Network Working Group M. Kosters Internet-Draft Network Solutions, Inc. Expires: May 18, 2001 November 17, 2000 DNSSEC Opt-in for Large Zones draft-kosters-dnsext-dnssec-opt-in-00.txt Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. 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." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire on May 18, 2001. Copyright Notice Copyright (C) The Internet Society (2000). All Rights Reserved. Abstract In order for DNSSEC to be deployed operationally with large zones and little operational impact, there needs to be included a mechanism that allows for the separation of secure versus unsecure views of zones. This needs to be done in a transparent fashion that allows DNSSEC to be deployed in an incremental manner. This document proposes the use of an extended RCODE to signify that a DNSSEC-aware requestor may have to re-query for the information, if and only if, the delegation is not yet secure. Thus, one can maintain two views of the zone and expand the DNSSEC zone as demand warrants. Kosters Expires May 18, 2001 [Page 1] Internet-Draft DNSSEC Opt In November 2000 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Protocol Additions . . . . . . . . . . . . . . . . . . . . . . . 4 4. Security Considerations . . . . . . . . . . . . . . . . . . . . 6 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 6 6. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 7 Full Copyright Statement . . . . . . . . . . . . . . . . . . . . 8 Kosters Expires May 18, 2001 [Page 2] Internet-Draft DNSSEC Opt In November 2000 1. Introduction DNS is an unsecure system. The key features that gives DNS its power can also be its chief weaknesses. One feature is the facility to delegate branches of information from one set of servers to another. Currently, this is done in a non-cryptographically verified way that allows spoofing attacks. For example, Eugene Kashpureff exploited this vulnerability to redirect www.netsol.com and www.internic.net to his own website. If this delegated information had been cryptographically verified, this attack would not have been able to occur. In recent years, there has been much work within the IETF regarding DNS security. There are a number of RFCs that integrate public key technology within DNS to enable cryptographically-verified answers. To this end, three new resource record types (RR's) have been defined: o KEY -- a public key of the zone o SIG - a signature of an accompanying RR o NXT - a negative response record Within the zone, each authoritative RR will have accompanying SIG RR's that can be verified with the KEY RR of the zone. Each KEY RR can be verified hierarchically with a SIG RR from the direct parent zone. For unsecure delegations, a null-KEY RR is inserted in the parent zone. Finally, NXT RR's and their accompanying SIG RR's are issued in the case of a negative reply. As a zone maintainer, transitioning to a secure zone has a high overhead in the following areas: KEY RR At a delegation point, the zone maintainer needs to place a NULL key and accompanying SIG RR's when the child zone is not known to be secure. NXT RR Each delegation needs to be lexigraphically ordered so that a NXT RR can be generated and signed with SIG RR's. For large zone operators, generating the zone file is a very time consuming process. In the resolution process, NXT lookups require that the server replace efficient hash structures with a lexigraphically ordered search structure which degrades lookup performance. This lookup performance is a critical element for a high-query rate DNS server. Thus, the net effect is when one initially secures a zone as defined in RFC2535[4], the net overhead is massive because of the following factors: Kosters Expires May 18, 2001 [Page 3] 1. Zone ordering and maintenance for large zones is difficult and expensive. 2. Adding null-KEY RR's, NXT RR's and their accompanying SIG RR's for unsecure delegations will consume large amounts of memory (6x the current memory requirements). 3. Having a less efficient look-up algorithm to provide answers to queries will degrade overall performance. 4. Very little initial payoff (anticipate only a small fraction of delegations to be signed. This equates to less than 1% over the first six months). 5. Unsecured delegations are more expensive at the parent than secure delegations (NULL KEY). 2. Rationale As DNSSEC is initially deployed, it is anticipated that DNSSEC adoption will be slow to materialize. It is also anticipated that DNSSEC security resolution will be top down. Thus for DNSSEC to be widely adopted, the root zone and GTLD zones will need to be signed. Based on the implications previously listed, a large zone maintainer such as the administrator of COM, needs to create an infrastructure that is an order of magnitude larger than its current state with very little initial benefit. This document proposes an alternative opt-in approach that minimizes the expense and complexity to ease adoption of DNSSEC for large zones by allowing for an alternate view of secured only delegations. 3. Protocol Additions The opt-in proposal allows for a zone operator to maintain two views of its delegations - one being non-DNSSEC and the other being DNSSSEC aware. The non-DNSSEC view will have all delegations - both secured and non-secured. The DNSSEC aware view will only have secured delegations. It is assumed that neither view will have any innate knowledge of the other's delegations. Thus, the cost of securing a zone is proportional to the demand of its delegations with the added benefit of no longer having to maintain NULL KEY RRs for unsecure delegations. To determine which view each DNS query packet is to be queried against, there is a simple algorithm to be followed: 1. The DNSSEC view is to be queried when the DO bit is set within the EDNS0 OPT meta RR as indicated in [6]. 2. The DNSSEC view is to be queried when the query type is SIG, KEY, or NXT and the RRs added match the query name and query type. If the query does not follow either case (1) or (2), the non-DNSSEC view MUST be consulted by default. Kosters Expires May 18, 2001 [Page 4] Internet-Draft DNSSEC Opt In November 2000 Since the DNSSEC view will have a subset of the actual delegations of that zone, it will not be able to respond to an unsecured delgation. To that end, a new extended RCODE MUST be set within the EDNS OPT RR for the resolver to retry again with the DO bit not set. This RCODE is referred to as "RETRY-NO-SEC" (RS). In the context of the EDNS0 OPT meta-RR, the RS value will be determined by IANA. Setting the RS RCODE in a response indicates to the resolver that the resolver is retrying the query again without the DO bit set. The behavior of the authority and additional records section being populated should be the same using the RS RCODE as the the RCODE being set to NXDOMAIN. Therefore, the resolver will be able to verify that the answer does not exist within the secure zone since the NXT RR will be sent in the Authority section. To avoid caching, the server SHOULD set the TTL on the NXT RR to 0. Example: Consider a zone with the secure names 3, 6, and 9, and unsecure names 2, 4, 5, 7, and 8. Unsecured zone Contents: @ SOA 2 NS 3 NS 4 NS 5 NS 6 NS 7 NS 8 NS 9 NS Secured zone Contents: @ SOA, SIG SOA, NXT(3), SIG NXT 3 NS, SIG NS, NXT(6), SIG NXT 6 NS, SIG NS, NXT(9), SIG NXT 9 NS, SIG NS, NXT(@), SIG NXT 1. Query for 5 RR type A with EDNS0 DO bit set, the response would return with the extended RCODE RS bit set: RCODE=RS Authority Section: SOA, SIG SOA, 3 NXT(6), SIG NXT Additional Section: KEY, SIG KEY Kosters Expires May 18, 2001 [Page 5] Internet-Draft DNSSEC Opt In November 2000 The source would then retry without the EDNS0 DO bit set which would return an answer as defined in RFC1035[2]. 2. Query for 55 RR type A with EDNS0 DO bit set, the response would return with the extended RCODE RS bit set: RCODE=RS Authority Section: SOA, SIG SOA, 3 NXT(6), SIG NXT Additional Section: KEY, SIG KEY The source would then retry without the EDNS0 DO bit set which would return an answer as defined in RFC1035[2]. 3. Query for 33 RR type KEY without EDNS DO bit set. The response would return with an answer as defined in RFC2535[4]. 4. Query for 3 RR type A, with EDNS0 DO bit set, the response would be the same as defined in RFC2535[4]. 4. Security Considerations This draft is different and separate from RFC2535[4] in that it allows for secured delegation paths to exist but does not allow for secure answers to unsecured delegations at the parent level. Increased exposure will be marginal given that the children are unsecure. 5. IANA Considerations Allocation of the most significant bit of the RCODE field in the EDNS0 OPT meta-RR is required. 6. Acknowledgements This document is based on a rough draft by Brian Wellington, and input from Olafur Gudmundsson. References [1] Mockapetris, P.V., "Domain names - concepts and facilities", RFC 1034, STD 13, Nov 1987. [2] Mockapetris, P.V., "Domain names - implementation and specification", RFC 1035, STD 13, Nov 1987. Kosters Expires May 18, 2001 [Page 6] Internet-Draft DNSSEC Opt In November 2000 [3] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", RFC 2119, BCP 14, March 1997. [4] Eastlake, D., "Domain Name System Security Extensions", RFC 2535, March 1999. [5] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", RFC 2671, August 1999. [6] Conrad, D. R., "Indicating Resolver Support of DNSSEC (work in progress)", August 2000. Author's Address Mark Kosters Network Solutions, Inc. 505 Huntmar Park Drive Herndon, VA 22070 US Phone: +1 703 948-3362 EMail: markk@netsol.com URI: http://www.netsol.com Kosters Expires May 18, 2001 [Page 7] Internet-Draft DNSSEC Opt In November 2000 Full Copyright Statement Copyright (C) The Internet Society (2000). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Acknowledgement Funding for the RFC editor function is currently provided by the Internet Society. Kosters Expires May 18, 2001 [Page 8]