]> Google

1600 Amphitheatre Parkway Mountain View, CA 94043 US warren@kumari.net

Loon

1600 Amphitheatre Parkway Mountain View, CA 94043 US ek@google.com

In many environments offering short-term or temporary Internet access (such as coffee shops), it is common to start new connections in a captive portal mode. This highly restricts what the customer can do until the customer has authenticated. This document describes a DHCP option (and a Router Advertisement (RA) extension) to inform clients that they are behind some sort of captive-portal device, and that they will need to authenticate to get Internet access. It is not a full solution to address all of the issues that clients may have with captive portals; it is designed to be used in larger solutions. The method of authenticating to, and interacting with the captive portal is out of scope of this document. [ This document is being collaborated on in Github at: https://github.com/wkumari/draft-ekwk-capport-rfc7710bis. The most recent version of the document, open issues, etc should all be available here. The authors (gratefully) accept pull requests. Text in square brackets will be removed before publication. ]

In many environments, users need to connect to a captive-portal device and agree to an Acceptable Use Policy (AUP) and / or provide billing information before they can access the Internet. It is anticipated that the IETF will work on a more fully featured protocol at some point, to ease interaction with Captive Portals. Regardless of how that protocol operates, it is expected that this document will provide needed functionality because the client will need to know when it is behind a captive portal and how to contact it. In order to present users with the payment or AUP pages, the captive-portal device has to intercept the user's connections and redirect the user to the captive portal, using methods that are very similar to man-in-the-middle (MITM) attacks. As increasing focus is placed on security, and end nodes adopt a more secure stance, these interception techniques will become less effective and/or more intrusive. This document describes a DHCP () option (Captive-Portal) and an IPv6 Router Advertisement (RA) () extension that informs clients that they are behind a captive-portal device and how to contact it.

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 .

The Captive Portal DHCP / RA Option informs the client that it is behind a captive portal and provides the URI to access an authentication page. This is primarily intended to improve the user experience by getting them to the captive portal faster; for the foreseeable future, captive portals will still need to implement the interception techniques to serve legacy clients, and clients will need to perform probing to detect captive portals. In order to support multiple "classes" of clients (e.g. IPv4 only, IPv6 only with DHCPv6 (), IPv6 only with RA) the captive portal can provide the URI via multiple methods (IPv4 DHCP, IPv6 DHCP, IPv6 RA). The captive portal operator should ensure that the URIs handed out are equivalent to reduce the chance of operational problems. The maximum length of the URI that can be carried in IPv4 DHCP is 255 bytes, so URIs longer than 255 bytes should not be used in IPv6 DHCP or IPv6 RA. In all variants of this option, the URI SHOULD be that of the captive portal API endpoint, conforming to the recommendations for such URIs [cite:API] (i.e. the URI SHOULD contain a DNS name and SHOULD reference a secure transport, e.g. https). A captive portal MAY do content negotiation [citation?] and attempt to redirect clients querying without an explicit indication of support for the captive portal API content type (i.e. without application/capport+json listed explicitly anywhere within an Accepts header [citation]). In so doing, the captive portal SHOULD redirect the client to the value associated with the "user-portal-url" API key. The URI SHOULD NOT contain an IP address literal. The URI parameter is not null terminated. Networks with no captive portals MAY explicitly indicate this condition by using this option with the IANA-assigned URI for this purpose (see ). Clients observing the URI value "urn:ietf:params:capport-unrestricted" MAY forego time-consuming forms of captive portal detection.

The format of the IPv4 Captive-Portal DHCP option is shown below.

Code: The Captive-Portal DHCPv4 Option (160) (one octet) Len: The length, in octets of the URI. URI: The URI for the captive portal API endpoint to which the user should connect (encoded following the rules in ).

The format of the IPv6 Captive-Portal DHCP option is shown below.

option-code: The Captive-Portal DHCPv6Option (103) (two octets) option-len: The length, in octets of the URI. URI: The URI for the captive portal API endpoint to which the user should connect (encoded following the rules in ). See , Section 5.7 for more examples of DHCP Options with URIs.

This section describes the Captive-Portal Router Advertisement option.

37 8-bit unsigned integer. The length of the option (including the Type and Length fields) in units of 8 bytes. The URI for the captive portal API endpoint to which the user should connect. This MUST be padded with NULL (0x00) to make the total option length (including the Type and Length fields) a multiple of 8 bytes.

Some captive portal network deployments may be unable to change, or unwilling to risk changing, the network infrastructure necessary to use any of the above options. In such deployments, when clear text HTTP intercept and redirection are used, a Link relation header (, Section 3.3) MAY be inserted to convey to a HTTP client (user agent) the associated Captive Portal API URI. HTTP user agents MUST ignore this link relation in any context other than when explicitly probing to detect the presence of a captive portal. Failure to do so could allow an attacker to inject a Captive Portal API URI other than the correct URI for a given network or for networks where there is no captive portal present at all.

This document requests two new IETF URN protocol parameter () entries. Thanks IANA!

Registry name: Captive Portal Unrestricted Identifier URN: urn:ietf:params:capport-unrestricted Specification: RFC TBD (this document) Repository: RFC TBD (this document) Index value: Only one value is defined (see URN above). No hierarchy is defined and therefore no sub-namespace registrations are possible.

Registry name: Captive Portal API Link Relation Type URN: urn:ietf:params:capport-api Specification: RFC TBD (this document) Repository: RFC TBD (this document) Index value: Only one value is defined (see URN above). No hierarchy is defined and therefore no sub-namespace registrations are possible.

An attacker with the ability to inject DHCP messages could include this option and so force users to contact an address of his choosing. As an attacker with this capability could simply list himself as the default gateway (and so intercept all the victim's traffic); this does not provide them with significantly more capabilities, but because this document removes the need for interception, the attacker may have an easier time performing the attack. As the operating systems and application that make use of this information know that they are connecting to a captive-portal device (as opposed to intercepted connections) they can render the page in a sandboxed environment and take other precautions, such as clearly labeling the page as untrusted. The means of sandboxing and user interface presenting this information is not covered in this document - by its nature it is implementation specific and best left to the application and user interface designers. Devices and systems that automatically connect to an open network could potentially be tracked using the techniques described in this document (forcing the user to continually authenticate, or exposing their browser fingerprint). However, similar tracking can already be performed with the standard captive portal mechanisms, so this technique does not give the attackers more capabilities. Captive portals are increasingly hijacking TLS connections to force browsers to talk to the portal. Providing the portal's URI via a DHCP or RA option is a cleaner technique, and reduces user expectations of being hijacked - this may improve security by making users more reluctant to accept TLS hijacking, which can be performed from beyond the network associated with the captive portal. By simplifying the interaction with the captive portal systems, and doing away with the need for interception, we think that users will be less likely to disable useful security safeguards like DNSSEC validation, VPNs, etc. In addition, because the system knows that it is behind a captive portal, it can know not to send cookies, credentials, etc. By handing out a URI using which is protected with TLS, the captive portal operator can attempt to reassure the user that the captive portal is not malicious.

This document is a -bis of RFC7710. Thanks to all of the original authors (Warren Kumari, Olafur Gudmundsson, Paul Ebersman, Steve Sheng), and original contributors. Also thanks to the CAPPORT WG for all of the discussion and improvements.

[RFC Editor: Please remove this section before publication ] From initial to -00. Import of RFC7710.