DNS Access Denied Error page
draft-reddy-dnsop-error-page-03

Abstract

When a DNS server filters a query the response conveys no detailed explanation of why the query was blocked, leading to end-user confusion. This document defines a method to return an URI that explains the reason the DNS query was filtered.

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."

This Internet-Draft will expire on February 26, 2021.

Copyright Notice

Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved.

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

• 1. Introduction
• 2. Terminology
• 3. Error page URI EDNS0 option format
• 4. Error page URL Processing
• 5. Sign and Verify
• 6. ERROR Page
• 7. Usability Considerations
• 8. Security Considerations
• 9. IANA Considerations
• 9.1. A New Error Page URL EDNS Option
• 10. Acknowledgements
• 11. References
• 11.1. Normative References
• 11.2. Informative References
• Authors' Addresses

1. Introduction

DNS filters are deployed for a variety of reasons including endpoint security, parental filtering, and filtering required by law enforcement. These are discussed in more detail below:

• Various network security services are provided by Enterprise networks to protect endpoints (e.g., Hosts including IoT devices). Network-based security solutions such as firewalls and Intrusion Prevention Systems (IPS) rely on network traffic inspection to implement perimeter-based security policies. The network security services may, for example, prevent malware download, block known malicious domains, block phishing sites, etc. These network security services act on DNS queries originating from endpoints. For example, DNS firewalls, a method of expressing DNS response policy information inside specially constructed DNS zones, known as Response Policy Zones (RPZs) allows DNS servers to modify DNS responses in real time to stop access to malware and phishing domains. Note that some of the commonly known types of malware are viruses, worms, trojans, bots, ransomware, backdoors, spyware, and adware.
• Network devices in a home network offer network security to protect the devices connected to the home network by performing DNS-based content filtering. The network security service may, for example, block access to specific domains to enforce parental control, block access to malware sites, etc.
• ISPs typically block access to some domains due to a requirement imposed by an external entity (e.g., Law Enforcement Agency) by performing DNS-based content filtering.

DNS responses can be filtered by sending a bogus ("forged") A or AAAA response, NXDOMAIN error or empty answer, or an extended error code defined in [I-D.ietf-dnsop-extended-error]. Each of these have advantages and disadvantages, discussed below:

1. The DNS response is forged providing IP addresses that points to a HTTP(S) server alerting the end user of the reason for blocking access to the domain (e.g., malware). When a HTTP(S) enabled domain name is blocked, the network security device presents a block page instead of the HTTP response from the content provider. If an HTTP enabled domain name is blocked, the network security device intercepts the HTTP request and returns a block page over HTTP. If an HTTPS enabled domain is blocked, the block page is also served over HTTPS. In order to return a block page over HTTPS, man in the middle (MITM) is enabled on endpoints by generating a local root certificate and an accompanying (local) public/private key pair. The local root certificate is installed on the endpoint, and the network security device(s) store a copy of the private key. During the TLS handshake, the network security device modifies the certificate provided by the server and (re)signs it with the private key from the local root certificate.
   • However, configuring the local root certificate on endpoints is not viable option in several deployments like Home networks, Schools, Small Office/Home Office (SOHO), and Small/ Medium Enterprise (SME). In these cases, the typical behavior is that the forged DNS response directs the user towards a server hosted to display the block page which breaks the TLS connection. For web-browsing this then results in an HTTPS certificate error message indicating that a secure connection could not be established, which gives no information to the end-user about the reason for the error. The typical errors are "The security certificate presented by this website was not issued by a trusted certificate authority" (Internet Explorer/Edge"), "The site's security certificate is not trusted" (Chrome), "This Connection is Untrusted" (Firefox), “Safari can't verify the identity of the website…” (Safari on MacOS)".
   • Enterprise networks do not assume that all the devices connected to their network are managed by the IT team or Mobile Device Management (MDM) devices, especially in the quite common BYOD ("Bring Your Own Device") scenario. In addition, the local root certificate cannot be installed on IoT devices without a device management tool.
   • An end user does not know why the connection was reset and, consequently, may repeatedly try to unsuccessfully reach the domain. Frustrated, the end user may use insecure interfaces to reach the domain, potentially compromising both security and privacy. Furthermore, certificate errors train users to click through certificate errors, which is poor security practice. To eliminate the need for an end user to click through certificate errors, an end user may manually install a local root certificate [Chrome-Install-Cert] on a host device. Doing so, however, is also poor security practice as it creates a security vulnerability that may be exploited by a MITM attack. When the manually installed local root certificate expires, the user has to (again) manually install the new local root certificate.
2. The DNS response is forged to provide a NXDOMAIN response to cause the DNS lookup to terminate in failure. In this case, an end user does not know why the domain cannot be reached, and may repeatedly try to unsuccessfully reach the domain. Frustrated, the end user may use insecure interfaces to reach the domain, potentially compromising both security and privacy.
3. The extended error codes Blocked, Censored, and Filtered defined in [I-D.ietf-dnsop-extended-error] can be returned by the DNS server to provide additional information about the cause of an DNS error. If the extended error code "Forged answer" defined in [I-D.ietf-dnsop-extended-error] is returned by the DNS server, the client can identify the DNS response is forged and the reason for HTTPS certificate error. These extended error codes do not suffer from the limitations discussed in (1) and (2) but the user still does not know the exact reason nor the user is aware of the exact entity blocking the access to the domain. For example, a DNS server may block access domain based on the content category like "Adult Content" to enforce parental control, "Violence & Terrorism" due to an external requirement imposed by an external entity (e.g., Law Enforcement Agency), etc. The content categories for domains cannot be standardized because the classification of domains into content categories is vendor specific, typically ranges from 40 to 100 types of categories depending on the vendor and the categories keep evolving. Further, the threat data used to categorize domains may sometimes mis-classify domains (e.g., Domains wrongly classified as DGA (Domain Generation Algorithm) by deep learning techniques, domain wrongly classified as phishing due to crowd sourcing, new domains not categorized by the threat data, etc.). The end user needs to know the contact details of the IT/InfoSec team to raise a complaint.

No matter which type of response is generated (forged IP address, NXDOMAIN or empty answer, or an extended error code), the user who generated the query has little chance to understand which entity filtered the query, how to report a mistake in the filter, or why the entity filtered it at all. This document describes a mechanism to provide a URI which, when accessed, provides such information to the user.

One of the other benefits of this document is eliminating the need to "spoof" block pages for HTTPS resources, as the block page no longer needs to create a signed certificate when blocking a destination. This avoids the need to install an local root certificate authority on those IT-managed devices.

2. Terminology

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119][RFC8174] when, and only when, they appear in all capitals, as shown here.

This document makes use of the terms defined in [RFC8499] and [I-D.ietf-dnsop-terminology-ter].

'Encrypted DNS' refers to DNS-over-HTTPS [RFC8484], DNS-over-TLS [RFC7858], or DNS-over-QUIC [I-D.ietf-dprive-dnsoquic].

3. Error page URI EDNS0 option format

This draft uses an EDNS0 ([RFC6891]) option to include the URI that gives additional information about the cause of blocking access to a domain in a DNS response. The option is structured as follows:

                                             1   1   1   1   1   1
     0   1   2   3   4   5   6   7   8   9   0   1   2   3   4   5
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   |                            OPTION-CODE                        |
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   |                           OPTION-LENGTH                       |
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   | ERROR-PAGE-URI-LENGTH  (fixed size)                           |
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   / ERROR-PAGE-URI (variable size)                                |
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   | ERROR-PAGE-SIG-ALG (fixed size)                               |
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+
   / ERROR-PAGE-SIG  (variable size)                               |
   +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+

Field definition details:

• OPTION-CODE, 2-octets/16-bits (defined in [RFC6891]) for Error page URI, value is TBD. [RFC Editor: change TBD to the proper code once assigned by IANA.]
• OPTION-LENGTH, 2-octets/16-bits (defined in [RFC6891]) contains the length of the payload (everything after OPTION-LENGTH) in octets. The variability of the option length stems from the variable-length ERROR-PAGE-URI and ERROR-PAGE-SIG fields.
• ERROR-PAGE-URI-LENGTH, 2-octets/16-bits representing the length of ERROR-PAGE-URI. It MUST NOT be set to 0.
• ERROR-PAGE-URI, a variable length UTF-8 encoded [RFC5198] text field containing the URI Template that gives additional information about the cause of blocking access to a domain. The ERROR-PAGE-URI field MUST NOT be zero octets in length.
• ERROR-PAGE-SIG-ALG, 16-bits, which is the signature algorithm used to generate the signature for the Error page URI Template. The values are defined in the TLS SignatureScheme [TLS-SIG-SCHEME] with limitations described in Section 5.
• ERROR-PAGE-SIG, a variable length field containing the signature of the Error page URI Template. The signature generation process is discussed in Section 5.

The Error page URI option MUST only be included in error responses SERVFAIL, NXDOMAIN or REFUSED to a query that included the OPT Pseudo-RR [RFC6891].

The URI Template defined in ERROR-PAGE-URI describes how to construct the URL to fetch the error page. The agent acting as HTTPS client on the endpoint encodes a FQDN to which access is denied into an HTTP GET request to retrieve the error page. The HTTPS server returning the error page defines the URI used by the HTTP GET request through the use of a URI template. The URI Template is processed with a defined variable "target-domain" whose value is set to the FQDN to which access is denied. The FQDN is encoded with base64url [RFC4648] and then provided as the variable value for "target-domain" to expand the URI Template into a URI reference in the HTTP GET request. Padding characters for base64url MUST NOT be included.

An example is illustrated below:

If the URI template is "https://block.example.net/block-page{?target-domain}" for the HTTPS server returning the error page and access to the target domain "example.com" is blocked by the encrypted DNS server, the variable "target-domain" has the value "example.com" base64url encoded into an HTTP GET request. In the above example, the expansion of the above URI Template is "https://block.example.net/block-page?target-domain=ZXhhbXBsZS5jb20".

HTTP/2 [RFC7540] is the minimum RECOMMENDED version of HTTP to use to retrieve the error page. The HTTPS client retrieving the error page MUST verify the entire certification path per [RFC5280]. The HTTPS client additionally uses validation techniques as described in [RFC6125] to compare the domain name in the error page URI to the server certificate provided in TLS handshake. See [RFC7525] for additional TLS recommendations.

4. Error page URL Processing

The DNS client MUST follow the following rules to process the Error page URI EDNS0 option:

• The Error page URI EDNS0 option is susceptible to forgery. In order to defend against this attack the DNS client MUST NOT process the DNS response with Error page URI EDNS0 option unless DNS messages exchanged are cryptographically protected using encrypted DNS.
• If an DNS client has enabled opportunistic privacy profile (Section 5 of [RFC8310]) for DoT, the DNS client will either fallback to an encrypted connection without authenticating the DNS server provided by the local network or fallback to clear text DNS, and cannot exchange encrypted DNS messages. The fallback adversely impacts security and privacy. If the DNS client has enabled opportunistic privacy profile for DoT, the client MUST NOT process the DNS response with Error page URI EDNS0 option.
• If an DNS client has enabled strict privacy profile (Section 5 of [RFC8310]) for DoT, the DNS client requires an encrypted connection and successful authentication of the DNS server; this mitigates both passive eavesdropping and client redirection (at the expense of providing no DNS service if an encrypted, authenticated connection is not available). If the DNS client has enabled strict privacy profile for DoT, the client can process the DNS response with Error page URI EDNS0 option. Note that the strict and opportunistic privacy profiles as defined in [RFC8310] only applies to DoT protocol, there has been no such distinction made for DoH protocol.
• If the DNS response contains more than one Error page URI EDNS0 option, the DNS client MUST discard multiple Error page URI EDNS0 options in the DNS response.
• The Error page URI EDNS0 option MUST be processed by the DNS client for a "Censored", "Blocked" or "Filtered" extended error codes and MUST be ignored for any other type of extended DNS error code. When "Censored", "Blocked" or "Filtered" extended error code is returned in conjunction with an Error page URI EDNS0 option, any other resource records in the answer MUST be ignored by clients supporting this specification.
• If the encrypted DNS server does not offer DNS filtering service (e.g., prevent access to malware sites or objectionable content (e.g., legal obligation)), the DNS client MUST reject the Error page URI EDNS0 option. The mechanism discussed in [I-D.reddy-add-server-policy-selection] can be used by the DNS client to assess whether the encrypted DNS server performs DNS-based content filtering.
• If the Error page URI EDNS0 option is included in a NOERROR RCODE message, the DNS client MUST reject the Error page URI EDNS0 option.
• The DNS client MUST reject the error page URI if the scheme is not "https".
• The DNS client verifies the signature in the ERROR-PAGE-SIG field following the mechanism discussed in Section 5. If the signature is valid, the client can positively identify that the Error page URI EDNS0 option has been generated by the encrypted DNS server and the encrypted DNS server did not forward the Error Page URI EDNS0 option from an upstream resolver. If signature validation fails, the DNS client MUST reject the Error page URI EDNS0 option.

Because EDNS is a hop-by-hop extension to DNS and EDNS responses are not cached (Section 6.2.1 of [RFC6891]), a DNS resolver MUST NOT propagate a received Error Page URI EDNS0 option, because an attacker could insert a bogus URI; instead, if access to a domain is denied, the DNS resolver MUST generate its own Error Page URI. Because DNS forwarders (or DNS proxies) are supposed to propagate unknown EDNS0 options (Section 4.1 and Section 4.4.1 of [RFC5625]), the Error Page URI EDNS0 option may get propagated. To detect both of these scenarios, the Error Page URI Template is protected with an object signature as described in Section 5.

5. Sign and Verify

The signature algorithms for generating signature for DNS resource record sets (RRsets) are defined in [DNSKEY-IANA]. The "mandatory-to-implement" algorithms are defined in [RFC8624] are RSA, ECDSA and EdDSA. Along similar lines, the encrypted DNS server's end-entity certificate's public key and the signature algorithm with which the key can be used are RSA, ECDSA and EdDSA [RFC8446]. If ECDSA is used, it is RECOMMENDED to use the deterministic digital signature generation procedure of the Elliptic Curve Digital Signature Algorithm (ECDSA), specified in [RFC6979].

The signature is generated by the encrypted DNS server using the Error page URI Template, private key of the encrypted DNS server's end-entity certificate as inputs to the signature algorithm. The signature algorithm in the ERROR-PAGE-SIG-ALG field MUST be compatible with the key in the DNS server's end-entity certificate. The implementation MUST support the same set of algorithms in the TLS client for validating the signature in the CertificateVerify message from the server in the TLS handshake and in the DNS client to validate the signature for the Error page URI Template. As a reminder, the server's end-entity certificate's public key will be compatible with the selected authentication algorithm from the client's "signature_algorithms" TLS extension (Section 4.4.2.2 of [RFC8624]).

If the signature algorithm in the ERROR-PAGE-SIG-ALG field is not compatible with the key in the DNS server's end-entity certificate, the DNS client MUST reject the Error page URI EDNS0 option. The DNS client verifies the signature using the signature in the ERROR-PAGE-SIG field, error page URI Template and DNS server's end-entity certificate's public key as inputs to the signature algorithm. For example, if Ed25519 is used, Ed25519 signature algorithm and verification of the Ed25519 signature are described in Sections 5.1.6 and 5.1.7 of [RFC8032], respectively.

6. ERROR Page

The following text outlines the RECOMMENDED contents of an error page to assist the operator developing the error page.

• The exact reason for blocking access to the domain. If the domain is blocked based on some threat data, the threat type associated with the blocked domain can be provided/displayed to the end user. For example, the reason can indicate the type of malware blocked like spyware and the damage it can do the security and privacy of the user.
• The domain name blocked.
• If query was blocked by regulation, a pointer to a regulatory text that mandates this query block.
• The entity (or organization) blocking the access to the domain and contact details of the IT/InfoSec team to raise a complaint.
• The blocked error page to not include Ads and dynamic content.

The content of the error page discussed above is non-normative, the above text only provides the guidelines and template for the error page and.

• Does not attempt to offer an exhaustive list for the contents of an error page.
• It is not intended to form the basis of any legal/compliance for developing the error page.

7. Usability Considerations

The error page SHOULD be returned in the user's preferred language as expressed by the Accept-Language header. If the error page is displayed in a language not known to the end user and assuming Internationalization features failed, browser extensions to translate to user's native language can be used. For example, "Google Translate" extension [Chrome-Translate] provided by Google on Chrome can be used by the user to translate the error page. The "Google Translate" extension automatically detects whether the language of a page is different from the language the user has selected. If it is in a different language, a banner appears at the top of the page. The user can click on the Translate button in the banner to have all the text on the page appear in the language selected by the user.

8. Security Considerations

Security considerations in [I-D.ietf-dnsop-extended-error] and [RFC8624] need to be taken into consideration.

The Error page URI EDNS0 option causes an HTTPS retrieval by the client. To prevent forgery of the Error page URI EDNS0 option, this specification requires it only be sent only over an encrypted DNS channel with an authorized DNS server.

The DNS client can learn the filtering capability of an encrypted DNS server using [I-D.reddy-add-server-policy-selection]. [I-D.reddy-add-server-policy-selection] also discusses how a DNS client can authenticate it is connecting to an encrypted DNS server hosted by a specific organization (e.g., ISP). This information is cryptographically signed to attest its authenticity. It is particularly useful when the encrypted DNS server is insecurely discovered and prevents the client from connecting to an attacker's encrypted DNS server.

To reduce threat surface when retrieving the Error page URL over HTTPS, the client SHOULD perform the retrieval using an isolated environment. Such isolation should prevent transmitting cookies, block JavaScript, block auto-fill of credentials or personal information, and be isolated from the user's normal environment. Browsers perform some of these limitations today when accessing captive portals [Safari-Cookie], during private browsing, or using containerization [Facebook-Container]).

The encrypted DNS session provides transport security for the interaction between the DNS client and server, but DNSSEC signing and validation is not possible for the Error Page URI EDNS0 option returning the error page URI template. For example, if access to a doman is blocked, the encrypted DNS resolver can rewrite the response to send NXDOMAIN error and Error page URI EDNS0 option in the DNS response, it will omit DNSSEC RRsets, because the modified responses cannot be verified by DNSSEC signatures. However, the signature in the Error Page URI EDNS0 option provides data origin authentication of the Error Page URI EDNS0 option. Nevertheless, the Error Page URI EDNS0 option returning the error page URI Template should be treated only as diagnostic information and MUST NOT alter DNS protocol processing.

By design, the object referenced by the error page URL potentially exposes additional information about the DNS resolution process that may leak information. An example of this is the reason for blocking the access to the domain name and the entity blocking access to the domain.

9. IANA Considerations

9.1. A New Error Page URL EDNS Option

This document defines a new EDNS(0) option, entitled "Error Page URI", assigned a value of TBD from the "DNS EDNS0 Option Codes (OPT)" registry [to be removed upon publication: [http://www.iana.org/assignments/dns-parameters/dns-parameters.xhtml#dns-parameters-11]

Value  Name                 Status    Reference
-----  ----------------     ------    ------------------
 TBD   Error Page URI      Standard       [ This document ]

10. Acknowledgements

Thanks to Vittorio Bertola, Wes Hardaker, Ben Schwartz, Erid Orth, Viktor Dukhovni and Bob Harold for the comments.

11. References

11.1. Normative References

11.2. Informative References

Authors' Addresses