Time Synchronisation

The WWVB time signal is a dedicated radio broadcast providing an accurate and reliable source of United States civil time, based on the global time scale UTC (Coordinated Universal Time), the WWVB signal is broadcast and maintained by the United States’ NIST laboratory (National Institute for Standards and Time).

The WWVB time signal can be utilised by anyone requiring accurate timing information although its main use is as a source of UTC time for administrators synchronising a computer network with a radio clock. Radio clocks are really another term for a network time server that utilises a radio transmission as a timing source.

Most radio based network time servers use NTP (Network Time Protocol) to distribute the timing information throughout the network.

The WWVB signal is broadcast from Fort Collins, Colorado. It is available 24 hours a day across most of the USA and Canada, although the signal is vulnerable to interference and local topography. Users of the WWVB service receive predominantly a ‘ground wave’ signal. However, there is also a residual ‘sky wave’ which is reflected off the ionosphere and is much stronger at night; this can result in a total received signal that is either stronger or weaker.

The WWVB signal is carried on a frequency of 60 kHz (to within 2 parts in 1012) and is controlled by a caesium atomic clock based at NIST

The signal’s field strength exceeds 100 µV/m (microvolts a meter) at a distance of 1000 km from Colorado – covering much of the USA.

The WWVB signal is in the form of a simple binary code containing time and date information The WWVB  time and date code includes the following information: year, month, day of month,  day of week,  hour, minute, Summer Time (in effect or imminent).

NTP (Network Time Protocol) is the most flexible, accurate and popular method of sending time over the Internet. It is perhaps the Internet’s oldest protocol having been around in one form or another since the mid 1980’s.

The main purpose of NTP is to ensure that all devices on a network are synchronised to the same time and to compensate for some network time delays. Across a LAN or WAN NTP manages to maintain an accuracy of a few milliseconds (Across the Internet, time transfer if far less accurate due to network traffic and distance).

NTP is by far the most widely used time synchronisation protocol (somewhere in the region of 95% of all time servers use NTP) and it owes much of its success to its continual updates and its flexibility. NTP will run on UNIX, LINUX, and Windows based operating systems (it is also free, another possible reason for its huge success).

NTP uses a single time source that it distributes among all devices on a network; it also checks each device for drift (the gaining or losing of time) and adjusts for each.  It is also hierarchical in that literally thousands of machines can be controlled using just one NTP server as each machine can in itself be used by neighbouring machines as a time server.

NTP is also highly secure (when using an external time reference not when using the Internet for a timing source) with an authentication protocol able to establish exactly where a timing source comes from.

For a network to be really effective most NTP time servers use an atomic clock as a basis for their time synchronisation. An international timescale based on the time told by atomic clocks has been developed for this very purpose. UTC (Coordinated Universal Time).

There are really two methods to receive a secure UTC atomic clock time signal to be utilised by NTP. The first being the time and frequency transmissions that several national physics laboratories broadcast on long wave around the world; the second (and by far the most readily available) is by using the timing information in the GPS satellite transmissions. These can be picked up anywhere on the globe and provide safe, secure and highly accurate timing information.

Time synchronisation can be conducted in multiple ways on a computer network. There are many time protocols but NTP (Network Time Protocol) is by far the most used (probably close to 99% of networks use NTP).

Most time servers receive a UTC time signal UTC (Coordinated Universal Time) yet, there are multiple places that NTP can receive a signal from but each has it downsides. The most widely used source of UTC is the Internet but no serious network administrator would dare to use the Internet as a timing source.

Internet timing sources can’t be authenticated using NTP’s security measure and an Internet timing source will exist outside a networks firewall so a TCP/IP port would need to be left open both of these problems could leave a network open to malicious attacks, abuse or even fraud. Another reason why Internet sources should not be used is that a recent survey found less than a third were accurate enough to UTC to be useful and those that were depended on the distance from the client.

Time and frequency transmissions broadcast in longwave are far more secure and far more accurate than any Internet timing source. These transmissions are broadcast by several national physics laboratories and provide a safe, secure and highly accurate method of receiving UTV. Unfortunately not every country has a timing broadcast and even in a country that does the signals are vulnerable to local geography and interference.

GPS however, is available everywhere on the planet, the only downside at all in using a GPS time server is that it needs an antenna to be situated on the roof, which in most cases is not normally an issue.

NTP time server (Network Time Protocol) abuse is quite often unintentional and fortunately thanks to the NTP pool is less frequent than it was although incidents still happen.

NTP server abuse is any act that violates the access rules of a NTP time server or an act that damages it in any way. Public NTP servers are those servers that can be accessed from across the Internet by devices and routers to use as a timing source to synchronise a network to. Most public NTP time servers are non-profit and set up as acts of generosity, mostly by University’s or other technical centres.

For this reason access rules have to be set up as huge amounts of traffic can generate giant bandwidth bills and can lead to the NTP time server being turned off permanently. Access rules are used to prevent too much traffic from accessing stratum 1 servers, by convention stratum 1 servers should only be accessed by stratum 2 servers which in turn can pass the timing information on down the line.

However, the worst cases of NTP server abuse have been where thousands of devices have sent requests for time, where in the hierarchical nature of NTP only one is needed.

Whilst most acts of NTP abuse are intentional some of the worst abuses of NTP time servers have been committed (albeit unintentionally) by large companies. The first large firm discovered to have been guilty of NTP abuse was Netgear, who, in 2003 released four routers that were all hard coded to use the University of Wisconsin’s NTP server, the resulting DDS (Distributed Denial of Service) reached nearly 150 megabits a second.

Even now, five years on and despite the release of several patches to fix the problem and the University being compensated by Netgear the problem still continues as some people have never patched their routers.

Similar incidents have been committed by SMC and D-Link. D-Link in particular caused controversy as when the matter was drawn to their attention they decided to bring the lawyers in. Only after it was discovered that they violated nearly 50 NTP servers did they attempt resolve the problem (and only after scathing press coverage did they relent).

The easiest way to avoid such problems is to use a dedicated external stratum 1 time server. These devices are relatively inexpensive, simple to install and far more accurate and secure than online NTP servers. These devices receive the time from atomic clocks either from the GPS network (Global Positioning System).

Time has always played an important part in civilisation. Understanding and monitoring time has been one of the pre-occupations of mankind since prehistory and the ability to keep track of time was as important to the ancients as it is to us.

Our ancestors needed to know when the best time was to plant crops or when to gather for religious celebrations and knowing the time means making sure it is the same as everyone else’s.

Time synchronization is the key to accurate time keeping as arranging an event at a particular time is only worthwhile if everybody is running at the same time. In the modern world, as business has moved from a paper-based system to an electronic one, the importance of time synchronisation and the search for ever better accuracy is even more crucial.

Computer networks are now communicating with each other from across the globe conducting billions of dollars worth of transactions every second, millisecond accuracy is now part of business success.

Computer networks can be comprised of hundreds and thousands of computers, servers and routers and while they all have an internal clock, unless they are synchronised perfectly together a myriad of potential problems could occur.

Security breaches, data loss, frequent crashes and breakdowns, fraud and customer credibility are all potential hazards of poor computer time synchronisation. Computers rely on time as the only point of reference between events and many applications and processes are time dependent.

Even discrepancies of a few milliseconds between devices can cause problems particularly in the world of global finance where millions are gained or lost in a second. For this reason most computer networks are controlled by a time server. These devices receive a time signal from an atomic clock. This signal is then distributed to every device on the network, ensuring that all machines have the identical time.

Most synchronisation devices are controlled by the computer program NTP (Network Time Protocol). This software regularly checks each device’s clock for drift (slowing or accelerating from the desired time) and corrects it ensuring the devices never waver from the synchronised time.

Installation of a NTP time server (network time protocol) is relatively straight forward. Like other computer servers a NTP time server can be mounted to a rack or left as a stand alone.

NTP time servers are the same as other servers in their installation except that there is one other consideration to take into account and that is where the time signal is coming from.

NTP time servers can utilise any timing signal but for the purposes of network time synchronisation a UTC (Coordinated Universal Time) time source is a prerequisite. UTC is available from several sources. The most common source for UTC is from across the Internet however, as these external time servers can’t be authenticated, their accuracy verified and exist outside the firewall, no serious network administrator would consider using them.

Fortunately two readily available source of UTC time do exist: the GPS network and the national time and frequency transmissions, these both offer secure, reliable and highly accurate sources of UTC time (technically GPS does not broadcast UTC but an NTP time server will convert the time into UTC).

Which one of these two systems you plumb for will affect the installation of the NTP time server. Most radio referenced time servers will receive the long wave transmission indoors although some moving around of the equipment to reduce interference may be necessary.

With GPS the only consideration is that the antenna has to have a clear view of the sky so ideally should be placed on the roof, this has obvious constraints as there is a finite length of cable that can be used.

NTP (Network Time Protocol) has been around as long as the Internet itself. Its longevity owes much to its original designer Professor David Mills of the University of Delaware in the USA.

Professor Mills saw the need for a time synchronisation protocol when the Internet was in its infancy and it was the development of NTP that has made possible the online global communication that we all take for granted.

Without synchronisation and a reliable synchronisation protocol many of the International transactions that we have taken for granted over recent years such as airline reservation, online banking, Internet auctions and trading on the stock markets, would be impossible.

The NTP time server is used throughout the world as a reliable method of synchronisation using a single time reference that is distributed amongst all devices on a network.

Most NTP time servers use a UTC time reference. UTC (Coordinated Universal Time) is  a global timescale based on the time told by atomic clocks making it a reliable and accurate timescale.

By using a NTP time server a computer network can be synchronised not just with itself but with every other computer network across the globe that uses UTC as its timing source.

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.

There are many protocols designed for distributing and synchronising time across a network. Yet out of every 100 network administrators, 99 will say they use NTP for this purpose.

While other time protocols do exist, NTP is without doubt the most widely used time synchronisation protocol and the reason for this is many fold.

Network Time Protocol, is not only the oldest time synchronisation protocol it is one of the oldest protocols on the Internet. This is one reason why NTP is so popular as it has been around since 1985 when Professor David Mills and his team from Delaware University developed the protocol.

Since then NTP has been continually updated. This constant support (currently NTP is on version 4) and development has kept NTP viable over 20 years later and provided the confidence that most administrators require.

NTP works by using a single time source (normally UTC – coordinated universal time) and distributes it amongst the network’s devices. A NTP time server will check how much each devices clock is drifting and make adjustments. NTP can typically keep a network running to within a few milliseconds of UTC, as long as there is a reliable UTC source.

Dedicated network time servers are on the whole reliable pieces of hardware and very rarely do they fail to receive and distribute a timing signal.  However, some common errors do occur:

Network time servers that receive a time signal from the GPS satellite system can often fail to find a signal if they have not been given enough time to lock on to a satellite. All GPS network time servers should be left on for at least 48 hours to ensure the time server can get a stable lock on a the satellites.

Network time servers that receive a radio signal can also fail, the most common reason is that the signal is being blocked by interference. Move the time server away from any other electrical equipment; even just tilting the network time server can increase the signal strength as the antenna should be perpendicular to the signal.

Radio signals are also disrupted for scheduled maintenance users of the UK’s MSF signal should check with the National Physical Laboratory for planned outages.

Network time servers are also vulnerable to over subscription if a network consists of a lot of machines all trying to receive a time signal then the bandwidth can be taken up with time requests. Ensure the correct use of stratum to alleviate the problem.