Can accurate delivery of time over optical transport answer data center synchronization challenges?
As computer technology continues to improve, data centers (large facilities where data is stored and processed) are able to process more and more data. This means that data is being handled and distributed across many different locations, rather than just in one central location. In order for this data to be useful, it needs to be organized in the correct order and all of the different parts of the data need to be synced up with each other. The current methods and technologies used to make this happen are not accurate or precise enough to keep up with the increasing amount of data being processed. This means that new and improved technologies are needed to ensure that data is being handled correctly and can be used effectively.
Reducing errors and enhancing reliability
The synchronization network technologies we currently use are becoming less effective as the amount of data being processed increases. One method that is currently used is satellite-based synchronization, which is generally accurate but has some vulnerabilities that make it less reliable. Another method is using packet transport networks, which can be designed to be very resilient but still have issues with variations in timing and accuracy. Additionally, the technology that is used to keep accurate time, called atomic cesium clock technology, is becoming less reliable as the need for greater stability and accuracy increases. Therefore, it is crucial that new and improved technologies are developed to enhance the accuracy and resilience of the synchronization process.
GNSS, or Global Navigation Satellite System, is a method of synchronizing data by using signals from satellites. However, this method is vulnerable to various types of attacks or interference that can disrupt the timing and accuracy of the data. These include malicious jamming and spoofing attacks, which can disrupt the signals from the satellites and cause outages, as well as interference from other radio signals or disruptions caused by the sun. To address these challenges, we need new types of GNSS receivers that can use multiple methods to ensure accuracy and reliability. By using multiple frequency bands and different types of satellite signals, the GNSS receivers can be made more resilient and less vulnerable to attacks. Additionally, using AI and machine learning technology, the receivers can analyze the performance of many different receivers and detect any problems or anomalies, and take action to prevent service disruptions before they happen.
Packet networks are networks that transmit data by breaking it down into small packets, which are then sent separately and reassembled at the destination. These networks can cause problems with timing, because the packets can be delayed or arrive out of order, which negatively impacts the quality of the timing signals. To address these issues, packet network devices need to actively measure and compensate for these delays in real-time. One way to do this is by using transparent or boundary clocks, which are devices that can measure the delay and adjust the timing signals accordingly. Another approach is to use a co-located synchronization device that can build an overlay synchronization network, which is a network that is built on top of the existing packet network and can provide more accurate timing. This can be achieved by using a bidirectional optical timing channel which directly interconnects boundary clocks with native Ethernet-carrying PTP traffic, eliminating delay differences from dual fiber approaches.
An example of such a synchronization overlay network interconnecting highly accurate boundary class D clocks is shown in the image below. (With 5ns accuracy, class D is the highest quality level defined by ITU-T G.8275.1.)
The overlay approach is a way to improve the accuracy of timing signals in packet networks by avoiding the delay variations caused by the technology used to transmit the data, specifically called OTN (Optical Transport Network) technology. This is achieved by using a special type of network called an optical timing channel, which operates separately from the main data network and directly connects the devices that measure and adjust the timing (called boundary clocks) with the data network (native Ethernet-carrying PTP traffic) in both directions.
This approach uses a single optical fibre to transmit data, instead of multiple fibers. By using a single fiber, it eliminates the delay differences that can occur when using multiple fibers, which can negatively impact the quality of timing signals. This results in more precise and accurate timing signals, which helps in synchronizing data in a distributed way. This is particularly useful in data centers where large amount of data is being processed.
A new type of atomic clock called the commercial optical cesium atomic clock has recently been released. This new technology is more accurate and stable than previous types of atomic clocks, and it also has a longer lifespan. Early tests have shown that the new clock performs better and maintains stability over longer periods of time. Because of its improved stability, it can be used to synchronize data even when other methods are not working properly. Additionally, this new clock also allows for remote management and is more secure.
Increasing timing accuracy in metrology and scientific research requires the development of atomic clocks, timing-transport networks and satellite-delivered synchronization technologies. This unique combination of advanced synchronization technologies is only provided by ADVA.