We rely on NTP to synchronize the clocks on all of our network-connected devices. In this video, you’ll learn about NTP servers, NTP clients, and how to configure NTP options on your workstation.
<< Previous Video: Configuring DHCP Next: Copper Cabling >>
All of the devices on our network have their own clock. This would be every server, every router, every switch, every workstation, and every other device. There is usually log files, authentication details, anything dealing with logging of information is all using the date and time that is gathered from this clock information. So, having this synchronized across all of these devices is very important. That would allow you to compare log files across multiple devices to track exactly what’s going on in your environment. And, of course, you would like this to occur automatically. You don’t want to have to go into every single device and update the clock or have to update the clock every time the device is rebooted. Fortunately, we’re able to do this using NTP– or the Network Time Protocol. This allows us to configure our devices with an NTP server and then they can check in to that NTP server to update their clocks on a regular basis. This is an extremely accurate way of updating these clocks. Usually, the accuracy is better than one millisecond when you’re updating the clocks with an NTP server on your local network.
If you’re configuring NTP on your network you may want to have an internal NTP server. This is a server that is in charge of the clock for all of the devices that need to get updates. And as an NTP server, it does not modify its own time. An NTP client is a device that will request these times from the NTP server and then update their clocks accordingly. Sometimes a device can be both an NTP client and an NTP server. So it can update its time based on the time that may be available on a more accurate NTP server and then it can provide those updates to other NTP clients on your network. If you’re planning to configure this on your network then you need to decide what devices will be NTP clients, which will be NTP servers, and which devices may be both an NTP client and an NTP server.
We can associate a value with an NTP server that designates how far it is away from the original time source. We call these stratum layers in NTP. That’s because the original reference clock would be stratum 0. This might be an atomic clock or a GPS clock that you’ve configured. The next server away from stratum 0 would be a stratum 1 server, which is going to be obtaining its timestamp from a stratum 0 server. You might also have stratum 2 servers on your network, which are going to be receiving their timestamps from a stratum 1 server.
On your client workstations you generally would configure your NTP client, which may specify an IP address or the domain name of a particular NTP server. And if this is inside of your network you may want to specify your own internal IP addresses. You can also use multiple NTP servers just in case one of those is not available. If this NTP service is going to be on your own internal network you’ll need at least one NTP server to act as the clock source. You’ll also specify in the configuration of that NTP server what stratum this device happens to be. So, if a client has the choice of synchronizing itself with an NTP server that has a stratum level of two and another NTP server that has a stratum level of one it will choose to synchronize itself with the server that has the lower stratum number or the stratum level one.