| NTP Working Group | D. Franke |
| Internet-Draft | Akamai |
| Intended status: Standards Track | D. Sibold |
| Expires: December 28, 2017 | K. Teichel |
| PTB | |
| June 26, 2017 |
Network Time Security for the Network Time Protocol
draft-ietf-ntp-using-nts-for-ntp-09
This memo specifies Network Time Security (NTS), a mechanism for using Transport Layer Security (TLS) and Authenticated Encryption with Associated Data (AEAD) to provide cryptographic security for the Network Time Protocol.
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 http://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 December 28, 2017.
Copyright (c) 2017 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 (http://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.
This memo specifies Network Time Security (NTS), a cryptographic security mechanism for network time synchronization. A complete specification is provided for application of NTS to the Network Time Protocol (NTP). However, certain sections of this memo are not inherently NTP-specific, and enable future work to apply them to other time synchronization protocols such as the Precision Time Protocol (PTP).
The objectives of NTS are as follows:
The Network Time Protocol includes many different operating modes to support various network topologies. In addition to its best-known and most-widely-used client-server mode, it also includes modes for synchronization between symmetric peers, a control mode for server monitoring and administration and a broadcast mode. These various modes have differing and contradictory requirements for security and performance. Symmetric and control modes demand mutual authentication and mutual replay protection, and for certain message types control mode may require confidentiality as well as authentication. Client-server mode places more stringent requirements on resource utilization than other modes, because servers may have vast number of clients and be unable to afford to maintain per-client state. However, client-server mode also has more relaxed security needs, because only the client requires replay protection: it is harmless for servers to process replayed packets. The security demands of symmetric and control modes, on the other hand, are in conflict with the resource-utilization demands of client-server mode: any scheme which provides replay protection inherently involves maintaining some state to keep track of what messages have already been seen.
In order to simulatenously serve these conflicting requirements, NTS is structured as a suite of three protocols:
It is intended that NTP implementations will use DTLS-encapsulated NTPv4 to secure symmetric mode and control mode, and use NTS-KE followed by NTS Extensions for NTPv4 to secure client/server mode. NTS does not support NTP's broadcast mode.
As previously stated, DTLS-encapsulated NTPv4 is trivial. The communicating parties establish a DTLS session and then exchange arbitrary NTP packets as DTLS Application Data.
The typical protocol flow for client/server mode is as follows. The client connects to the server on the NTS TCP port and the two parties perform a TLS handshake. Via the TLS channel, the parties negotiate some additional protocol parameters and the server sends the client a supply of cookies. The parties use TLS key export to extract key material which will be used in the next phase of the protocol. This negotiation takes only a single round trip, after which the server closes the connection and discards all associated state. At this point the NTS-KE phase of the protocol is complete.
Time synchronization proceeds over the NTP UDP port. The client sends the server an NTP client packet which includes several extension fields. Included among these fields are a cookie (previously provided by the server), and an authentication tag, computed using key material extracted from the NTS-KE handshake. The server uses the cookie to recover this key material (previously discarded to avoid maintaining state) and send back an authenticated response. The response includes a fresh, encrypted cookie which the client then sends back in the clear with its next request. (This constant refreshing of cookies is necessary in order to achieve NTS's unlinkability goal.)
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 RFC 2119.
Network Time Security makes use of both TLS (for NTS Key Establishment) and DTLS (for NTS-encapsulated NTPv4). In either case, the requirements and recommendations of this section are similar. The notation "(D)TLS" refers to both TLS and DTLS.
Since securing time protocols is (as of 2017) a novel application of (D)TLS, no backward-compatibility concerns exist to justify using obsolete, insecure, or otherwise broken TLS features or versions. We therefore put forward the following requirements and guidelines, roughly representing 2017's best practices.
Implementations MUST NOT negotiate (D)TLS versions earlier than 1.2.
Implementations willing to negotiate more than one possible version of (D)TLS SHOULD NOT respond to handshake failures by retrying with a downgraded protocol version. If they do, they MUST implement [RFC7507].
(D)TLS clients MUST NOT offer, and (D)TLS servers MUST not select, RC4 cipher suites. [RFC7465]
(D)TLS clients SHOULD offer, and (D)TLS servers SHOULD accept, the TLS Renegotiation Indication Extension. Regardless, they MUST NOT initiate or permit insecure renegotiation. (*)
(D)TLS clients SHOULD offer, and (D)TLS servers SHOULD accept, the TLS Session Hash and Extended Master Secret Extension. (*)
Use of the Application-Layer Protocol Negotation Extension is integral to NTS and support for it is REQUIRED for interoperability.
(*): Note that (D)TLS 1.3 or beyond may render the indicated recommendations inapplicable.
The NTS-encapsulated NTPv4 protocol proceeds in two parts. The two endpoints carry out a DTLS handshake in conformance with Section 3, with the client offering (via an ALPN extension), and the server accepting, an application-layer protocol of "ntp/4". Second, once the handshake is successfully completed, the two endpoints use the established channel to exchange arbitrary NTPv4 packets as DTLS-protected Application Data.
In addition to the requirements specified in Section 3, implementations MUST enforce the anti-replay mechanism specified in Section 4.1.2.6 of RFC 6347 (or an equivalent mechanism specified in a subsequent revision of DTLS). Servers wishing to enforce access control SHOULD either demand a client certificate or use a PSK-based handshake in order to establish the client's identity.
The NTS-encapsulated NTPv4 protocol is the RECOMMENDED mechanism for cryptographically securing mode 1 (symmetric active), 2 (symmetric passive), and 6 (control) NTPv4 traffic. It is equally safe for mode 3/4 (client/server) traffic, but is NOT RECOMMENDED for this purpose because it scales poorly compared to using NTS Extensions for NTPv4.
Since DTLS-encapsulated NTPv4 sessions may carry arbitrary NTP packets, there is no prescribed implication from an implementation's role as a DTLS client vs. DTLS server, to its role in the application-level Network Time Protocol. For example, it is entirely permissible for an implementation to initiate a DTLS handshake (thus acting in the role of DTLS client), and then once the handshake is completed, act as an NTP server with the DTLS server acting as an NTP client. The following guidelines are offered as sensible default behavior. Implementations may depart from this guidance if the user configures them to do so.
Implementations typically should not use DTLS-encapsulated NTPv4 for client/server mode, instead preferring to use NTS-KE and NTS Extensions for NTPv4. If DTLS-encapsulated NTPv4 is used for client/server mode, then the NTP client (mode 3) should be the DTLS client and the NTP server (mode 4) should be the DTLS server.
For control mode (6), the party sending queries should be the DTLS client and the party responding to the queries should be the DTLS server.
For symmetric operation between an active (mode 1) and passive (mode 2) peer, the active peer should be the DTLS client and the passive peer should be the DTLS server.
For symmetric operation between two active (mode 1) peers, both parties should attempt to initiate a DTLS session with their peer. If one handshake fails and the other succeeds, the successfully-established session should be used for traffic in both directions. If both handshakes succeed, either session may be used and packets should receive identical dispositon regardless of which of the two sessions they arrived over. Inactive sessions may be timed out but the redundant session should not be proactively closed.
If, likely as a result of user error, party A is configured as a symmetry active peer of party B, but party B is neither accepting DTLS handshakes from party A nor initiating one with it, then after a suitable number of failed attempts, party A may fall back to acting as an NTP client (mode 3) of party B using NTS-KE and NTS Extensions for NTPv4.
The NTS key establishment protocol is conducted via TCP port [[TBD1]]. The two endpoints carry out a TLS handshake in conformance with Section 3, with the client offering (via an ALPN extension), and the server accepting, an application-layer protocol of "ntske/1". Immediately following a successful handshake, the client SHALL send a single request (as Application Data encapsulated in the TLS-protected channel), then the server SHALL send a single response followed by a TLS "Close notify" alert and then discard the channel state.
The client's request and the server's response each SHALL consist of a sequence of records formatted according to Figure 1. The sequence SHALL be terminated by a "End of Message" record, which has a Record Type of zero and a zero-length body. Furthermore, requests and non-error responses each SHALL include exactly one NTS Next Protocol Negotiation record.
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|C| Record Type | Body Length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
. .
. Record Body .
. .
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 1
The requirement that all NTS-KE messages be terminated by an End of Message record makes them self-delimiting.
The fields of an NTS-KE record are defined as follows:
The following NTS-KE Record Types are defined.
The End of Message record has a Record Type number of 0 and an zero-length body. It MUST occur exactly once as the final record of every NTS-KE request and response. The Critical Bit MUST be set.
The NTS Next Protocol Negotiation record has a record type of 1. It MUST occur exactly once in every NTS-KE request and response. Its body consists of a sequence of 16-bit unsigned integers in network byte order. Each integer represents a Protocol ID from the IANA Network Time Security Next Protocols registry. The Critical Bit MUST be set.
The Protocol IDs listed in the client's NTS Next Protocol Negotiation record denote those protocols which the client wishes to speak using the key material established through this NTS-KE session. The Protocol IDs listed in the server's response MUST comprise a subset of those listed in the request, and denote those protocols which the server is willing and able to speak using the key material established through this NTS-KE session. The client MAY proceed with one or more of them. The request MUST list at least one protocol, but the response MAY be empty.
The Error record has a Record Type number of 2. Its body is exactly two octets long, consisting of an unsigned 16-bit integer in network byte order, denoting an error code. The Critical Bit MUST be set.
Clients MUST NOT include Error records in their request. If clients receive a server response which includes an Error record, they MUST discard any negotiated key material and MUST NOT proceed to the Next Protocol.
The following error code are defined.
The Warning record has a Record Type number of 3. Its body is exactly two octets long, consisting of an unsigned 16-bit integer in network byte order, denoting a warning code. The Critical Bit MUST be set.
Clients MUST NOT include Warning records in their request. If clients receive a server response which includes an Warning record, they MAY discard any negotiated key material and abort without proceeding to the Next Protocol. Unrecognized warning codes MUST be treated as errors.
This memo defines no warning codes.
The AEAD Algorithm Negotiation record has a Record Type number of 4. Its body consists of a sequence of unsigned 16-bit integers in network byte order, denoting Numeric Identifiers from the IANA AEAD registry. The Critical Bit MAY be set.
If the NTS Next Protocol Negotiation record offers "ntp/4",this record MUST be included exactly once. Other protocols MAY require it as well.
When included in a request, this record denotes which AEAD algorithms the client is willing to use to secure the Next Protocol, in decreasing preference order. When included in a response, this record denotes which algorithm the server chooses to use, or is empty if the server supports none of the algorithms offered. In requests, the list MUST include at least one algorithm. In responses, it MUST include at most one. Honoring the client's preference order is OPTIONAL: servers may select among any of the client's offered choices, even if they are able to support some other algorithm which the client prefers more.
Server implementations of NTS extensions for NTPv4 MUST support AEAD_AES_SIV_CMAC_256 (Numeric Identifier 15). That is, if the client includes AEAD_AES_SIV_CMAC_256 in its AEAD Algorithm Negotiation record, and the server accepts the "ntp/4" protocol in its NTS Next Protocol Negotiation record, then the server's AEAD Algorithm Negotation record MUST NOT be empty.
The New Cookie for NTPv4 record has a Record Type number of 5. The contents of its body SHALL be implementation-defined and clients MUST NOT attempt to interpret them. See Section 7 for a RECOMMENDED construction.
Clients MUST NOT send records of this type. Servers MUST send at least one record of this type, and SHOULD send eight of them, if they accept "ntp/4" as a Next Protocol. The Critical Bit SHOULD NOT be set.
Following a successful run of the NTS-KE protocol, key material SHALL be extracted according to RFC 5705. Inputs to the exporter function are to be constructed in a manner specific to the negotiated Next Protocol. However, all protocols which utilize NTS-KE MUST conform to the following two rules:
Following a successful run of the NTS-KE protocol wherein "ntp/4" is selected as a Next Protocol, two AEAD keys SHALL be extracted: a client-to-server (C2S) key and a server-to-client (S2C) key. These keys SHALL be computed according to RFC 5705, using the following inputs.
Implementations wishing to derive additional keys for private or experimental use MUST NOT do so by extending the above-specified syntax for per-association context values. Instead, they SHOULD use their own disambiguating label string. Note that RFC 5705 provides that disambiguating label strings beginning with "EXPERIMENTAL" MAY be used without IANA registration.
In general, an NTS-protected NTPv4 packet consists of:
Always included among the authenticated or authenticated-and-encrypted extensions are a cookie extension and a unique-identifier extension. The purpose of the cookie extension is to enable the server to offload storage of session state onto the client. The purpose of the unique-identifier extension is to protect the client from replay attacks.
The Unique Identifier extension has a Field Type of [[TBD2]]. When the extension is included in a client packet (mode 3), its body SHALL consist of a string of octets generated uniformly at random. The string SHOULD be 32 octets long. When the extension is included in a server packet (mode 4), its body SHALL contain the same octet string as was provided in the client packet to which the server is responding. Its use in modes other than client/server is not defined.
The Unique Identifier extension provides the client with a cryptographically strong means of detecting replayed packets. It may also be used standalone, without NTS, in which case it provides the client with a means of detecting spoofed packets from off-path attackers. Historically, NTP's origin timestamp field has played both these roles, but for cryptographic purposes this is suboptimal because it is only 64 bits long and, depending on implementation details, most of those bits may be predictable. In contrast, the Unique Identifier extension enables a degree of unpredictability and collision-resistance more consistent with cryptographic best practice.
The NTS Cookie extension has a Field Type of [[TBD3]]. Its purpose is to carry information which enables the server to recompute keys and other session state without having to store any per-client state. The contents of its body SHALL be implementation-defined and clients MUST NOT attempt to interpret them. See Section 7 for a RECOMMENDED construction. The NTS Cookie extension MUST NOT be included in NTP packets whose mode is other than 3 (client) or 4 (server).
The NTS Cookie Placeholder extension has a Field Type of [[TBD4]]. When this extension is included in a client packet (mode 3), it communicates to the server that the client wishes it to send additional cookies in its response. This extension MUST NOT be included in NTP packets whose mode is other than 3.
Whenever an NTS Cookie Placeholder extension is present, it MUST be accompanied by an NTS Cookie extension, and the body length of the NTS Cookie Placeholder extension MUST be the same as the body length of the NTS Cookie Extension. (This length requirement serves to ensure that the response will not be larger than the request, in order to improve timekeeping precision and prevent DDoS amplification). The contents of the NTS Cookie Placeholder extension's body are undefined and, aside from checking its length, MUST be ignored by the server.
The NTS Authenticator and Encrypted Extensions extension is the central cryptographic element of an NTS-protected NTP packet. Its Field Type is [[TBD5]] and the format of its body SHALL be as follows:
The Ciphertext field SHALL be formed by providing the following inputs to the negotiated AEAD Algorithm:
The NTS Authenticator and Encrypted Extensions extension MUST NOT be included in NTP packets whose mode is other than 3 (client) or 4 (server).
A client sending an NTS-protected request SHALL include the following extensions:
The client MAY include one or more NTS Cookie Placeholder extensions, which MUST be authenticated and MAY be encrypted. The number of NTS Cookie Placeholder extensions that the client includes SHOULD be such that if the client includes N placeholders and the server sends back N+1 cookies, the number of unused cookies stored by the client will come to eight. When both the client and server adhere to all cookie-management guidance provided in this memo, the number of placeholder extensions will equal the number of dropped packets since the last successful volley.
The client MAY include additional (non-NTS-related) extensions, which MAY appear prior to the NTS Authenticator and Encrypted Extensions extension (therefore authenticated but not encrypted), within it (therefore encrypted and authenticated), or after it (therefore neither encrypted nor authenticated). In general, however, the server MUST discard any unauthenticated extensions and process the packet as though they were not present. Servers MAY implement exceptions to this requirement for particular extensions if their specification explicitly provides for such.
Upon receiving an NTS-protected request, the server SHALL (through some implementation-defined mechanism) use the cookie to recover the AEAD Algorithm, C2S key, and S2C key associated with the request, and then use the C2S key to authenticate the packet and decrypt the ciphertext. If the cookie is valid and authentication and decryption succeed, then the server SHALL include the following extensions in its response:
The server MAY include additional (non-NTS-related) extensions, which MAY appear prior to the NTS Authenticator and Encrypted Extensions extension (therefore authenticated but not encrypted), within it (therefore encrypted and authenticated), or after it (therefore neither encrypted nor authenticated). In general, however, the client MUST discard any unauthenticated extensions and process the packet as though they were not present. Clients MAY implement exceptions to this requirement for particular extensions if their specification explicitly provides for such.
If the server is unable to validate the cookie or authenticate the request, it SHOULD respond with a Kiss-o'-Death packet (see RFC 5905, Section 7.4)) with kiss code "NTSN" (meaning "NTS NAK"). Such a response MUST include exactly one Unique Identifier extension whose contents SHALL echo those provided by the client. It MUST NOT include any NTS Cookie or NTS Authenticator and Encrypted Extensions extension.
Upon receiving an NTS-protected response, the client MUST verify that the Unique Identifier matches that of an outstanding request, and that the packet is authentic under the S2C key associated with that request. If either of these checks fails, the packet MUST be discarded without further processing.
Upon receiving an NTS NAK, the client MUST verify that the Unique Identifier matches that of an outstanding request. If this check fails, the packet MUST be discarded without further processing. If this check passes, the client SHOULD wait until the next poll for a valid NTS-protected response and if none is received, discard all cookies and AEAD keys associated with the server which sent the NAK and initiate a fresh NTS-KE handshake.
This section provides a RECOMMENDED way for servers to construct NTS cookies. Clients MUST NOT examine the cookie under the assumption that it is constructed according to this section.
The role of cookies in NTS is closely analagous to that of session cookies in TLS. Accordingly, the thematic resemblance of this section to RFC 5077 is deliberate, and the reader should likewise take heed of its security considerations.
Servers should select an AEAD algorithm which they will use to encrypt and authenticate cookies. The chosen algorithm should be one such as AEAD_AES_SIV_CMAC_256 which resists accidential nonce reuse, and it need not be the same as the one that was negotiated with the client. Servers should randomly generate and store a master AEAD key `K`. Servers should additionally choose a non-secret, unique value `I` as key-identifier for `K`.
Servers should periodically (e.g., once daily) generate a new pair (I,K) and immediately switch to using these values for all newly-generated cookies. Immediately following each such key rotation, servers should securely erase any keys generated two or more rotation periods prior. Servers should continue to accept any cookie generated using keys that they have not yet erased, even if those keys are no longer current. Erasing old keys provides for forward secrecy, limiting the scope of what old information can be stolen if a master key is somehow compromised. Holding on to a limited number of old keys allows clients to seamlessly transition from one generation to the next without having to perform a new NTS-KE handshake.
The need to keep keys synchronized across load-balanced clusters can make automatic key rotation challenging. However, the task can be accomplished without the need for central key-management infrastructure by using a ratchet, i.e., making each new key a deterministic, cryptographically pseudo-random function of its predecessor. A recommended concrete implementation of this approach is to use HKDF to derive new keys, using the key's predecessor as Input Keying Material and its key identifier as a salt.
To form a cookie, servers should first form a plaintext `P` consisting of the following fields:
Servers should the generate a nonce `N` uniformly at random, and form AEAD output `C` by encrypting `P` under key `K` with nonce `N` and no associated data.
The cookie should consist of the tuple `(I,N,C)`.
To verify and decrypt a cookie provided by the client, first parse it into its components `I`, `N`, and `C`. Use `I` to look up its decryption key `K`. If the key whose identifier is `I` has been erased or never existed, decryption fails; reply with an NTS NAK. Otherwise, attempt to decrypt and verify ciphertext `C` using key `K` and nonce `N` with no associated data. If decryption or verification fails, reply with an NTS NAK. Otherwise, parse out the contents of the resulting plaintext `P` to obtain the negotiated AEAD algorithm, S2C key, and C2S key.
IANA is requested to allocate two entries, identical except for the Transport Protocol, in the Service Name and Transport Protocol Port Number Registry as follows:
IANA is requested to allocate the following two entries in the Application-Layer Protocol Negotation (ALPN) Protocol IDs registry:
IANA is requested to allocate the following entry in the TLS Exporter Label Registry:
| Value | DTLS-OK | Reference | Note |
|---|---|---|---|
| EXPORTER-network-time-security/1 | Y | [[this memo]] |
IANA is requested to allocate the following entry in the registry of NTP Kiss-o'-Death codes:
| Code | Meaning |
|---|---|
| NTSN | NTS NAK |
IANA is requested to allocate the following entries in the NTP Extensions Field Types registry:
| Field Type | Meaning | Reference |
|---|---|---|
| [[TBD2]] | Unique Identifier | [[this memo]] |
| [[TBD3]] | NTS Cookie | [[this memo]] |
| [[TBD4]] | NTS Cookie Placeholder | [[this memo]] |
| [[TBD5]] | NTS Authenticator and Encrypted Extensions | [[this memo]] |
IANA is requested to create a new registry entitled "Network Time Security Key Establishment Record Types". Entries SHALL have the following fields:
The policy for allocation of new entries in this registry SHALL vary by the Type Number, as follows:
Applications for new entries SHALL specify the contents of the Description, Set Critical Bit and Reference fields and which of the above ranges the Type Number should be allocated from. Applicants MAY request a specific Type Number, and such requests MAY be granted at the registrar's discretion.
The initial contents of this registry SHALL be as follows:
| Field Number | Description | Critical | Reference |
|---|---|---|---|
| 0 | End of message | MUST | [[this memo]] |
| 1 | NTS next protocol negotiation | MUST | [[this memo]] |
| 2 | Error | MUST | [[this memo]] |
| 3 | Warning | MUST | [[this memo]] |
| 4 | AEAD algorithm negotiation | MAY | [[this memo]] |
| 5 | New cookie for NTPv4 | SHOULD NOT | [[this memo]] |
| 16384–32767 | Reserved for Private & Experimental Use | MAY | [[this memo]] |
IANA is requested to create a new registry entitled "Network Time Security Next Protocols". Entries SHALL have the following fields:
Applications for new entries in this registry SHALL specify all desired fields, and SHALL be granted upon approval by a Designated Expert. Protocol IDs 32768-65535 SHALL be reserved for Private or Experimental Use, and SHALL NOT be registered.
The initial contents of this registry SHALL be as follows:
| Protocol Name | Human-Readable Name | Reference |
|---|---|---|
| 0 | Network Time Protocol version 4 | [[this memo]] |
| 1 | Precision Time Protocol version 2 | Reserved by [[this memo]] |
| 32768-65535 | Reserved for Private or Experimental Use | Reserved by [[this memo]] |
IANA is requested to create two new registries entitled "Network Time Security Error Codes" and "Network Time Security Warning Codes". Entries in each SHALL have the following fields:
The policy for allocation of new entries in these registries SHALL vary by their Number, as follows:
The initial contents of the Network Time Security Error Codes Registry SHALL be as follows:
| Number | Description | Reference |
|---|---|---|
| 0 | Unrecognized Critical Extension | [[this memo]] |
| 1 | Bad Request | [[this memo]] |
The Network Time Security Warning Codes Registry SHALL initially be empty.
Certain non-standard and/or deprecated features of the Network Time Protocol enable clients to send a request to a server which causes the server to send a response much larger than the request. Servers which enable these features can be abused in order to amplify traffic volume in distributed denial-of-service (DDoS) attacks by sending them a request with a spoofed source IP. In recent years, attacks of this nature have become an endemic nuisance.
NTS is designed to avoid contributing any further to this problem by ensuring that NTS-related extensions included in server responses will be the same size as the NTS-related extensions sent by the client. In particular, this is why the client is required to send a separate and appropriately padded-out NTS Cookie Placeholder extension for every cookie it wants to get back, rather than being permitted simply to specify a desired quantity.
NTS's security goals are undermined if the client fails to verify that the X.509 certificate chain presented by the server is valid and rooted in a trusted certificate authority. [RFC5280] and [RFC6125] specifies how such verification is to be performed in general. However, the expectation that the client does not yet have a correctly-set system clock at the time of certificate verification presents difficulties with verifying that the certificate is within its validity period, i.e., that the current time lies between the times specified in the certificate's notBefore and notAfter fields, and it may be operationally necessary in some cases for a client to accept a certificate which appears to be expired or not yet valid. While there is no perfect solution to this problem, there are several mitigations the client can implement to make it more difficult for an adversary to successfully present an expired certificate:
Additional standardization work and infrastructure development is necessary before NTS can be used with public NTP server pools. First, a scheme needs to be specified for determining what constitutes an acceptable certificate for a pool server, such as establishing a value required to be contained in its Extended Key Usage attribute, and how to determine, given the DNS name of a pool, what Subject Alternative Name to expect in the certificates of its members. A more important matter, however, is that pool operators need procedures for establishing and maintaining trust in their members. Pools in existence as of 2017 are volunteer-run, with minimal requirements for admission and no organized effort to monitor pool servers for misbehavior. Without any sort of policing in place, there is nothing to prevent an adversary from going through normal channels to obtain a valid certificate for participation in a pool and then proceeding to serve maliciously inaccurate time.
In a packet delay attack, an adversary with the ability to act as a man-in-the-middle delays time synchronization packets between client and server asymmetrically [RFC7384]. Since NTP's formula for computing time offset relies on the assumption that network latency is roughly symmetrical, this leads to the client to compute an inaccurate value [Mizrahi]. The delay attack does not reorder or modify the content of the exchanged synchronization packets. Therefore, cryptographic means do not provide a feasible way to mitigate this attack. However, the maximum error that an adversary can introduce is bounded by half of the round trip delay.
[RFC5905] specifies a parameter called MAXDIST which denotes the maximum round-trip latency (including not only the immediate round trip between client and server but the whole distance back to the reference clock as reported in the Root Delay filed) that a client will tolerate before concluding that the server is unsuitable for synchronization. The standard value for MAXDIST is one second, although some implementations use larger values. Whatever value a client chooses, the maximum error which can be introduced by a delay attack is MAXDIST/2.
Usage of multiple time sources, or multiple network paths to a given time source [Shpiner], may also serve to mitigate delay attacks if the adversary is in control of only some of the paths.
At various points in NTS, the generation of cryptographically secure random numbers is required. See [RFC4086] for guidelines concerning this topic.
Unlinkability prevents a device from being tracked when it changes network addresses (e.g. because said device moved between different networks). In other words, unlinkability thwarts an attacker that seeks to link a new network address used by a device with a network address that it was formerly using, because of recognizable data that the device persistently sends as part of an NTS-secured NTP association. This is the justification for continually supplying the client with fresh cookies, so that a cookie never represents recognizable data in the sense outlined above.
NTS's unlinkability objective is merely to not leak any additional data that could be used to link a device's network address. NTS does not rectify legacy linkability issues that are already present in NTP. Thus, a client that requires unlinkability MUST also minimize information transmitted in a client query (mode 3) packet as described in the draft [I-D.ietf-ntp-data-minimization].
The unlinkability objective only holds for time synchronization traffic, as opposed to key exchange traffic. This implies that it cannot be guaranteed for devices that function not only as time clients, but also as time servers (because the latter can be externally triggered to send authentication data).
It should also be noted that it could be possible to link devices that operate as time servers from their time synchronization traffic, using information exposed in (mode 4) server response packets (e.g. reference ID, reference time, stratum, poll). Also, devices that respond to NTP control queries could be linked using the information revealed by control queries.
NTS does not protect the confidentiality of information in NTP's header fields. When clients implement [I-D.ietf-ntp-data-minimization], client packet headers do not contain any information which the client could conceivably wish to keep secret: one field is random, and all others are fixed. Information in server packet headers is likewise public: the origin timestamp is copied from the client's (random) transmit timestamp, and all other fields are set the same regardless of the identity of the client making the request.
Future extension fields could hypothetically contain sensitive information, in which case NTS provides a mechanism for encrypting them.
The authors would like to thank Richard Barnes, Steven Bellovin, Sharon Goldberg, Russ Housley, Martin Langer, Miroslav Lichvar, Aanchal Malhotra, Dave Mills, Danny Mayer, Karen O'Donoghue, Eric K. Rescorla, Stephen Roettger, Kurt Roeckx, Kyle Rose, Rich Salz, Brian Sniffen, Susan Sons, Douglas Stebila, Harlan Stenn, Martin Thomson, and Richard Welty for contributions to this document. on the design of NTS.
| [IEC.61588_2009] | IEEE/IEC, "Precision clock synchronization protocol for networked measurement and control systems", IEEE 1588-2008(E), IEC 61588:2009(E), DOI 10.1109/IEEESTD.2009.4839002, February 2009. |
| [Mizrahi] | Mizrahi, T., "A game theoretic analysis of delay attacks against time synchronization protocols", in Proceedings of Precision Clock Synchronization for Measurement Control and Communication, ISPCS 2012, pp. 1-6, September 2012. |
| [RFC4086] | Eastlake 3rd, D., Schiller, J. and S. Crocker, "Randomness Requirements for Security", BCP 106, RFC 4086, DOI 10.17487/RFC4086, June 2005. |
| [RFC5077] | Salowey, J., Zhou, H., Eronen, P. and H. Tschofenig, "Transport Layer Security (TLS) Session Resumption without Server-Side State", RFC 5077, DOI 10.17487/RFC5077, January 2008. |
| [RFC5280] | Cooper, D., Santesson, S., Farrell, S., Boeyen, S., Housley, R. and W. Polk, "Internet X.509 Public Key Infrastructure Certificate and Certificate Revocation List (CRL) Profile", RFC 5280, DOI 10.17487/RFC5280, May 2008. |
| [RFC5869] | Krawczyk, H. and P. Eronen, "HMAC-based Extract-and-Expand Key Derivation Function (HKDF)", RFC 5869, DOI 10.17487/RFC5869, May 2010. |
| [RFC7384] | Mizrahi, T., "Security Requirements of Time Protocols in Packet Switched Networks", RFC 7384, DOI 10.17487/RFC7384, October 2014. |
| [Shpiner] | "Multi-path Time Protocols", in Proceedings of IEEE International Symposium on Precision Clock Synchronization for Measurement, Control and Communication (ISPCS), September 2013. |