The OpenConnect VPN Protocol Version 1.1

The OpenConnect protocol combines the TLS protocol , Datagram TLS protocol and HTTP protocols to provide an Internet-Layer VPN channel. The channel is designed to operate using UDP packets, and fallback on TCP if that's not possible. In brief the protocol initiates an HTTP over TLS connection on a known port, where client authentication is performed. After this step, the client initiates an HTTP CONNECT command to establish a VPN channel over TCP. A secondary VPN channel over UDP will be established using information provided by the server using HTTP headers. At that point the raw IP packets flow, over the VPN channels.

The client and server establish a TLS connection over a known port, typically over 443, the port used for HTTPS. The client SHOULD negotiate TLS 1.1 or later, and support the following TLS protocol extensions. Server Name Indication : the client SHOULD provide the DNS name of the server in the TLS handshake. Application-Layer Protocol Negotiation : the client MAY provide this protocol name. The protocol name to be used is defined in .

In the OpenConnect VPN protocol, the server is always authenticated using its certificate. Once a client establishes a TCP connection to the server's well known port, it initiates the TLS protocol. In the first connection to the server, the client SHOULD verify the provided by the server certificate, and SHOULD store its public key for verification of subsequent sessions. Thus, subsequent sessions SHOULD check whether the server's key match the initial. The server's identity in the certificate SHOULD be placed in the certificate's SubjectAlternativeName field, and unless a special profile is assumed, it will be of type DNSName.

The OpenConnect VPN protocol allows for the following types of client authentication, or combinations of them. Password: a user can authenticate itself using a password. Certificate: a user can authenticate itself using a PKIX certificate it possesses. HTTP SPNEGO: a user can authenticate itself using a Kerberos ticket, or any other mechanism supported by SPNEGO (i.e., GSSAPI). The server is authenticated to the client using a PKIX certificate presented during the TLS negotiation. It is important to note that during the password and HTTP SPNEGO authentication methods, any headers allowed by the HTTP protocol can be present. In fact, there are legacy clients which assume that the server will keep a state using cookies, and send their username and password in different TLS and HTTP connections. This practice prevents the server from binding the TLS channel with the VPN session , and is discouraged. It is RECOMMENDED for clients to complete authentication in the same TLS session, and rely on TLS session resumption if reconnections to the server are needed. After the TLS session is established the client irrespective of the supported authentication methods, should send an HTTP POST request on "/" with a config-auth XML structure of type 'init'. An example of its contents follow.

v5.01 ]]> The precise DTD declarations for the contents of XML messages defined in this document are listed in . Also the HTTP Content-Type to be used for these XML structures MUST be 'text/xml'.

During the initial TLS protocol handshake the server may require a client certificate to be presented, depending on its configuration. Because the client certificate is sent in the clear during the handshake it SHOULD NOT contain other identifying information other than a username, or a pseudonymus identifier. It is RECOMMENDED to place the user identifier in the DN field of the certificate, using the UID object identifier (0.9.2342.19200300.100.1.1) . After the TLS session is established and the the config-auth XML structure of type 'init' is sent, the server should send it reply. If the certificate sent by the client was successfully validated, it should reply using the HTTP response code 200, and the contents of the reply should be a config-auth XML structure of type 'complete', as follows.

0.1(1) SSL VPN Service ]]> In that case the client should proceed to the establishment of the primary channel as in .

After the TLS session is established and the the config-auth XML structure of type 'init' is sent, the server will reply using forms the client software should prompt the user to fill in. Its reply utilizes a config-auth XML structure of type 'auth-request'.

Please enter your username ]]> The client may be asked to provide the information in separate forms as above, or may be asked combined as below.

Please enter your username ]]> The client software will then fill in the provided form and sent it back to the server using an HTTP POST on the location specified by the server (in the above examples it was "/auth"). The reply would then be of type 'auth-reply' as in the following example.

v5.01 test ]]> As mentioned above, the server may ask repeatedly for information until it believes the user is authenticated. For example, the server could present a second form asking for the password, after the username is provided, or ask for a second password if that is necessary. In these cases the server should respond with an HTTP 200 OK status code, and proceed sending its new request. If client authentication fails, the server MUST respond with an HTTP 401 unauthorized status code. Otherwise, on successful authentication the server should reply with a 200 HTTP code and use the 'complete' config-auth XML structure as in . Note, that sending the username and password in different messages will reveal the length of them to a passive eavesdropper. For that is is RECOMMENDED for clients to use the 'X-Pad' HTTP header, which will contain arbitrary printable data to make the message length a multiple of 64 bytes. An example session is shown in figure .

| | | | | TLS handshake Finished | | | <----------------------------------- | | | | | HTTP POST config-auth init | ,--------------------!. | -----------------------------------> |This is an HTTP over|_\ | | |TLS session. | | | `----------------------' | config-auth auth-request | | | <----------------------------------- | | | | | HTTP POST config-auth auth-reply | | | -----------------------------------> | | | | | config-auth complete | | | <----------------------------------- | | | | | HTTP CONNECT | | | -----------------------------------> | | | | | | | | =================================== | ====================== CSTP VPN session is established ======================= | =================================== | | | | | | ,-------------------------!. | TLS record packet with CSTP payload| |These packets show |_\ | -----------------------------------> |that IP traffic can start | | | |prior to the DTLS channel | | | |establishment. | | | `---------------------------' | TLS record packet with CSTP payload| | | <----------------------------------- | | | | | DTLS handshake Client Hello | | - - - - - - - - - - - - - - - - - - - - - - - - - - - > | | | | DTLS handshake Finished | | <- - - - - - - - - - - - - - - - - - - - - - - - - - - - | | | | | | | =================================== | ====================== DTLS VPN channel is established ======================= | =================================== | | | | | DTLS record packet with payload | | - - - - - - - - - - - - - - - - - - - - - - - - - - - > | | | | DTLS record packet with payload | | <- - - - - - - - - - - - - - - - - - - - - - - - - - - - Client ,--+---. ,----+-----. ,-. |Server| |ServerDTLS| `-' `------' `----------' /|\ | / \ ]]>

That type of authentication is performed using the HTTP SPNEGO protocol , a method which is available using the Generic Security Service API . The following approach is used to advertise the availability of the HTTP SPNEGO protocol by the client. A client which supports the HTTP SPNEGO protocol, SHOULD indicate it using the following header on in its initial request to the server with the config-auth 'init' XML structure.

After that the server would report a "401 Unauthorized" status code and authentication would proceed as specified in the HTTP SPNEGO protocol. The server may utilize the following header, to indicate that alternative authentication methods are available (e.g., with plain password), if authentication fails.

If client authentication fails, the server MUST respond with an HTTP 401 unauthorized status code. In that case, a client which received the previous header should retry authenticating to the server without sending the "X-Support-HTTP-Auth: true" header. Otherwise, on successful authentication the server should reply with a 200 HTTP code and use the 'complete' config-auth XML structure as in .

By the receipt of a success XML structure, the client SHOULD issue an HTTP CONNECT request. In addition it may provide the following headers. X-CSTP-Address-Type: A comma separated list of the requested address types. IPv4: when the client only supports IPv4 addresses. IPv6: when the client only supports IPv6 addresses. IPv4,IPv6: when the client supports both types of IP addresses. X-CSTP-Base-MTU: The MTU of the link as estimated by the client. X-CSTP-Accept-Encoding: A comma separated list of accepted compression algorithms for the CSTP channel. User-Agent: A string identifying the client software. For the options related to compression see for more information. An example CONNECT request is shown below.

After a successful receipt of an HTTP CONNECT request, the server should reply and provide the client with configuration parameters. The available options follow. X-CSTP-Address: The IPv4 address of the client, if IPv4 has been requested. X-CSTP-Netmask: An IPv4 netmask to be pushed to the client, if IPv4 has been requested. This should contain the mask on the P-t-P link and is RECOMMENDED the server address to be the first in defined network. X-CSTP-Address-IP6: The IPv6 address of the client in CIDR notation, if IPv6 has been requested. The prefix length is RECOMMENDED to be set to 127-bits according to . X-CSTP-DNS: The IP address of a DNS server that can be used for that session. X-CSTP-Default-Domain: The DNS domains the provided DNS servers respond for. X-CSTP-Split-Include: The network address of a route which is provided by this server. X-CSTP-Split-Exclude: The network address of a route that is not provided by this server. X-CSTP-Base-MTU: The MTU of the link as estimated by this server. X-CSTP-DynDNS: Set to "true" if the server is operating with a dynamic DNS address. X-CSTP-Content-Encoding: if present is it set to one of the values presented by the client in 'X-CSTP-Accept-Encoding' header. It will be the compression algorithm used in the CSTP channel. X-DTLS-Content-Encoding: if present is it set to one of the values presented by the client in 'X-DTLS-Accept-Encoding' header. It will be the compression algorithm used in the DTLS channel. The client is expected to treat the received parameters as his networking settings. If no "X-CSTP-Split-Include" headers are present, the client is expected to assign its default route through the VPN.

The previous HTTP message is the last HTTP message sent by the server. After that message, the established TCP channel is used to transport IP packets between the client and the server. The transferred packets encoding is discussed in . This channel will be referred as CSTP in the rest of this document.

To establish the secondary UDP-based channel, which will be referred to as the DTLS channel, the client must advertise support for it during the issue of the HTTP CONNECT request (see ). This is done by appending the following headers to the request. X-DTLS-Accept-Encoding: A comma separated list of accepted compression algorithms for the DTLS channel. X-DTLS-CipherSuite: Must contain the keyword PSK-NEGOTIATE. The DTLS channel utilizes the PSK key exchange method. The key material for this session is a 256-bit value generated with an exporter. The key material exporter uses the label "EXPORTER-openconnect-psk" without the quotes, and without any context value. In its client hello message the client must copy the value received in the 'X-DTLS-App-ID' header (after hex decoding it), to the session ID field of the DTLS client hello. That identifier, is not used for session resumption, and is used by the server to associate the DTLS channel with the CSTP channel. The following headers are used by the server's response to CONNECT, and are related to the DTLS channel establishment. X-DTLS-App-ID: A hex encoded value to be used as a DTLS application-specific identifier by the client. It serves as an identifier for the server to associate the incoming DTLS session with the TLS session. X-DTLS-Port: The port number to which the client should send UDP packets for DTLS. X-DTLS-CipherSuite: It must contain the value "PSK-NEGOTIATE" without any quotes. X-DTLS-Rekey-Time: The time (in seconds) after which the DTLS session should rekey, see . Only considered if applicable to the negotiated DTLS protocol. X-DTLS-Rekey-Method: The method used in DTLS rekey, see . Only considered if applicable to the negotiated DTLS protocol. Note that in future versions of the Datagram TLS protocol (see ), clients should supply the value in 'X-DTLS-App-ID' header as a PSK identity after hex decoding it.

Previous versions of this protocol utilized a special DTLS protocol negotiation, based on an unpublished description of the DTLS protocol. This section attempts to summarize this negotiation, but may not be entirely accurate. To establish the legacy UDP-based channel, the client must advertise support for it during the issue of the HTTP CONNECT request (see ). This is done by appending the following headers to the request. X-DTLS-Accept-Encoding: A comma separated list of accepted compression algorithms for the DTLS channel. X-DTLS-Master-Secret: A hex encoded pre-master secret to be used in the legacy DTLS session negotiation. X-DTLS-CipherSuite: A colon-separated list of ciphers (e.g., the string PSK-NEGOTIATE:AES256-SHA:AES128-SHA:DES-CBC3-SHA). The DTLS channel utilizes session resumption as a method for preshared-key authentication. That is the value presented in X-DTLS-Master-Secret is set as a master secret to be resumed. The session ID value is sent by the server on the response to CONNECT using the 'X-DTLS-Session-ID' header. That header provides a hex-encoded value of the DTLS session ID to be used by the client. The following headers are used by the server's response to CONNECT, and are related to the DTLS channel establishment. X-DTLS-Session-ID: A hex encoded value to be used as a DTLS session ID by the client. It also serves as an identifier for the server to associate the incoming DTLS session with the TLS session. X-DTLS-Port: The port number to which the client should send UDP packets for DTLS. X-DTLS-CipherSuite: The ciphersuite selected by the server. It should be one of the options present in the client's X-DTLS-CipherSuite header. X-DTLS-Rekey-Time: The time (in seconds) after which the DTLS session should rekey, see . X-DTLS-Rekey-Method: The method used in DTLS rekey, see . The following table lists the ciphers negotiated via the X-DTLS-CipherSuite header, and the corresponding DTLS ciphersuite. OpenConnect cipher DTLS ciphersuite DTLS version DES-CBC3-SHATLS_RSA_WITH_3DES_EDE_CBC_SHA1DTLS 0.9 (pre-draft version) AES128-SHATLS_RSA_WITH_AES_128_CBC_SHA1DTLS 0.9 (pre-draft version) AES256-SHATLS_RSA_WITH_AES_256_CBC_SHA1DTLS 0.9 (pre-draft version) OC-DTLS1_2-AES128-GCMTLS_RSA_WITH_AES_128_GCM_SHA256DTLS 1.2 OC-DTLS1_2-AES256-GCMTLS_RSA_WITH_AES_256_GCM_SHA256DTLS 1.2 The legacy DTLS protocol negotiation described in this section, is similar to DTLS 1.0 except for the following deviations: The negotiated protocol version for the handshake and record headers is 1.0 instead of 254.255. The Hello Verify and Hello verify request messages are included in the handshake hashes. The handshake header is not included as part of the handshake hashes. The ChangeCipherSpec message is 3 byte long instead of 1, and contains the handshake sequence number (2-bytes long) appended to the message id.

The format of the packets sent over the primary channel consists of an 8-bytes header followed by data. The whole packet in encapsulated in a TLS record (see ). The bytes of the header indicate the type of data that follow, and their contents are explained in . byte value 0fixed to 0x53 (S) 1fixed to 0x54 (T) 2fixed to 0x46 (F) 3fixed to 0x01 4-5The length of the packet that follows this header in big endian order 6The type of the payload that follows (see for available types) 7fixed to 0x00 The available payload types are listed in . Value Description 0x00DATA: the TLS record packet contains an IPv4 or IPv6 packet 0x03DPD-REQ: used for dead peer detection. Once sent the peer should reply with a DPD-RESP packet, that has the same contents as the original request. 0x04DPD-RESP: used as a response to a previously received DPD-REQ. 0x05DISCONNECT: sent by the client (or server) to terminate the session. No data is associated with this request. The session will be invalidated after such request. 0x07KEEPALIVE: sent by any peer. No data is associated with this request. 0x08COMPRESSED DATA: a Data packet which is compressed prior to encryption. 0x09TERMINATE: sent by the server to indicate that the server is shutting down. No data is associated with this request.

The format of the packets sent over the UDP channel consists of an 1-byte header followed by data. The header byte indicates the type of data that follow as in . The header and the data are encapsulated in a DTLS record packet (see ).

During the exchange of session parameters (), the server advertizes the methods available for session rekey using the "X-CSTP-Rekey-Method" and "X-DTLS-Rekey-Method" HTTP headers. The available options for both the server and client are listed below. none: no rekey; the session will go on until 2^48 DTLS records have been exchanged, or 2^64 TLS records. ssl: a TLS or DTLS rehandshake will be performed periodically. new-tunnel: the session will tear down and the client will reconnect periodically. When the value is other than "none" the rekey period is determinated by the "X-CSTP-Rekey-Time" and "X-DTLS-Rekey-Time" headers. These headers contain the time in seconds after which a session should rekey. It should be noted that when the "ssl" rekey option is used, care must be taken by both the client and the server to ensure that either safe renegotiation is used (), or that the identity of the peer remained the same.

In OpenConnect there are two packet types that can be used for keep-alive or dead peer detection, as shown in . These are the DPD-REQ and KeepAlive packets. The timings of the transmission of these packets are set by the server, and they for the DPD are advisory to a client. However, any peer receiving these packets MUST response with the appropriate packet. For DPD-REQ packets, the response MUST be DPD-RESP, and for KeepAlive packets the response must be another KeepAlive packet. The main difference between these two types of packets, is that the DPD packets similarly to are sent when there is no traffic or when the other party requests them, and allow for arbitrary data to be attached, making them suitable for Path MTU detection. The server advertizes the suggested periods during the exchange of session parameters (). The available headers are listed below. X-CSTP-DPD: applicable to CSTP channel; contains a relative time in seconds. X-CSTP-Keepalive: applicable to CSTP channel; contains a relative time in seconds. X-DTLS-DPD: applicable to DTLS channel; contains a relative time in seconds. X-DTLS-Keepalive: applicable to DTLS channel; contains a relative time in seconds.