ported to the development branch of OpenSSL, and in 2007 it was back-ported to OpenSSL 0.9.8 (first released in 0.9.8f[20]).

For an application program to implement SNI, the TLS library it uses must implement it and the application must pass the hostname to the TLS library. Further complicating matters, the TLS library may either be included in the application program or be a component of the underlying operating system. Because of this, some browsers implement SNI when running on any operating system, while others implement it only when running on certain operating systems.

• RFC 6066 (obsoletes RFC 4366, which obsoleted RFC 3546)