Home > NTP configuration

Archive for the ‘NTP configuration’ Category

The World in Perfect Synchronization

Wednesday, May 13th, 2009

Synchronization is something we are familiar with everyday of our lives. From driving down the highway to walking crowded street; we automatically adapt our behaviour to synchronize with those around us. We drive in the same direction or walk the same thoroughfares as other commuters as failing to do so would make our journey a lot more difficult (and dangerous).

When it comes to timing, synchronisation is even more important. Even in our day to day dealings we expect a reasonable amount of synchronisation from people. When a meeting starts at 10am we expect everybody to be there within a few minutes.

However, when it comes to computer transactions across a network, accuracy in synchronisation becomes even more important where accuracy to a few seconds is too inadequate and synchronisation to the millisecond becomes essential.

Computers use time for every transaction and process they do and you only have to think back to the furore caused by the millennium bug to appreciate the importance computer’s place on time. When there is not precise enough synchronisation then all sorts of errors and problems can occur, particularly with time sensitive transactions.

Its not just transactions that can fail without adequate synchronisation but time stamps are used in computer log files so if something goes wrong or if a malicious user has invaded (which is very easy to do without adequate synchronisation) it can take a long time to discover what went wrong and even longer to fix the problems.

A lack of synchronisation can also have other effects such as data loss or failed retrieval it can also leave a company defenceless in any potential legal argument as a badly or unsynchronised network can be impossible to audit.

Millisecond synchronisation is however, not the headache many administrators assume it is going to be. Many opt to take advantage of many of the online timeservers that are available on the internet but in doing so can generate more problems than it solves such as having to leave the UDP port open in the firewall (to allow the timing information through) not-to-mention no guaranteed level of accuracy from the public time server.

A better and simpler solution is to use a dedicated network time server that uses the protocol NTP (Network Time Protocol). A NTP time server will plug straight into a network and use the GPS (Global Positioning System) or specialist radio transmissions to receive the time direct from an atomic clock and distribute it amongst the network.

Why the Need for NTP

Friday, May 8th, 2009

Network Time Protocol is an Internet protocol used to synchronize computer clocks to a stable and precise time reference. NTP was originally developed by Professor David L. Mills at the University of Delaware in 1985 and is an Internet standard protocol and is used in most network time servers, hence the name NTP server.

NTP was developed to solve the problem of multiple computers working together and having the different time. Whilst, time usually just advances, if programs are running on different computers time should advance even if you switch from one computer to another. However, if one system is ahead of the other, switching between these systems would cause time to jump forward and back.

As a consequence, networks may run their own time, but as soon as you connect to the Internet, effects become visible. Just Email messages arrive before they were sent, and are even replied to before they were mailed!

Whilst this sort of problem may seem innocuous when it comes to receiving email, however, in some environments a lack of synchronisation can have disastrous results this is why air traffic control was one of the first applications for NTP.

NTP uses a single time source and distributes it amongst all devices on a network it does this by using an algorithm that works out how much to adjust a system clock to ensure synchronisation.

NTP works on a hierarchical basis to ensure there are no network traffic and bandwidth problems. It uses a single time source, normally UTC (coordinated universal time) and receives time requests from the machines on the top of the hierarch which then pass the time on further down the chain.

Most networks that utilise NTP will use a dedicated NTP time server to receive their UTC time signal. These can receive the time from the GPS network or radio transmissions broadcast by national physics laboratories. These dedicated NTP time servers are ideal as they receive time direct from an atomic clock source they are also secure as they are situated externally and therefore do not require interruptions in the network firewall.

NTP has been an astronomical success and is now used in nearly 99 per cent of time synchronisation devices and a version of it is included in most operating system packages.

NTP owes much of its success to the development and support it continues to receives nearly three decades after its inception which is why t is now used throughout the world in NTP servers.

Increased Accuracy of Dual NTP Server Systems

Wednesday, May 6th, 2009

The NTP time server has revolutionised the synchronisation of computer networks over the last twenty years. NTP (Network Time Protocol) is the software  that  is responsible for distributing time from the time server to the entire network, adjusting machines for drift and assuring accuracy.

NTP can reliable maintain system clocks to within a few millimetres of UTC (Coordinated Universal Time) or whatever timescale it is fed with.

However NTP can only be as reliable as the time source that it receives and as UTC  is the global civil timescale it depends on where the UTC source comes from.

National time and frequency transmissions from physics labs like NIST in the USA or NPL in the UK are extremely reliable sources of UTC and NTP time servers are designed specifically for them. However, the time signals are not guaranteed, they can drop off throughout the day and are susceptible to interference; they are also regularly turned of for maintenance.

For most applications a few hours of your network relying on crystal oscillators will probably not cause too much problems in synchronisation. However, GPS (Global Positioning System) is far more reliable source for UTC time in that a GPS satellite is always overhead. They do require a line-of-sight reception which means an antenna has to go on the roof or outside an open window.

For applications where accuracy and reliability are essential the safest solution is to invest in a dual system NTP time server, these device can receive both the radio transmissions such  as MSF, DCF-77 or WWVB and the GPS signal.

On a dual system NTP server, NTP will take both time sources and to synchronise a network to ensuring increased accuracy and reliability.

What is the Best Source of UTC Time?

Sunday, May 3rd, 2009

UTC (Coordinated Universal Time) is the world’s global timescale and replaced the old time standard GMT (Greenwich Meantime) in the 1970’s.

Whilst GMT was based on the movement of the Sun, UTC is based on the time told by atomic clocks although it is kept inline with GMT by the addition of ‘Leap Seconds’ which compensates for the slowing of the Earth’s rotation allowing both UTC and GMT to run side by side (GMT is often mistakenly referred to as UTC – although as there is no actual difference it doesn’t really matter).

In computing, UTC allows computer networks all over the world to synchronise to the same time making possible time sensitive transactions from across the globe. Most computer networks used dedicated network time servers to synchronise to a UTC time source. These devices use the protocol NTP (Network Time Protocol) to distribute the time across the networks and continually checks to ensure there is no drift.

The only quandary in using a dedicated NTP time server is selecting where the time source comes from which will govern the type of NTP server you require. There are really three places that a source of UTC time can be easily located.

The first is the internet. In using an internet time source such as time.nist.gov or time.windows.com a dedicated NTP server is not necessarily required as most operating systems have a version of NTP already installed (in Windows just double click the clock icon to see the internet time options).

*NB it must be noted that Microsoft, Novell and others strongly advise against using internet time sources if security is an issue. Internet time sources can’t be authenticated by NTP and are outside the firewall which can lead to security threats.

The second method is to use a GPS NTP server; these devices use the GPS signal (most commonly used for satellite navigation) which is actually a time code generated by an atomic clock (from onboard the satellite). Whilst this signal is available anywhere on the globe, a GPS antenna does need a clear view of the sky which is the only drawback in using GPS.

Alternatively, many countries’ national physics laboratories such as NIST in the USA and NPL in the UK, transmit a time signal from their atomic clocks. These signals can be picked up with a radio referenced NTP server although these signals are finite and vulnerable to local interference and topography.

Galileo and the GPS NTP Server

Thursday, April 23rd, 2009

Currently there is only one Global Navigation Satellite System (GNSS) the NAVSTAR GPS which has been open for civilian use since the late 1980’s.

Most commonly, the GPS system is thought to provide navigational information allowing drivers, sailors and pilots to pinpoint their position anywhere in the world.

In fact, the only information beamed from a GPS satellite is the time which is generated by the satellites internal atomic clock. This timing signal is so accurate that a GPS receiver can use the signal from three satellites and pinpoint the location to within a few metres by working out how long each precise signal took to arrive.

Currently a GPS NTP server can use this timing information to synchronise entire computer networks to providing accuracy to within a few milliseconds.

However, the European Union is currently working on Europe’s own Global Navigation Satellite System called Galileo, which will rival the GPS network by providing its own timing and positioning information.

However, Galileo is designed to be interoperable with GPS meaning that a current GPS NTP server will be able to receive both signals, although some software adjustments may have to be made.

This interoperability will provide increased accuracy and may make national time and frequency radio broadcasts obsolete as they will not be able to produce a comparable accuracy.

Furthermore, Russia, China and India are currently planning their own GNSS systems which may provide even more accuracy. GPS has already revolutionised the way the world works not only by allowing precise positioning but also enabling entire globe to synchronise to the same timescale using a GPS NTP server. It is expected that even more advances in technology will emerge once the next generation of GNSS begin their transmissions.

Choosing the Right Time Signal for Your Network

Wednesday, April 22nd, 2009

Computer network synchronisation is essential in the modern world. Many of the world’s computer networks are all synchronised to the same global timescale UTC (Coordinated Universal Time).

To govern synchronisation the protocol NTP (Network Time Protocol) is used in most cases as it is able to reliably synchronise a network to a few milliseconds off UTC time.

However, the accuracy of time synchronisation is solely dependent on the accuracy of whatever time reference is selected for NTP to distribute and here lies one of the fundamental errors made in synchronising computer networks.

Many network administrators rely on Internet time references as a source of UTC time, however, apart from the security risks they pose (being as they are on the wrong side of a network firewall) but also their accuracy can not be guaranteed and recent studies have found less than half of them providing any useful accuracies at all.

For a secure, accurate and reliable method of UTC there really are just two choices. Utilise the time signal from the GPS network or rely on the long wave transmissions broadcast by national physics laboratories such as NPL and NIST.

To select which method is best then the only factor to consider is the location of the NTP server that is to receive the time signal.

GPS is the most flexible in that the signal is available literally everywhere on the planet but the only downside to the signal is that a GPS antenna has to be situated on the roof as it needs a clear view of the sky. This may prove problematic if the time server is located in the lower floors of a sky scraper but on the whole most users of GPS time signals find that they are very reliable and incredibly accurate.

If GPS is impractical then the national time and frequencies provide an equally accurate and secure method of UTC time. These longwave signals are not broadcast by every country however, although the US WWVB signal broadcast by NIST in Colorado is available in most of North America including Canada.

There are various versions of this signal broadcast throughout Europe including the German DCF and the UK MSF which prove to be the most reliable and popular. These signals can often be picked up outside the nation’s borders too although it must be noted long wave transmissions are vulnerable to local interference and topography.

For complete peace of mind, dual system NTP servers that receive signals from both the GPS and national physics laboratories are available although they tend to be a little more expensive than single systems although utilising more than one time signal makes them doubly reliable.

Atomic Clocks Explained

Monday, April 20th, 2009

Is an Atomic Clock Radioactive?

An atomic clock keeps time better than any other clock. They even keep time better than the rotation of the Earth and the movement of the stars. Without the atomic clock, GPS navigation would be impossible, the Internet would not synchronise, and the position of the planets would not be known with enough accuracy for space probes and landers to be launched and monitored.

An atomic clock is not radioactive, it doesn’t rely on atomic decay. Rather, an atomic clock has an oscillating mass and a spring, just like ordinary clocks.

The big difference between a standard clock in your home and an atomic clock is that the oscillation in an atomic clock is between the nucleus of an atom and the surrounding electrons. This oscillation is not exactly a parallel to the balance wheel and hairspring of a clockwork watch, but the fact is that both use oscillations to keep track of passing time. The oscillation frequencies within the atom are determined by the mass of the nucleus and the gravity and electrostatic “spring” between the positive charge on the nucleus and the electron cloud surrounding it.

What Are The Types of Atomic Clock?

Today, though there are different types of atomic clock, the principle behind all of them remains the same. The major difference is associated with the element used and the means of detecting when the energy level changes. The various types of atomic clock include:

The Cesium atomic clock employs a beam of cesium atoms. The clock separates cesium atoms of different energy levels by magnetic field.

The Hydrogen atomic clock maintains hydrogen atoms at the required energy level in a container with walls of a special material so that the atoms don’t lose their higher energy state too quickly.

The Rubidium atomic clock, the simplest and most compact of all, use a glass cell of rubidium gas that changes its absorption of light at the optical rubidium frequency when the surrounding microwave frequency is just right.

The most accurate commercial atomic clock available today uses the cesium atom and the normal magnetic fields and detectors. In addition, the cesium atoms are stopped from zipping back and forth by laser beams, reducing small changes in frequency due to the Doppler effect.

When Was The Atomic Clock Invented? atomic clock

In 1945, Columbia University physics professor Isidor Rabi suggested that a clock could be made from a technique he developed in the 1930s called atomic beam magnetic resonance. By 1949, the National Bureau of Standards (NBS, now the National Institute of Standards and Technology, NIST) announced the world’s first atomic clock using the ammonia molecule as the source of vibrations, and by 1952 it announced the first atomic clock using cesium atoms as the vibration source, NBS-1.

In 1955, the National Physical Laboratory (NPL) in England built the first cesium-beam atomic clock used as a calibration source. Over the next decade, more advanced forms of the atomic clocks were created. In 1967, the 13th General Conference on Weights and Measures defined the SI second on the basis of vibrations of the cesium atom; the world’s time keeping system no longer had an astronomical basis at that point! NBS-4, the world’s most stable cesium atomic clock, was completed in 1968, and was used into the 1990s as part of the NPL time system.

In 1999, NPL-F1 began operation with an uncertainty of 1.7 parts in 10 to the 15th power, or accuracy to about one second in 20 million years, making it the most accurate atomic clock ever made (a distinction shared with a similar standard in Paris).

How Is Atomic Clock Time Measured?

The correct frequency for the particular cesium resonance is now defined by international agreement as 9,192,631,770 Hz so that when divided by this number the output is exactly 1 Hz, or 1 cycle per second.

The long-term accuracy achievable by modern cesium atomic clock (the most common type) is better than one second per one million years. The Hydrogen atomic clock shows a better short-term (one week) accuracy, approximately 10 times the accuracy of a cesium atomic clock. Therefore, the atomic clock has increased the accuracy of time measurement about one million times in comparison with the measurements carried out by means of astronomical techniques.

Synchonising to an Atomic Clock

The simplest way to synchonise to an atomic clock is to use a dedicated NTP server. These devices will receive either the GPS ataomic clock signal or radio waves from places like NIST or NPL.

Features of Network Time Protocol

Thursday, April 16th, 2009

NTP is reliant on a reference clock and all clocks on the NTP network are synchronised to that time. It is therefore imperative that the reference clock is as accurate as possible. The most accurate timepieces are atomic clocks. These large physics lab devices can maintain accurate time over millions of years without losing a second.

An NTP server will receive the time from an atomic clock either from across the internet, the GPS network or radio transmissions. In using a atomic clock as a reference an NTP network will be accurate to within a few milliseconds of the world’s global timescale UTC (Coordinated Universal Time).

NTP is a hierarchical system. The closer a device is to the reference clock the higher on the NTP strata it is. An atomic clock reference clock is a stratum 0 device and a NTP server that receives the time from it is a stratum 1 device, clients of the NTP server are stratum 2 devices and so on.

Because of this hierarchical system, devices lower down the strata can also be used as a reference which allows huge networks to operate while connected to just one NTP time server.

NTP is a protocol that is fault tolerant. NTP watches out for errors and can process multiple time sources and the protocol will automatically select the best.   Even when a reference clock is temporarily unavailable, NTP can use past measurements to estimate the current time..

Receiving the Time and Finding the Correct Time Source

Monday, April 6th, 2009

So you have decided to synchronize your network to UTC (Coordinated Universal Time), you have a time server that utilizes NTP (Network Time Protocol) now the only thing to decide on is where to receive the time from.

NTP servers do not generate time they simply receive a secure signal from an atomic clock but it is this constant checking of the time that keeps the NTP server accurate and in turn the network that it is synchronizing.

Receiving an atomic clock time signal is where the NTP server comes into its own. There are many sources of UTC time across the Internet but these are not recommended for any corporate use or for whenever security is an issue as internet sources of UTC are external to the firewall and can compromise security – we will discuss this in more detail in future posts.

Commonly, there are two types of time server. There are those that receive an atomic clock source of UTC time from long wave radio broadcasts or those that use the GPS network (Global Positioning System) as a source.

The long wave radio transmissions are broadcast by several national physics laboratories. The most common signals are the USA’s WWVB (broadcast by NIST – National Institute for Standards and Time), the UK’s MSF (broadcast by the UK National Physical Laboratory) and the German DCF signal (Broadcast by the German National Physics Laboratory).

Not every country produces these time signals and the signals are vulnerable to interference from topography. However, in the USA the WWVB signal is receivable in most areas of North America (including Canada) although the signal strength will vary depending on local geography such as mountains etc.

The GPS signal on the other hand is available literally everywhere on the planet as along as the GPS antenna attached to the GPS NTP server can have a clear view of the sky.

Both systems are a truly reliable and accurate method of UTC time and using either will allow synchronization of a computer network to within a few milliseconds of UTC.

Difficulties in telling the time!

Friday, April 3rd, 2009

Precision in telling the time has never been as important as it is now. Ultra precise atomic clocks are the foundation for many of the technologies and innovations of the twentieth century. The internet, satellite navigation, air traffic control and global banking all just a few of the applications that is reliant on particularly accurate timekeeping.

The problem we have faced in the modern age is that our understanding exactly of what time is has changed tremendously over the last century. Previously it was thought that time was constant, unchanging and that we travelled forward in time at the same rate.

Measuring the passing of time was straight forward too. Each day, governed by the revolution of the Earth was divided into 24 equal amounts – the hour.  However, after the discoveries of Einstein during the last century, it was soon discovered time was not at all constant and could vary for different observers as speed and even gravity can slow it down.

As our timekeeping became more precise another problem became apparent and that was the age old method of keeping track of the time, by using the Earth’s rotation, was not an accurate method.

Because of the Moon’s gravitational influence on our oceans, the Earth’s spin is sporadic, sometimes falling short of the 24 hour day and sometimes running longer.

Atomic clocks were developed to try to keep time as precise as possible. They work by using the unchanging oscillations of an atom’s electron as they change orbit. This ‘ticking’ of an atom occurs over nine billion times a second in caesium atoms which makes them an ideal basis for a clock.

This ultra precise atomic clock time (known officially as International Atomic Time – TAI) is the basis for the world’s official timescale, although because of the need to keep the timescale in parallel with the rotation of the Earth (important when dealing with extra terrestrial bodies such as astronomical objects or even satellites) addition seconds, known as leap second, are added to TAI, this altered timescale is known as UTC – Coordinated Universal Time.

UTC is the timescale used by businesses, industry and governments all around the world. As it is governed by atomic clocks it means the entire world can communicate using the same timescale, governed by the ultra-precise atomic clocks. Computer networks all over the world receive this time using NTP servers (Network Time Protocol) ensuring that everybody has the same time to within a few milliseconds.