Time Synchronisation

NTP servers are a vital tool for any business that needs to communicate globally and securely. NTP servers distribute Coordinated Universal Time (UTC), the world’s global timescale based on the highly accurate time told by atomic clocks.

NTP (Network Time Protocol) is the protocol used to distribute the UTC time across a network it also ensures all time is accurate and stable. However, there are many pitfalls in setting up a NTP network, here are the most common:

Using the correct time source

Attaining the most suitable time source is fundamental in setting up a NTP network. The time source is going to be distributed amongst all machines and devices on a network so it is vital that it is not only accurate but also stable and secure.

Many system administrators cut corners with a time source. Some will decide to use an Internet based time source although these are not secure as the firewall will require an opening and also many internet sources are either wholly inaccurate or too far away to afford any useful precision.

There are two highly secure methods of receiving a UTC time source. The first is to utilise the GPS network which although doesn’t transmit UTC, GPS time is based on International atomic time and is therefore easy for NTP to convert. GPS time signals is also readily a

The second method is to use the long wave radio signals broadcast by some national physical laboratories. These signals, however, are not available in every country and they have a finite range and are susceptible to interference and local topography.

Organising Strata

Stratum levels describe the distance between a device and the reference clock. For instance an atomic clock based in a physics laboratory or GPS satellite is a stratum 0 device. A stratum 1 device is a time server that receives time from a stratum 0 device so any dedicated NTP server is stratum 1. Devices that receive the time from the time server such as computers and routers are stratum 2 devices.

NTP can support up to 16 stratum levels and although there is a drop-off in accuracy the further away you go stratum levels are designed to allow huge networks to all receive a time from a single NTP server without causing network congestion or a blockage in the bandwidth.

When using a NTP server it is important to not overload the device with time requests so the network should be divided with a select number of machines taking requests from the NTP server (the NTP server manufacturer can recommend the number of requests it can handle). These stratum 2 devices can ten be used as time references for other devices (which become stratum 3 devices) on very large networks these can then be used as time references themselves.

NTP GPS servers are now one of the most common ways to receive and distribute a UTC time source. This time is passed from the reference clock (stratum 0 device – the GPS satellite) to the NTP GPS server (stratum 1 device). Servers synchronised to the stratum 1 server will be stratum 2 devices.

The synchronisation of a client to a network server involves several packet exchanges where each exchange is a pair of request and reply. When sending out a request, the client stores its own original timestamp into the packet being sent. When a server receives the packet, it will in turn store its own timestamp into the packet, and the packet will be returned after putting a transmission timestamp into the packet. When receiving the reply, the receiver will once more log its own receipt time to estimate the travelling time of the packet. The travelling time (delay) is estimated to be half of “the total delay minus remote processing time”, assuming symmetrical delays.

The time is not trusted by the NTP GPS server until several packet exchanges have taken place, each passing a set of tests. Only if the replies from a server satisfy the predefined conditions in the protocol specification, the server is considered valid.

Usually it takes about five minutes (five good samples) until a NTP server is accepted as synchronisation source. Interestingly, this is also true for local reference clocks that have no delay at all by definition.

After initial synchronisation, the quality estimate of the client usually improves over time. As a client becomes more accurate, one or more potential servers may be considered invalid after some time.

Information courtesy of the NTP pool.

A NTP time server is simply a time server that uses the protocol NTP (Network Time Protocol) to distribute a single time source around a network. There are plenty of other protocols for distributing time across a network but NTP is by far the most widely used.

Time is essential for computer networks as it is the only reference a computer has to establish the sequence of events. When time differs on two machines on a network it does not become noticeable until either an event occurs that require the reconstruction of past events with respect to time or an event does not take place because it was dependent on other events that should have been executed in a particular time sequence.

A NTP time server will receive a UTC time source (coordinated universal time) from either across the Internet, from a radio transmission or via the GPS network. NTP will then distribute this time to all devices on its network. NTP will assess the drift and error rate of a system clock and advance or retreat it until it is matched to the UTC source.

NTP is also hierarchical so that devices that have received a time signal from the NTP time server can also distribute it themselves. NTP can support up to 16 levels in its hierarchy (known as stratum levels) although distance does deteriorate the accuracy.

Time synchronisation is now a critical aspect of network management enabling time sensitive applications to be conducted from across the globe. Without correct synchronisation computer systems would be unable to communicate with each other and transactions such as seat reservation, Internet auctions and online banking would be impossible.

For effective time synchronisation the global timescale UTC (Coordinated Universal Time) is a prerequisite. While a computer network can be synchronised to any single time source, UTC is employed by computer networks all over the world. By synchronising to a UTC time source a computer network can therefore be synchronised to every other computer network across the globe that also use UTC as their time source.

Receiving a reliable UTC time source is not as easy as it sounds. Many network administrators opt to use a UTC Internet time source. Whilst many of these time sources are accurate enough, they can be too far away to provide reliability and there are plenty of Internet time sources that are vastly inaccurate.

Another reason why Internet time sources should not be used as a source of time synchronisation is because an Internet time source is outside of a firewall and leaving a gap in the firewall to receive timing information can leave a system open to abuse.

So that UTC time can be opted as a civil time throughout the world several national physics laboratories broadcast a UTC timing signal that can be received and utilised as a network time source. Unfortunately, however, these time signals are not available in every country and even in those areas where a signal exists; they can be quite often obstructed by interference and local topography.

Another method for receiving a source of UTC time is to use the GPS satellite network. Strictly speaking the Global Positioning System (GPS ) does not relay UTC but it is a time based on International Atomic Time (TAI) with a predefined offset. A GPS NTP clock can simply convert the GPS time into UTC for synchronisation purposes.

The main advantage of using GPS is that a GPS signal is available anywhere on the planet providing that there is a clear view of the sky above (GPS transmissions are broadcast via line-of-sight) so UTC synchronisation can be conducted anywhere.

A network time server will synchronise all machines and devices on a network to UTC time (coordinated universal time). They are essential pieces of hardware in modern computer networks, without which, time sensitive transactions would be impossible.

Most network time servers will use NTP (Network Time Protocol) to distribute a single UTC time source (Coordinated Universal Time) around a network. The way a network time sever synchronises devices is by allowing them to ask it the time (poll) if the device is slow or fast compared to UTC then NTP will advance or retreat the system clock.

You would think that on larger networks with hundreds of machines all requesting timing information from a single network time server will cause congestion of the network and the time server. However, NTP is hierarchical, the distance from the time source is split into strata, the closer to the time source the lower the stratum level. So for example a stratum 0 device is an atomic clock while a dedicated network time server that receives the time from an atomic clock is a stratum 1 device while a computer that receives the time from a network time server is a stratum 2 device.

Fortunately stratum 2 devices can also be used to distribute time. Devices that receive time from a stratum 2 device become stratum 3 stratum device. These devices too can be used to distribute time to lower stratum levels, although the accuracy of the devices will lessen the further away from the original time source.

Arranging a network into stratum levels means that all devices can be synchronised using just the one network time server.

What is a network time server?

A network time server is a hardware device that utilises a single time source and distributes it amongst a network to ensure all computers and devices are telling the same time.

What is NTP?

Network Time Protocol is a protocol designed to distribute time across computers. NTP uses an algorithm to work out drift and inaccuracies and compensates for them ensuring all devices on a network are synchronised.

What is UTC?
Coordinated Universal Time (UTC) is a global timescale used by computer networks throughout the world. It is based on the time told by atomic clocks but compensates for the slowing of the Earth’s rotation by adding Leap Seconds, this also keeps it inline with GMT (Greenwich Meantime).

How does a network time server receive UTC?

UTC is broadcast by several national time and frequency transmissions on long-wave. Not every country has one and the signals are finite and vulnerable to interference. Alternatively as GPS satellites (Global Positioning System) all transmit timing information from their onboard atomic clocks this can be used also as a timing source for network time servers.

Can’t I use the Internet as a timing source?
There are many sources of UTC time on the Internet but many are not very accurate and for those that are the distance away can cause a drop-off in precision. More importantly, Internet time sources can’t be authenticated. Authentication is a security measure that ensures that a timing reference is what it says it is.

Which method is best for me a radio referenced time server or one that utilises the GPS network?

This depends primarily on where the network time server is to be situated. If a national radio signal is available and local topography doesn’t cause interference then a radio referenced network time server is probably the best option. However, if a radio signal is unavailable then the GPS network would be an obvious solution. While GPS signals are available everywhere on the planet they do have a downside in that their antennas have to have a clear view of the sky to receive the signal, this can be problematic  if a server room is on the ground floor of a skyscraper – although it is possible to receive a GPS signal via window.

How accurate are time servers?

A radio referenced network time server can provide UTC time to an accuracy of 100 microseconds, while the GPS network can fare even better with accuracies of a few milliseconds reasonably possible.

The NTP server is an essential network tool. Whilst other protocols do exist, NTP is by far the standard time synchronisation protocol and is utilised in the majority of time servers.

A NTP server is reliant on a single time source it is this time reference that it uses to distribute amongst the network and synchronise to. This timing reference tends to be a UTC time source (coordinated universal time) which is a global time source based on the time told by atomic clocks.

There are only two viable options for receiving a UTC timing source. Although the Internet can be used, the signal can’t be authenticated this is a security measure used by NTP to ensure the reference is what it says it is. Also by using an Internet time source a hole must be left open in the network firewall to allow for communication to the server, this has its own security risks.

The only two secure methods for receiving a UTC time signal is to either use the GPS network or national time and frequency transmissions that are broadcast by several countries’ national physics laboratories.

In selecting a timing source for a NTP server, location is the key consideration. The national time and frequency transmissions are not available in every country. Whilst the USA, UK, Germany, France, Japan and Finland have a signal there are many countries that do not. Furthermore being a long wave radio transmission it can easily be blocked by local topography, although the radio aerial can pick op a signal indoors which is something a GPS NTP server can’t do.

GPS antennas have to be situated on a roof. This can have logistical problems if the server room is in the basement of a high storey building but on the plus inside the GPS signal can be received literally anywhere in the world.

MSF is the name given to the dedicated time broadcast provided by the National Physical Laboratory in the UK, It is an accurate and reliable source of UK civil time, based on the time scale UTC (Coordinated Universal Time).

MSF is used throughout the UK and indeed other parts of Europe to receive a UTC time source which can be used by radio clocks and to synchronise computer networks by using a NTP time server.

It is available 24 hours a day across the whole of the UK although in some areas the signal can be weaker and it is susceptible to interference and local topography. The signal operates on a frequency of 60 kHz and carries a time and date code which relays the following information in binary format: Year, month, day of month,  day of week,  hour,  minute,  British Summer Time (in effect or imminent) and DUT1 (the difference between UTC and UT1 which is based on the Earths rotation)

The MSF signal is transmitted from Anthorn Radio Station in Cumbria but was only recently moved there after residing in Rugby, Warwickshire since it was started in the 1960’s. The signal’s carrier frequency is at 60 kHz, controlled by caesium atomic clocks at the radio station.

Caesium atomic clocks are the most reliably accurate atomic clocks anywhere, neither losing nor gaining a second in several millions of years.

To receive the MSF signal simple radio clocks can be used to display the exact UTC time or alternatively MSF referenced time servers can receive the long-wave transmission and distribute the timing information around computer networks using NTP (Network Time Protocol).

The only real alternative to the MSF signal in the UK is to use the onboard caesium clocks of the GPS network (Global Positioning System) that relay accurate time information that can be used as a UTC time source.

The GPS (Global Positioning System) is a Global Navigational Satellite System (GNSS) controlled and run by the United States of America.

GNSS systems work by using satellites several thousand miles above the Earth’s surface that beam timing information down to a GNSS receiver (like the satellite navigation unit in our cars). It is this information that is used by the GPS receiver to triangulate an exact position. They can only do this by having onboard their own highly accurate atomic clock as the distance the satellites are away from the Earth, even an inaccuracy of a second or two could mean a sat navigation’s location could be miles out.

As a consequence of having this accurate time sources, GPS and the new breed of GNSS systems can all be used to receive an absolute or UTC (Universal Coordinated Time) time source. This time source can be used by computer networks running a NTP server (Network Time Protocol) to synchronise all machines and devices to the same time.

NTP is a protocol designed to synchronise computers and network devices to an external timing reference.

GPS is an ideal time and frequency reference because it can provide highly accurate time anywhere in the world using relatively cheap components. Each GPS satellite transmits in two frequencies L2 for the military use and L1 for use by civilians transmitted at 1575 MHz, Low-cost GPS antennas and receivers are now widely available and dedicated NTP GPS servers are now relatively low cost.

The radio signal transmitted by the satellite can pass through windows but can be blocked by buildings so the ideal location for a GPS antenna is on a rooftop with a good view of the sky. The more satellites it can receive from the better the signal. However, roof-mounted antennas can be prone to lighting strikes or other voltage surges so a suppressor is highly recommend being installed inline on the GPS cable.

A NTP GPS Server is ideal in providing NTP time servers or stand-alone computers with a highly accurate external reference for synchronisation. Even with relatively low cost equipment, accuracy of hundred nanoseconds (a nanosecond = a billionth of a second) can be reasonably achieved using GPS as an external reference.

Computer networking can seem an intimidating undertaking. However, a computer network is really just a number of machines connected together for ease of data transfer and security. They can be very small such as two computers in a home network to really large networks consisting of hundreds and thousands of machines.

When a computer or device is connected to a network then there is only one point of reference that the computers can use to establish the order of events and applications and that is time.

Time, in the form of time stamps are used by most applications and this is when problems in computer networks can occur.

Computers tell the time by using a software clock. This is based on a system clock that keeps time when the computer is off. However, computers internal clocks are wholly inaccurate. They tend to drift up to several seconds a week. On a network when there is more than one machine, this can cause severe problems if the machines are drifting at different rates.

Emails may arrive before they have been sent and the whole network can be vulnerable to security threats and even fraud!

A network time server is used to synchronize a computer network to a single time source. This time source can be anything from an internal clock on a computer to the time told by a wrist watch. However, to ensure perfect accuracy and to keep a network synchronized to the rest of the world then a UTC time source should be used.

UTC (Coordinated Universal Time) is a global timescale based on the time told by atomic clocks. A network time server can receive a UTC time source from across the Internet (although unsecured), via the GPS (global positioning system) network or via specialist radio transmission from national physics laboratories.

Most network time servers use NTP (Network Time Protocol) to distribute the timing reference throughout the network. NTP is not the only timing protocol designed to do this although it is, however, by far the most widely used.