Time server – Using NTP for Synchronisation

A time server is a generic term for a device that receives and distributes time.  Time servers are available across the Internet or dedicated devices can be bought that offer higher accuracy and more security.

Whilst many manufacturers produce time server devices and other synchronisation tools they all have one thing in common in that they all utilise Network Time Protocol (NTP).

Whilst NTP is not the only time synchronisation software available it is used in the vast majority of synchronisation tools with some 99% of time synchronisation being conducted using NTP. NTP is a protocol used by everyone from NASA to NASDAQ and owes much of its dominance or the time synchronisation market to the dedicated team that continue to update, upgrade and develop NTP (many of whom do not get paid).

NTP is in fact almost as old as the Internet itself having been unveiled as early as 1979. It was developed by Professor David Mills and his team from Delaware University who continue to update and develop it alongside hundreds of dedicated hobbyists referred to as the Internet timekeepers.

NTP is now on version 4 and versions of NTP are included as standard with most operating systems allowing any Windows or Linux computer to run as a time server. However, for those serious about time synchronisation or wanting to keep a network secure there is no real substitute than a dedicated time server that can distribute time sent directly from an atomic clock using NTP.

Time Server – FAQS

What is a time server?

A time server is a hardware device that distributes a single time source amongst a network to ensure all computers and devices are synchronised

What is NTP?

Network Time Protocol is a set of software instructions designed to distribute time across computers. NTP uses a complex algorithm to work out d inaccuracies and compensates for them by advancing or retarding the system clock.

What is UTC?

Coordinated Universal Time (UTC) is the offical global timescale. It is 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) sometimes referred to as UT1.

How does a time server receive UTC?

UTC can be received by either tuning in to a national time and frequency broadcast on long-wave administered by national physics laboratories, although not every country has one. Alternatively as GPS satellites (Global Positioning System) all carry an onboard atomic clock which transmits timing information this can be used also as a timing source for network time servers.

Can’t I use the Internet as a timing source?

You can, and there are many sources of UTC time on the Internet but very few offer any useful accuracy and for those that do the distance away can cause a drop-off in precision. More importantly, Internet time sources are exterior to your firewall and therefore a port needs to be left open to ensure communication, this can be taken advantage of by malicious users. Also Internet timing 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 on location. While GPS signals are available everywhere on the planet the antennas do have to have a clear view of the sky to receive the signal, alternatively radio transmissions are limited in their availability although where available the signals can be received indoors.

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.

Time Server – security and legal protection

Dedicated time servers are used throughout the world to ensure computer networks are synchronised. Not only does a synchronised network offer more security but synchronisation using a dedicated time server can offer legal protection.

Because dedicated time servers use such a secure and reliable method of keeping time (atomic clock signals from either the GPS network or radio broadcasts) the time can be legally audited to few milliseconds. However, networks synchronised using an Internet timing source can not be audited and this can leave these systems open to abuse or even fraud.

Dedicated time servers are also based behind a firewall ensuring the system is protected from any intrusion while an internet based timing source has to have access through the firewall through an open port.

Any company that requires security in its network or protection from fraud must ensure they are using either a  GPS or radio referenced time server.

Time Server – The need for a reliable time source

A time server is a necessary part of any computer network. Time servers receive an authoritative time source that they then distribute amongst all devices on a network. Most time servers use a protocol such as NTP (Network Time Protocol) to distribute the single time source to all devices.

The accuracy of a time server depends entirely on the accuracy of the time source that it uses. Most networks require a UTC time source (coordinated universal time). UTC is a global timescale based on International atomic clock time (TAI) and used throughout the globe. Utilising UTC means that a computer network can be synchronised to the exact same time as a network on the other side of the globe allowing precise communications and time sensitive transactions to take place.

A time server can receive a UTC time source from only three places: the Internet, via a radio transmission or the GPS network.

The internet is by far the most commonly used source of UTC time. There are over a thousand time servers all claiming to relay UTC time. However, in surveys, less than half of these internet time sources are accurate and those that are, can be too far away to provide and real use as a timing source.

Using an internet timing source will also leave a computer network vulnerable. The hole left in the firewall to access the timing source can be used by malicious users and the built-in security method of NTP, authentication, can not ne utilised from across the Internet.

Fortunately, the two other methods of receiving UTC time offer both secure and reliable methods of receiving UTC time. The GPS network is available anywhere where there is a clear view of the sky whilst although not every country has a national time and frequency transmission, these longwave signals can be picked up by a radio referenced network time server from within a building.

The Time Server – References and Strata

A time server is a computer server that reads the time from an accurate clock and distributes this information to its clients across a computer network.

The reference used by a time server is normally a UTC (coordinated universal time) time source. UTC is a global time scale adopted all over the world and based on the time told by atomic clocks The most common source for UTC time is now the GPS system (global positioning system).

The most widely-used protocol for distributing and synchronising time is Network Time Protocol (NTP) which has been around almost as long as the Internet itself, having been developed in 1985 by Professor David Mills.

NTP receives the time from the time server and then checks or clocks on its networks to see if they need advancing or retreat.

To prevent overload of networks requesting timing information, NTP is hierarchical.  The term “stratum” is used to label the nearness to a stratum 0 server – that atomic clock. The higher the stratum number the further away the server is. A GPS time server is normally a stratum 1 device as it receives time from a stratum 0 device. However, stratum 2 and stratum 3 devices can still synchronise with each other.

Types of Dedicated Time Server

Dedicated time servers are used despite the numerous Internet time sources available. The advantages of a dedicated time server is that the device is behind a firewall so preventing any security issues, dedicated time servers can also guarantee accuracy that is just not possible over the Internet.

The dedicated time server checks the time stamp from the UTC source (Coordinated Universal Time) and uses the information to calculate if the network clocks are drifting and adds or subtracts a second to match. The time server will do this at set intervals. Most time servers utilise NTP (network time protocol) although other protocols do exist NTP is by far the most widespread.

For NTP to distribute the time it first needs a timing source and this is where the differences in dedicated time servers come in. As UTC is based on the time told by atomic clocks, it is an atomic clocks that is the source for a time servers timing reference. Atomic clocks are extremely expensive, large, and require a multitude of maintenance engineers. For this reason they are only to be found in large scale physics laboratories.

However, the time from an atomic clock can still be utilised by using either a radio clock attached to a time server or a GPS clock. A radio clock is just a simple radio receiver that picks-up the long-wave signal broadcast by several national physics laboratories. This UTC signal is not available everywhere and the broadcast can be blocked by buildings, mountains and bad weather etc.

The other type of dedicated time server utilise the signals sent by the GPS satellites (global positioning system). These signals are available literally everywhere on the globe although the only downside is that a GPS antenna needs to have a clear view of the sky and therefore ideally needs to be situated on a roof which can cause some problems if the server room is along way from the top of the building.

Some dedicated time servers can actually receive both signals; the radio and GPS. In doing so they can continue receiving time signals even if one signal fails or gets temporarily blocked.

Probably the most inexpensive time servers in the World!

Galleon Systems and their The NTS 4000 MSF S and the NTS 4000 GPS S are possibly the two least expensive time servers available anywhere.

They are simple to use NTP time servers receiving accurate time from the a radio time signal (MSF) or the GPS network and using NTP can provide this synchronised time via a Ethernet port to any computer, server, switch, etc that conforms to the Network Time Protocol V3. And unbelievably they can process 1,000 NTP requests per minute.

The identical looking units are attractive pieces of kit too:

Time Servers – Using the MSF Broadcast

Time servers are essential in keeping computer networks synchronised without them many time sensitive transactions would be impossible to conduct. Time servers work by receiving a signal timing reference and distribute it to all devices on a network using the protocol NTP (Network Time Protocol). The time signal used by most time servers comes from a UTC (Coordinated Universal Time) source. UTC is based on the time told by atomic clocks and is used globally, allowing computers from across the world to be synchronized to the same time reference.

There are three methods that time servers can receive the UTC signal from. Firstly, the internet, although unsecured and without any guarantees of accuracy Internet time references are only suitable for networks where precision and security are not a concern.

The second method is via the GPS network which is available everywhere but only where an antenna has a clear view of the sky.

The third and simplest method to receive an accurate and secure UTC time reference is to use the national time and frequency broadcasts. In the UK the National Physical Laboratory transmits the British signal from Anthorn in Cumbria. This signal, known as MSF, can be received in most places in the UK although local topography can interfere with the transmission.

To receive the MSF broadcast a radio referenced NTP time server is required. This will receive the radio transmission via 65 kHz in long wave and distribute it amongst the network.
A radio based NTP server usually consists of a rack-mountable time server, and an antenna, consisting of a ferrite bar inside a plastic enclosure, which receives the radio time and frequency broadcast. The antenna should always be mounted horizontally at a right angle toward the transmission for optimum signal strength.

Similar national timing transmissions are broadcast from other countries in the US the signal is referred to as WWVB and is broadcast by the NIST (National Institute for Standards and Technology) in Fort Collins, Colorado, other systems are broadcast in Frankfurt, Germany (DCF-77), Japan (JJY) and France ( TDF).

NTP Server – Why the Need for a Dedicated Time Server

Just as your computer is only as good as the software it is running, a time server is only as useful as the timing source it uses. Time servers are similar to other servers in that they are located on a network but the primary function of a time server is to supply timing information rather than storing or routing data.

For computers a time stamp is the only point of a reference a computer can use to establish a sequence of events. Timestamps are used in everything from sending an email to debugging a system and ensuring your network is secure.

Nearly all time servers rely on NTP (Network Time Protocol) to synchronise devices once a timing source is selected Whilst it is not the only protocol designed for synchronisation it is by far the most widely used, primarily because it has been under constant development since its inception nearly three decades ago.

A NTP server works by receiving a time reference from an authoritative source, normally a UTC time source. UTC (Coordinated Universal Time) is an International timescale based on the time told by atomic clocks.

Unfortunately many administrators opt to use an Internet timing source to run their time server with. However, several studies of Internet timing references discovered that nearly two thirds were inaccurate by over several seconds (almost an eternity in the world of NTP) and those that weren’t were often too far away from a client to provide reliable accuracy.

Probably the most accurate and widely used method of retrieving a reliable time reference is to use a dedicated GPS Time Server, many national physics laboratories also broadcast the time told by their atomic clocks via long wave radio. These signals can be received by a time server if it is equipped with a radio receiver although not every country broadcasts one and the transmissions are susceptible to interference and local topography.

Once this secure UTC time reference is received the time server can then distribute the time across the network. The timestamp itself is represented as a single ascending number. The number is the amount of seconds from a prime epoch (start number) normally 01 January 1900. The time server checks the system clocks on its clients and advances or retracts the time until all machines on the network are synchronised.

From the Sundial to the Time Server a Brief History of Horology

Keeping track of time is something most people take for granted, yet the science of timekeeping has a long and fascinating history.

Keeping track of time was always based on the relationship between the Earth, Moon and Sun. The first timekeeping devices are thought to be monuments like Stone Henge in the UK that would recognise the winter or summer solstice allowing early man to calculate when to plant crops.

Dividing the day up into hours and being able to keep track of them has proved more difficult to civilisations.

The first timing devices were sundials, obelisks and water clocks but it wasn’t until the development of mechanical clocks in the middle-ages that time-telling started to become more accurate.

Mechanical clocks continued to develop until the turn of the twentieth century when they were bettered in accuracy by electronic oscillators that would use the resonance of a crystal (often quartz) to keep a stable time.

While electronic clocks provided accuracy to within a second a day, the atomic clock that uses the resonance of an atom (in most cases caesium -133) and was developed in the 1950′s demonstrated millisecond accuracy – not losing a second in several thousands of years.

Now atomic clocks are approaching nano-second accuracy (one second every billion years) with new developments like strontium. The atomic clock has also made horologists realise that basing a time system on the movement of the Earth and celestial bodies is unreliable as the Earth slows and speeds up.

UTC (coordinated universal time) was developed to combat this by adding leap seconds to keep atomic time in line with GMT (Greenwich Meantime). Now computer networks all over the world can synchronise to UTC and atomic clocks by using a time server.

A time server will receive a time from an atomic clock source and synchronise an entire network to this time. Without time servers, atomic clocks and UTC, technologies such as satellite communication, the Internet and global trading would be near impossible.