The Network Time Protocol (NTP) synchronizes time and coordinates time distribution throughout a TCP/IP network. NTP provides accurate and dependable timekeeping for hosts on TCP/IP networks. TCP/IP Services NTP software is an implementation of the NTP Version 4 specification and maintains compatibility with NTP Versions 1, 2, and 3.

Time synchronization is important in client/server computing. For example, systems that share common databases require coordinated transaction processing and timestamping of instrumental data.

NTP provides synchronization that is traceable to clocks of high absolute accuracy and avoids synchronization to clocks that keep incorrect time.

This chapter reviews key concepts and describes:

• How to start up and shut down NTP ( Section 13.3)
• How to configure the NTP host ( Section 13.4)
• How to configure the host as a backup time server ( Section 13.5)
• NTP event logging ( Section 13.6)
• How to configure NTP authentication ( Section 13.7)
• How to use NTP utilities ( Section 13.8)
• How to solve NTP problems ( Section 13.10)

Synchronized timekeeping means that hosts with accurate system timestamps send time quotes to each other. Hosts that run NTP can be either time servers or clients, although they are often both servers and clients.

NTP does not attempt to synchronize clocks to each other. Rather, each server attempts to synchronize to Coordinated Universal Time (UTC) using the best available source and the best available transmission paths to that source. NTP expects that the time being distributed from the root of the synchronization subnet will be derived from some external source of UTC (for example, a radio clock).

If your network is isolated and you cannot access other NTP servers on the internet, you can designate one of your nodes as the reference clock to which all other hosts will synchronize.

13.1.1 Time Distributed Through a Hierarchy of Servers

In the NTP environment, time is distributed through a hierarchy of NTP time servers. Each server adopts a stratum that indicates how far away it is operating from an external source of UTC. NTP times are an offset of UTC. Stratum 1 servers have access to an external time source, usually a radio clock. A stratum 2 server is one that is currently obtaining time from a stratum 1 server; a stratum 3 server gets its time from a stratum 2 server; and so on. To avoid long-lived synchronization loops, the number of strata is limited to 15.

Stratum 2 (and higher) hosts might be company or campus servers that obtain time from some number of primary servers and provide time to many local clients. In general:

• Lower-strata hosts act as time servers.
• Higher-strata hosts are clients that adjust their time clocks according to the servers.

Internet time servers are usually stratum 1 servers. Other hosts connected to an internet time server have stratum numbers of 2 or higher and can act as time servers for other hosts on the network. Clients usually choose one of the lowest accessible stratum servers from which to synchronize.

13.1.2 How Hosts Negotiate Synchronization

The identifying stratum number of each host is encoded within UDP datagrams. Peers communicate by exchanging these timestamped UDP datagrams. NTP uses these exchanges to construct a list of possible synchronization sources, then sorts them according to stratum and synchronization distance. Peers are accepted or rejected, leaving only the most accurate and precise sources.

NTP evaluates any new peer to determine whether it qualifies as a new (more suitable) synchronization source.

NTP rejects the peer under the following conditions:

• The peer is not synchronized.
• The stratum is higher than the current source's stratum.
• The peer is synchronized to the local node.

NTP accepts the peer under the following conditions:

• There is no current time source.
• The current source is unreachable.
• The current source is not synchronized.
• The new source's stratum is lower than the current source.
• The new source's stratum is the same as the current source, but its distance is closer to the synchronization source by more than 50 percent.

The OpenVMS system clock is maintained as a software timer with a resolution of 100 nanoseconds, updated at 10-millisecond intervals. A clock update is triggered when a register, loaded with a predefined value, has decremented to zero. Upon reaching zero, an interrupt is triggered that reloads the register, and the process is repeated.

The smaller the value loaded into this register, the more quickly the register reaches zero and triggers an update. Consequently, the clock runs more quickly. A larger value means more time between updates; therefore, the clock runs more slowly. A clock tick is the amount of time between clock updates.

13.1.4 How NTP Makes Adjustments to System Time

Once NTP has selected a suitable synchronization source, NTP compares the source's time with that of the local clock. If NTP determines that the local clock is running ahead of or behind the synchronization source, NTP uses a general drift mechanism to slow down or speed up the clock as needed. NTP accomplishes this by issuing a series of new clock ticks. For example, if NTP detects that the local clock is drifting ahead by +0.1884338 second, it issues a series of new ticks to reduce the difference between the synchronization source and the local clock.

If the local system time is not reasonably correct, NTP does not set the local clock. For example, if the new time is more than 1000 seconds off in either direction, NTP does not set the clock. In this case, NTP logs the error and shuts down.

NTP maintains a record of the resets it makes along with informational messages in the NTP log file, TCPIP$NTP_RUN.LOG. For details about event logging and for help interpreting an NTP log file, see Section 13.6.

13.1.5 Configuring the Local Host

The system manager of the local host, determines which network hosts to use for synchronization and populates an NTP configuration file with a list of the participating hosts.

NTP hosts can be configured in any of the following modes:

• Client/server mode
  This mode indicates that the local host wants to obtain time from the remote server and will supply time to the remote server. This mode is appropriate in configurations involving a number of redundant time servers interconnected through diverse network paths. Internet time servers generally use this mode.
  Indicate this mode with a peer statement in the configuration file, as shown in the following example:

  peer 18.72.0.3

• Client mode
  This mode indicates that the local host wants to obtain time from the remote server but will not provide time to the remote server. Client mode is appropriate for file server and workstation clients that do not provide synchronization to other local clients. A host with higher stratum generally uses this mode.
  Indicate client mode with the server statement in the configuration file, as shown in the following example:

  server 18.72.0.3

• Broadcast mode
  This mode indicates that the local server will send periodic broadcast messages to a client population at the broadcast/multicast address specified. This specification normally applies to the local server operating as a sender.
  Indicate this mode with a broadcast statement in the configuration file, as shown in the following example:

  broadcast 18.72.0.255

• Multicast mode
  A multicast client is configured using the broadcast statement, but with a multicast group (class D) address instead of a local subnet broadcast address. However, there is a subtle difference between broadcasting and multicasting. Broadcasting is specific to each interface and local subnet address. If more than one interface is attached to a machine, a separate broadcast statement applies to each one.
  IP multicasting is a different paradigm. A multicast message has the same format as a broadcast message and is configured with the same broadcast statement, but with a multicast group address instead of a local subnet address. By design, multicast messages travel from the sender via a shortest-path or shared tree to the receivers, which might require these messages to emit from one or all interfaces but to carry a common source address. However, it is possible to configure multiple multicast group addresses using multiple broadcast statements. Other than these differences, multicast messages are processed just like broadcast messages. Note that the calibration feature in broadcast mode is extremely important, since IP multicast messages can travel far different paths through the IP routing fabric than can ordinary IP unicast messages.
  The Internet Assigned Number Association (IANA) has assigned multicast group address 224.0.1.1 to NTP, but you should use this address only where the multicast span can be reliably constrained to protect neighbor networks. In general, you should use group addresses that have been given out by your administrator, as described in RFC 2365, or GLOP group addresses, as described in RFC 2770.
• Manycast mode
  Manycasting is an automatic discovery and configuration paradigm new to NTP Version 4. See Section 13.2.
• Iburst mode
  The iburst keyword can be configured when it is important to set the clock quickly, such as when an association is either first mobilized or first becomes reachable, or when the network attachment requires an initial calling or training procedure. The burst is initiated only when the server first becomes reachable. It results in good accuracy with intermittent connections typical of PPP and ISDN services. Outlyers caused by initial dialup delays and other factors are avoided, and the client sets the clock within 30 seconds after the first message.
  The burst keyword can be configured in cases of excessive network jitter or when the network attachment requires an initial calling or training procedure. The burst is initiated at each poll interval when the server is reachable. The burst does produce additional network overhead and can cause trouble if used indiscriminately. It should be used only if the poll interval is expected to settle to values equal to or greater than 1024 seconds.

Manycasting is an automatic discovery and configuration paradigm new to NTP Version 4. It is intended as a means for a client to survey the nearby network neighborhood to find cooperating servers, validate them using cryptographic means and evaluate their time values with respect to other servers that might be lurking in the vicinity. The intended result is that each client mobilizes associations with a given number of the "best" nearby servers, yet automatically reconfigures to sustain this number of servers should one or another fail.

Manycasting can be used with either symmetric key or public key cryptography. Public key cryptography offers the best protection against compromised keys and is generally considered stronger. By default, either of these two means is required, but this can be overridden by the disable auth command.

A manycast client association is configured using the manycastclient configuration command, which is similar to the server configuration command, but with a broadcast or multicast address. Depending on address family, the manycast client sends ordinary client mode messages, but with a broadcast address rather than a unicast address. It sends only if less than a given threshold of servers have been found and then only at the minimum feasible rate and minimum feasible time-to-live (TTL) hops. There can be as many manycast client associations as different broadcast addresses, each one serving as a template for a future unicast client/server association.

Manycast servers configured with the manycastserver command listen on the specified broadcast address for manycast client messages. If a manycast server is in scope of the current TTL and is itself synchronized to a valid source and operating at a stratum level equal to or lower than the manycast client, it replies to the manycast client message with an ordinary unicast server message.

The manycast client receiving this message mobilizes a preemptable client association according to the matching manycast client template, but only if cryptographically authenticated and the server stratum is less than or equal to the client stratum. The client runs the NTP mitigation algorithms, which act to demobilize all but a threshold number of associations according to stratum and synchronization distance. The surviving associations then continue in ordinary client/server mode.

If for some reason the number of available servers falls below the threshold, the manycast client resumes sending broadcast messages. The polling strategy is designed to reduce as much as possible the volume of broadcast messages and the effects of implosion due to near-simultaneous arrival of manycast server messages. The strategy is determined by the tos and ttl configuration commands described below.

It is possible and frequently useful to configure a host as both manycast client and manycast server. A number of hosts configured this way and sharing a common group address will automatically organize themselves in an optimum configuration based on stratum and synchronization distance.

For example, consider an NTP subnet of two primary servers and several secondary servers and a number of dependent clients. All servers and clients have identical configuration files including both multicastclient and multicastserver commands using, for instance, multicast group address 239.1.1.1. Each primary server configuration file must include commands for the primary reference source such as a GPS receiver.

The remaining configuration files for all secondary servers and clients have the same contents, except for the tos command, which is specific for each stratum level. For stratum 1 and stratum 2 servers, that command is not necessary. For stratum 3 and above servers the tos floor value is set to the intended stratum number. Thus, all stratum 3 configuration files use tos floor 3 , all stratum 4 files use tos floor 4 , and so forth.

Once operations have stabilized, the primary servers will find the primary reference source and each other, because they both operate at the same stratum (1), but not with any secondary server or client, since these operate at a higher stratum. The secondary servers will find the servers at the same stratum level. If one of the primary servers loses its GPS receiver, it will continue to operate as a client and other clients will time out the corresponding association and re-associate accordingly.

13.2.1 Manycast Options

Following are options that can be used with manycast.

• tos [ ceiling ceiling | cohort {0 | 1} | floor floor | minclock minclock | minsane minsane ]
  This command affects the clock selection and clustering algorithms. It can be used to select the quality and quantity of peers used to synchronize the system clock and is most useful in manycast mode.
  The variables operate as follows:
  • ceiling ceiling
    Servers with stratum at or above ceiling will be discarded if there are at least minclock peers remaining. This value defaults to 15, but can be changed to any number from 1 to 15.
  • cohort 0 | 1
    This is a binary flag which enables (0) or disables (1) manycast server replies to manycast clients with the same stratum level. This is useful to reduce implosions where large numbers of clients with the same stratum level are present. The default is to enable these replies.
  • floor floor
    Peers with strata below floor will be discarded if there are at least minclock peers remaining. This value defaults to 1, but can be changed to any number from 1 to 15.
  • minclock minclock
    The clustering algorithm repeatedly casts out outlyer associations until no more than minclock associations remain. This value defaults to 3, but can be changed to any number from 1 to the number of configured sources.
  • minsane minsane
    This is the minimum number of candidates available to the clock selection algorithm in order to produce one or more truechimers for the clustering algorithm. If fewer than this number are available, the clock is undisciplined and allowed to run free. The default is 1 for legacy purposes. However, according to principles of Byzantine agreement, minsane should be at least 4 in order to detect and discard a single falseticker.
• ttl hop...
  This command specifies a list of TTL values in increasing order. Up to 8 values can be specified. In manycast mode these values are used in turn in an expanding-ring search. The default is eight multiples of 32 starting at 31.

The NTP service can be shut down and started independently of TCP/IP Services. The following files are provided:

• SYS$STARTUP:TCPIP$NTP_STARTUP.COM allows you to start the NTP service.
• SYS$STARTUP:TCPIP$NTP_SHUTDOWN.COM allows you to shut down the NTP service.

To preserve site-specific parameter settings and commands, create the following files. These files are not overwritten when you reinstall TCP/IP Services:

• SYS$STARTUP:TCPIP$NTP_SYSTARTUP.COM can be used as a repository for site-specific definitions and parameters to be invoked when the NTP service is started.
• SYS$STARTUP:TCPIP$NTP_SYSHUTDOWN.COM can be used as a repository for site-specific definitions and parameters to be invoked when the NTP service is shut down.

The NTP configuration file TCPIP$NTP.CONF contains a list of hosts your system will use for time synchronization. Before configuring your host, you must do the following:

1. Select time sources.
2. Obtain the IP addresses or host names of the time sources.
3. Obtain the version number of NTP that the hosts are running.

To ensure reliable synchronization, select multiple time sources that you are certain provide accurate time and that are synchronized to an Internet time server.

To minimize common points of failure, avoid synchronizing the following:

• The local host to another peer at the same stratum, unless the latter is receiving time from a lower stratum source to which the local host cannot connect.
• More than one host in a particular administrative domain to the same time server outside that domain.

To simplify configuration file maintenance, avoid configuring peer associations with higher-stratum servers.

13.4.1 Creating the Configuration File

To create a configuration file for your local host, edit a copy of the file TCPIP$NTP.TEMPLATE (located in SYS$SPECIFIC:[TCPIP$NTP]) to add the names of participating hosts, then save the file as SYS$SPECIFIC:[TCPIP$NTP]TCPIP$NTP.CONF. This file is not overwritten when you install subsequent versions of TCP/IP Services.

The various modes are determined by the command keyword and the required IP address. Addresses are classed by type as (s) , a remote server, or peer (IPv4 class A, B and C); (b) the broadcast address of a local interface; (m) a multicast address (IPv4 class D); or (r) a reference clock address (127.127.x.x).

If IPv6 is enabled on the system, support for the IPv6 address family is generated in addition to the default support of the IPv4 address family. IPv6 addresses can be identified by the presence of colons (:) in the address field. IPv6 addresses can be used nearly everywhere that IPv4 addresses can be used, with the exception of reference clock addresses, which are always IPv4. Note that in contexts where a host name is expected, a -4 qualifier preceding the host name forces DNS resolution to the IPv4 namespace, while a -6 qualifier forces DNS resolution to the IPv6 namespace.

There are three types of associations: persistent, preemptable and ephemeral. Persistent associations are mobilized by a configuration command and never demobilized. Preemptable associations, which are new to NTPv4, are mobilized by a configuration command which includes the prempt flag and are demobilized by timeout or error. Ephemeral associations are mobilized upon arrival of designated messages and demobilized by timeout or error.

The following four commands specify the time server name or address to be used and the mode in which to operate. The address can be either a DNS name or an IP address in dotted-quad notation. Additional information on association behavior can be found in Section 13.1.5.

peer address [options ...] server address [options ...] broadcast address [options ...] manycastclient address [options ...]

Following are options that can be used with these commands:

• peer address [key ID | autokey] [version number] [prefer] [minpoll interval] [maxpoll interval]
  server address [key ID | autokey] [version number] [prefer ][burst] [iburst] [minpoll interval] [maxpoll interval]
  broadcast address [key ID | autokey] [version number][minpoll interval][ttl nn]
  manycastclient address [key ID | autokey] [version number][[minpoll interval] [maxpoll interval][ttl nn]
  These four statements specify the time server name or address to be used and the mode in which to operate. The address can be either a DNS name or an IP address.
  • peer --- For type s addresses only, this statement mobilizes a persistent symmetric-active mode association with the specified remote peer. This statement should not be used for type b , type m , or type r addresses.
  • server --- For type s and type r addresses only. This statement mobilizes a persistent client mode association with the specified remote server or local reference clock. This statement should not be used for type b or type m addresses.
  • broadcast --- For type b and type m addresses only. This statement mobilizes a persisent broadcast mode association. Multiple statements can be used to specify multiple local broadcast interfaces (subnets) or multiple multicast groups. Note that local broadcast messages go only to the interface associated with the subnet specified, but multicast messages go to all interfaces.
  • manycastclient --- For type m addresses only. This statement mobilizes a manycast client mode association for the multicast address specified. In this case, a specific address must be supplied that matches the address used on the manycastserver statement for the designated manycast servers.
    The manycastclient statement specifies that the local server is to operate in client mode with the remote servers that are discovered as the result of broadcast/multicast messages. The client broadcasts a request message to the group address associated with the specified address and specifically enabled servers respond to these messages. The client selects the servers providing the best time and continues as with the server statement. The remaining servers are discarded as if never heard.
  
  The following table describes the options to the NTP configuration statements:

  Option	Description
  autokey	All packets sent to and received from the server or peer are to include authentication fields encrypted using the autokey scheme described in Section 13.7. This option is valid with all commands.
  key ID	For all packets sent to the address, includes authentication fields encrypted using the specified key identifier, an unsigned 32-bit integer. The default is no encryption.
  version number	Specifies the version number to be used for outgoing NTP packets. Versions 1, 2, 3, and 4 are the choices. The default is 4.
  prefer	Marks the server as preferred. This host will be chosen for synchronization from a set of correctly operating hosts.
  burst	When the server is reachable, send a burst of eight packets instead of the usual one. The packet spacing is normally 2 s; however, the spacing between the first and second packets can be changed with the calldelay command to allow additional time for a modem or ISDN call to complete. This option is valid with only the server command and is a recommended option with this command when the maxpoll option is 11 or greater.
  iburst	When the server is unreachable, send a burst of eight packets instead of the usual one. The packet spacing is normally 2 s; however, the spacing between the first and second packets can be changed with the calldelay command to allow additional time for a modem or ISDN call to complete. This option is valid with only the server command and is a recommended option with this command.
  noselect	Marks the server as unused, except for display purposes. The server is discarded by the selection algorithm. This option is valid only with the server and peer commands.
  minpoll interval	Specifies the minimum polling interval for NTP messages, in seconds to the power of 2. The allowable range is 4 (16 seconds) to 14 (16384 seconds), inclusive. This option is not applicable to reference clocks. The default is 6 (64 seconds).
  maxpoll interval	Specifies the maximum polling interval (in seconds), for NTP messages. The allowable range is 4 (16 seconds) to 14 (16384 seconds) inclusive. The default is 10 (1024 seconds). This option does not apply to reference clocks.
  ttl nn	Specifies the time-to-live for multicast packets. Used only with broadcast and manycast modes.

• broadcastclient [novolley]
  This statement enables reception of broadcast server messages to any local interface (type b) address. Ordinarily, upon receiving a message for the first time, the broadcast client measures the nominal server propagation delay using a brief client/server exchange with the server, after which it continues in listen-only mode. If the novolley keyword is present, the exchange is not used and the value specified in the broadcastdelay command is used or, if the broadcastdelay command is not used, the default 4.0 ms. Note that, in order to avoid accidental or malicious disruption in this mode, both the server and client should operate using symmetric key or public key authentication as described in Section 13.7. Note that the novolley keyword is incompatible with public key authentication.
• broadcastdelay seconds
  The broadcast and multicast modes require a special calibration to determine the network delay between the local and remote servers. Usually, this is done automatically by the initial protocol exchanges between the client and server. In some cases, the calibration procedure fails because of network or server access controls. This statement specifies the default delay to be used under these circumstances. Typically (for Ethernet), a number between 0.003 and 0.007 second is appropriate. When this statement is not used, the default is 0.004 second.
• calldelay delay
  This option controls the delay in seconds between the first and second packets sent in burst or iburst mode to allow additional time for a modem or ISDN call to complete.
• multicastclient address
  This statement enables reception of multicast server messages to the multicast group address(es) (type m) specified. Upon receiving a message for the first time, the multicast client measures the nominal server propagation delay using a brief client/server exchange with the server, then enters the broadcast client mode, in which it synchronizes to succeeding multicast messages. Note that, in order to avoid accidental or malicious disruption in this mode, both the server and client should operate using symmetric key or public key authentication as described in Section 13.7.
• manycastserver address
  This statement enables reception of manycast client messages to the multicast group addresses (type m ) specified. At least one address is required. The Internet Assigned Number Association (IANA) has assigned multicast group address 224.0.1.1 to NTP, but you should use this address only where the multicast span can be reliably constrained to protect neighbor networks. In general, you should use group addresses that have been given out by your administrator, as described in RFC 2365, or GLOP group addresses, as described in RFC 2770. Note that to avoid accidental or malicious disruption in this mode, both the server and client should use authentication and the same trusted key and key identifier.
• driftfile file-specification
  This statement specifies the name of the file used to record the frequency offset of the local clock oscillator. If the file exists, it is read at startup to set the initial frequency offset, and then is updated hourly with the current frequency computed by the NTP server.
  If the file does not exist or if the driftfile statement is not specified in the configuration file, the initial frequency offset is assumed to be zero. If the file does not exist but the driftfile keyword is specified without a parameter, the default, SYS$SPECIFIC:[TCPIP$NTP]TCPIP$NTP.DRIFT is used.
  In these cases, it might take some hours for the frequency to stabilize and for the residual timing errors to subside.
  The drift file TCPIP$NTP.DRIFT consists of a single floating-point number that records the frequency of the offset measured in parts per million (ppm).
• enable auth | bclient | monitor | ntp | stats
  disable auth | bclient | monitor | ntp | stats
  These statements enable and disable the following server options:

  auth	Controls synchronization with unconfigured peers only if the peer has been correctly authenticated using a trusted key and key identifier. By default, auth is enabled.
  bclient	Controls the server to listen for messages from broadcast or multicast servers. By default, bclient is disabled.
  monitor	Controls the monitoring facility. By default, monitor is enabled.
  ntp	Enables the server to adjust its local clock by means of NTP. If disabled, the local clock free runs at its intrinsic time and frequency offset. This statement is useful in case the local clock is controlled by some other device or protocol and NTP is used only to provide synchronization to other clients. In this case, the local clock driver can be used to provide this function and also certain time variables for error estimates and leap indicators. The default for this flag is enable.
  stats	Enables the statistics facility. By default, stats is enabled.

• logconfig configkeyword
  This statement controls the amount and type of output written to the system log file. By default, all output is turned off. All configkeyword keywords can be prefixed with a plus sign (+) to add messages or a minus sign (-) to remove messages. Messages can be controlled in four classes ( clock , peer , sys , and sync ). Within these classes, four types of messages can be controlled. Informational messages ( info ) control configuration information. Event messages ( events ) control logging of events (reachability, synchronization, alarm conditions). Statistics messages ( statistics ) control statistical output. The final message group is the status ( status ) messages. This message group describes mainly the synchronization status.
  Configuration keywords are formed by concatenating the message class with the event class. The all prefix can be used instead of a message class. A message class can also be followed by the all keyword to enable or disable all messages of the respective message class. Therefore, a minimal log configuration might look like the following example:

  ---

  Previous

  Next

  Contents

  Index