Network Time Protocol Version 5

Network Time Protocol (NTP) is a protocol which enables computers to synchronize their clocks over network. Time is distributed from primary time servers to clients, which can be servers for other clients, and so on. Clients can use multiple servers simultaneously. NTPv5 is similar to NTPv4. The main differences are: The protocol specification (this document) doesn't describe any algorithms. NTPv5 drops support for the symmetric active, symmetric passive, broadcast, control, and private modes. Only the client and server modes are supported. Timestamps are clearly separated from values used as cookies. NTPv5 messages can be extended only with extension fields. The MAC field is wrapped in an extension field. Extension fields can be of any length, even indivisible by 4, but are padded to 4 octets. Extension fields specified for NTPv4 are compatible with NTPv5. NTPv5 adds support for other timescales than UTC. The NTP era number is exchanged in the protocol, which extends the unambiguous interval of the client from 136 years to about 35000 years. NTPv5 adds new measurement modes to provide clients with more accurate transmit timestamps.

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 when, and only when, they appear in all capitals, as shown here.

NTPv5 uses few different data types. They are all in the network order. Beside signed and unsigned integers, it has also three fixed-point types. A 16-bit fixed-point type contains values in seconds. It has 1 signed integer bit (i.e. it is just the sign) and 15 fractional bits. The minimum value is -1.0, the maximum value is 32767/32768, and the resolution is about 30 microseconds. A 32-bit fixed-point type contains values in seconds. It has 4 unsigned integer bits and 28 fractional bits. The maximum value is 16 seconds and the resolution is about 3.7 nanoseconds. A 64-bit fixed-point type is used for timestamps. It has 32 signed integer bits and 32 fractional bits. It spans an interval of about 136 years and has a resolution of about 0.23 nanoseconds. It can be used in different timescales. In the UTC timescale it is the number of SI seconds since 1 Jan 1972 plus 2272060800, excluding leap seconds. Timestamps in the TAI timescale are the same except they include leap seconds and extra 10 seconds for the original difference between TAI and UTC in 1972, when leap seconds were introduced.

NTPv5 servers and clients exchange messages as UDP datagrams. Clients send requests to servers and servers send them back responses. The format of the UDP payload is shown in Figure .

Each NTPv5 message has a header containing the following fields: A 2-bit field which can have the following values: 0 (normal), 1 (leap second inserted at the end of the month), 2 (leap second deleted at the end of the month), 3 (not synchronized). The values 1 and 2 are set at most 14 days in advance before the leap second. In requests it is always 0. A 3-bit field containing the value 5. A 3-bit field containing the value 3 (request) or 4 (response). A 4-bit identifier of the timescale. In requests it is the requested timescale. In responses it is the timescale of the receive and transmit timestamps. Defined values are: 0: UTC 1: TAI 2: UT1 3: Leap-smeared UTC A 4-bit field containing the stratum of host. Primary time servers have a stratum of 1, their clients have a stratum of 2, and so on. The value of 0 indicates an unknown or infinite stratum. In requests it is always 0. An 8-bit signed integer containing the polling interval as a rounded log2 value in seconds. In requests it is the current polling interval. In responses it is the minimum allowed polling interval. An 8-bit signed integer containing the precision of the timestamps included in the message as a rounded log2 value in seconds. In requests, which don't contain any timestamps, it is always 0. An 8-bit integer that can contain the following flags: In requests it is a request for a multi-message response. In responses it indicates the response may, but does not have to, have multiple messages. In responses it indicates it is not the first message of a multi-message response. In requests it is a request for a response in the interleaved mode. In responses it indicates the response is in the interleaved mode. An 8-bit unsigned NTP era number corresponding to the receive timestamp. In requests it is always 0. A 16-bit value specific to the selected timescale, which is referenced to the receive timestamp. In requests it is always 0. In the UTC (0) and TAI (1) timescales it is the TAI-UTC offset as a signed integer, or 0x8000 if unknown. In the UT1 timescale (2) it is the UT1-UTC offset as the 16-bit fixed-point number, or 0x8000 (-1.0) if unknown. In the leap-smeared UTC, it is the current offset between the leap smeared time and UTC as the 16-bit fixed-point number, or 0x8000 (-1.0) if unknown. A field using the 32-bit fixed-point type. In responses it is the server's root delay. In requests it is always 0. A field using the 32-bit fixed-point type. In responses it is the server's root dispersion. In requests it is always 0. A 32-bit field that enables follow-up messages and interleaved mode. It allows the server to verify that the client is actually receiving the responses. In requests it is either 0 or a copy of the server cookie from the last response. A 32-bit field containing a random number generated by the client. Responses contain a copy of the field from the corresponding request, which allows the client to verify that the responses arevalid responses to the requests. A field using the 64-bit fixed-point type. In requests it is always 0. In responses it is the time when the request was received. The timestamp corresponds to the end of the reception. A field using the 64-bit fixed-point type. In requests it is always 0. In responses it is the time when a response to the client was transmitted. The specific response depends on the selected mode (basic, follow-up, interleaved). The timestamp corresponds to the beginning of the transmission. The header has 48 octets, which is the minimum length of a valid NTPv5 message. A message can contain zero, one, or multiple extension fields. The maximum length is not specified, but the length is always divisible by 4.

The format of NTPv5 extension fields is shown in Figure .

Each extension field has a header which contains a 16-bit type and 16-bit length. The length is in octets and it includes the header. The minimum length is 4, i.e. an extension field doesn't have to contain any data. If the length is not divisible by 4, the extension field is padded with zeroes to 4 octets. Generally, if a request contains an extension field, the client is asking the server to include the same extension field in the response. Exceptions to this rule are allowed. Extension fields specified for NTPv4 can be included in NTPv5 messages as specified for NTPv4. The rest of this section describes new extension fields specified for NTPv5. Clients are not required to use or support any of these extension fields, but servers are required to support some extension fields.

This field is used by servers to pad the response to the same length as the request if the response doesn't contain all requested extension fields, or some have a variable length. It can have any length. This field MUST be supported on server.

This field authenticates the NTPv5 message with a symmetric key. Implementations SHOULD use the MAC specified in RFC8573. The extension field MUST be the last extension field in the message unless an extension field is specifically allowed to be placed after a MAC or another authenticator field.

This fields allows servers to prevent synchronization loops, i.e. synchronizing to one of its direct or indirect clients. It contains a set (bloom filter) of reference IDs. TODO This field MUST be supported on server.

This field allows switches and routers to make corrections in NTPv5 messages to allow clients to compensate for queueing and processing delays in the network. TODO (reuse draft-mlichvar-ntp-correction-field?)

This fields contains the time of the last update of the clock. It has a fixed length of 12 octets. In requests, the timestamp is always 0. (Is this really needed? It was mostly unused in NTPv4.)

This field contains an extra receive timestamp with a 32-bit step counter from a clock which doesn't have adjusted phase and can be used for a frequency transfer, e.g. to stabilize synchronization in a long chain of servers. It has a constant length of 16 octets. In requests, the counter and timestamp are always 0.

An NTPv5 client can use one or multiple servers. It has a separate association with each server. It makes periodic measurements of its offset and delay to the server. It can filter the measurements and compare measurements from different servers to select and combine the best servers for synchronization. It can adjust its clock in order to minimize its offset and keep the clock synchronized. These algorithms are not specified in this document. The polling interval can be adjusted for the network conditions and stability of the clock. When polling a public server on Internet, the client SHOULD use at least a polling interval of 64 seconds, increasing up to at least 1024 seconds. The client can make measurements in the following modes: basic, follow-up, interleaved. The follow-up and interleaved modes are two different mechanisms with different trade-offs that allow the server to provide the client with a more accurate transmit timestamp which is available only after the server forms or sends the response. Each successful measurement provides the client with an offset, delay and dispersion. When combined with the server's root delay and dispersion, it gives the client an estimate of the maximum error. On each poll, the client: Generates a new random cookie. Formats a request with necessary extension fields and the fields in the header all zero except: Version is set to 5. Mode is set to 3. Scale is set to the timescale in which the client wants to operate. Poll is set to the rounded log2 value of the current client's polling interval in seconds. Flags are set according to the requested mode. The Multi-message response flag enables the server to respond with multiple messages. The interleaved mode flag requests a response in the interleaved mode. Server cookie is set only in the follow-up and interleaved modes. If a valid response from the server was received previously, it is set to the Server Cookie from the previous response. Client cookie is set to the newly generated cookie. Sends the request to the server to the UDP port 123 and captures a transmit timestamp. Waits for a valid response from the server and captures a receive timestamp. A valid response has version 5, mode 4, client cookie equal to the cookie from the request, and passes authentication if enabled. The client MUST ignore all invalid responses. In the basic mode and interleaved mode it MUST accept at most one valid response. In the follow-up mode, if the first received response has the "Multi-message response" or "Follow-up message" flag set, the client SHOULD wait for some time (e.g. 10 milliseconds) and process all valid responses that have the same receive timestamp as the first received message. As the local receive timestamp, the client SHOULD use the timestamp of the message which didn't have the "Follow-up message" flag, or the message which it received first. Checks whether the response is usable for synchronization of the clock. Such a response has a leap indicator not equal to 3, stratum between 0 and 16, root delay and dispersion both smaller than a specific value, e.g. 16 seconds, and timescale equal to the requested timescale. If the response is in a different timescale, the client can switch to the provided timescale, convert the timestamps if the offset between the timescales is provided or known, or drop the response. Saves the server's receive and transmit timestamps. In the follow-up mode, if multiple messages were received, only the latest transmit timestamp is saved. If the client internally counts seconds using a type wider than 32 bits, it SHOULD expand the timestamps with the provided NTP era. Calculates the offset, delay, and dispersion: offset = ((T2 + T3) - (T4 + T1)) / 2 delay = |(T4 - T1) - (T3 - T2)| dispersion = |T4 - T1| * DR where T1 is the client's transmit timestamp T2 is the server's receive timestamp T3 is the server's transmit timestamp T4 is the client's receive timestamp DR is the client's estimated dispersion rate

A server receives requests on the UDP port 123. The server MUST support measurements in the basic mode. It MAY support the follow-up and/or interleaved modes. For the basic and follow-up modes the server doesn't need to keep any client-specific state. For the interleaved mode it needs to save transmit timestamps and be able to identify them by a cookie. The server maintains its leap indicator, stratum, root delay, and root dispersion: Leap indicator MUST be 3 if the clock is not synchronized or its maximum error cannot be estimated with the root delay and dispersion. Otherwise, it MUST be 0, 1, 2, depending on whether a leap second will be inserted or deleted in the next 14 days. Stratum SHOULD be one larger than stratum of the best server it uses for its own synchronization. Root delay SHOULD be the best server's root delay in addition to the measured delay to the server. Root dispersion SHOULD be the best server's root dispersion in addition to an estimate of the maximum drift of its own clock since the last update of the clock. The server has a randomly generated reference ID and it MUST track reference IDs of its servers using the Reference IDs Extension Field. For each received request, the server: Captures a receive timestamp. Checks the version in the request. If it is larger than 5, it MUST drop the request. If it is smaller than 5, it SHOULD either drop the request, or handle it according to the corresponding specification. Drops the request if the format is not valid, mode is not 3, or authentication fails if the MAC Extension Field or another authenticator field is present. Server forms a response with requested extension fields and sets the fields in the header as follows: Leap Indicator, Stratum, Root delay, and Root dispersion, are set to the current server's values. Version is set to 5. Scale is set to the client's requested timescale if it is supported by the server. If not, the server SHOULD respond in any timescale it supports. Flags are set according to the requested mode and supported mode. If the follow-up mode was requested, it is supported by the server, and the provided server cookie is valid, the response may have multiple messages. They have the same length and content except the flags, transmit timestamps (captured at different points of the first message's transmission), and MAC or authenticator field if present. The first message in the response has the Multi-message response flag set and all follow-up messages have the Follow-up flag set. If the interleaved mode is requested and a response in the interleaved mode is possible (i.e. a transmit timestamp is associated with the server cookie), the Interleaved mode flag is set. Era is set to the NTP era of the receive timestamp. Timescale Offset is set to the timescale-specific offset, or 0x8000 if unknown. Server Cookie is set when the follow-up or interleaved mode is requested and it is supported by the server, even if the response cannot be in the requested mode yet due to the request having an invalid server cookie. In the follow-up mode the cookie SHOULD be derived from the client's address and a secret key, which is rotated frequently (e.g. once per hour). It MUST prevent multi-message responses to addresses that didn't receive a previous response (other than by guessing the 64-bit number). In the interleaved mode, the cookie identifies a more accurate transmit timestamp, which can be retrieved by the client later with another request. Client Cookie is set to the Client Cookie from the request. Receive Timestamp is set to the server's receive timestamp of the request. Transmit Timestamp is set to a value which depends on the measurement mode. In the basic mode and the first message of a multi-message response in the follow-up mode, it is the server's current time when the message if formed. In the second and following messages in the follow-up mode it is a more accurate transmit timestamp of the first message captured after it was formed or sent. In the interleaved mode it is the transmit timestamp of the previous response identified by the server cookie in the request, captured at some point after the message was formed. Drops the response if it is longer than the request to prevent traffic amplification. Sends the response. In the follow-up mode, the follow-up messages are sent as soon as more accurate transmit timestamps are available. If the server can predict which transmit timestamps it will get, it SHOULD send only one follow-up message containing the most accurate timestamp. Saves the transmit timestamp and server cookie, if the interleaved mode was requested and is supported by the server.

Some ideas were taken from a different NTPv5 design proposed by Daniel Franke. The follow-up mode was inspired by PTP (IEEE 1588-2008).

This memo includes no request to IANA.