The demand for network appliances to be interconnected at the lowest latency level and synchronized to sub-microsecond precision has never been higher. Such capabilities enable the most efficient consumption of resources by applications such as radio access networks (RANs) in a world that is increasingly relying on virtualization.
In fact, interconnecting RANs is one of the most challenging applications. To deliver the best user experience, it is fundamental that radio base stations are timed as accurately as possible. As networks evolve to 5G, aligning mobile cells by both frequency and time will be essential for high-speed transmission and reception of data between multiple base stations and mobile user devices. Operators typically find it difficult to attain accuracy in frequency and phase synchronization, which is mandatory for LTE-TDD and LTE-Advanced Pro technology. They also struggle to gain an understanding of how accurately master clocks are being tracked by slave clocks residing in base station devices.
To support data rates of up to 1 Gbps and make more effective use of the radio spectrum, it is critical that there is reliable and secure delivery of frequency and precise phase synchronization. A more efficient use of the radio spectrum is important, using it to send more information per transmission and more transmissions per spectrum and using the correct spectrum for the transmission.
Network operators are increasingly looking to technology that will future proof them as the evolution from 4G to 4.5G and ultimately 5G networks takes place. Additionally they are moving to make transport networks central to their infrastructure, and optical transport is becoming increasingly popular due to capacity demands.
To distribute accurate timing across the network, the importance of metro and regional networks, as well as optical and Optical Transport Network (OTN) network infrastructure, is growing.
How Can Precision Timing Be Achieved?
In distributed networks there are two ways to synchronize frequency and time to a sub-microsecond level.
One technique uses the Global Navigation Satellite Systems (GNSS). This method can supply a time signal accurate to better than 100 nsec. The applications being timed can recover precise timing information from GNSS receivers deployed at each location.
However, this will not be a common solution for much longer. GNSS-jamming devices have become increasingly available, posing serious risks to timing accuracy. Additionally, GNSS cannot meet highest availability, scalability, and minimum operational complexity requirements. The acquisition of roofing rights, deployment of hardware and dedicated cables, and operation of timing gear is costly and difficult and cannot be expanded to the hundreds of thousands of nodes that would be required.
Alternatively, the underlying data network can be used to distribute timing information (Figure 1). This can be done either as single source or in tandem with GNSS to overcome the availability limitations in performance-critical applications. Accuracy can be increased to just several hundred nanoseconds and operational complexity can be significantly simplified when using network-based timing. With this in mind, operators have started to reduce the number of GNSS applications in favor of network-based timing approaches. This evolution helps them cater to the demand for precise synchronization from RANs as well as financial trading applications and smart power grids.
|Figure 1. Operators have begun to explore use of the underlying data network to|
distribute timing information.
The industry has chosen IEEE 1588-2008 Precision Time Protocol (PTP) to distribute timing with sub-microsecond accuracy over packet networks. PTP is based on the exchange of timestamped packets between a master and a slave clock. However, it remains essential that the underlying network experiences as little dynamic delay as possible; but this can be an ever-changing challenge.
What Will Happen to Optical Networks?
Current optical networks need to be augmented to enable transport of precise synchronization for all applications and ensure efficient consumption of resources, deliver a good consumer experience, and adhere to regulations. Existing OTN technology is asynchronous and therefore not well suited for the distribution of precise timing information when it comes to phase and time synchronization.
This may pose a problem for OTN network operators regardless of whether offering wholesale transport services to a mobile network operator or deploying infrastructure serving a packet-switched architecture on top. They must ensure that data is forwarded with maximum timing integrity, particularly the embedded flows of PTP frames carrying timing information.
An optical timing channel for out-of-band transfer and time-sensitive OTN technology for in-band transportation of PTP data will provide network operators with a reliable basis for precision timing. Synchronization information can be sent across optical networks with maximum integrity and without the fluctuation dynamic delay asymmetries cause.
The metro and core layer of an optical network's infrastructure allow network operators to efficiently transport PTP flows from a grandmaster clock over longer distances. This can be done using optical timing channel technology (Figure 2). Using this newer transfer technique bypasses both service multiplexing and grooming generating dynamic latencies. The approach enables data packets executing one or more PTP flows to be treated in a dedicated manner with a bi-directional single-fiber wavelength. This reduces timing errors and can be terminated at a distribution node at the edge of the network, removing issues that usually arise from asynchronous digital bit stream processing.
|Figure 2. Operators can use optical timing channel technology to transport PTP flows from a grandmaster clock over long distances.|
The precision of PTP can be further improved by strategically deploying boundary clocks at certain nodes to compensate for static link asymmetries. This strategy leads to intrinsically low delay asymmetries, as the functional blocks of DWDM network that are responsible for creating the asymmetry are completely bypassed.
|Figure 3. The use of boundary clocks can compensate for static link asymmetries.|
Operators also can create a highly stable timing core – not reliant on continuous availability of GNSS – by implementing enhanced primary reference time clocks (ePRTC) at a limited number of strategic sites across the core of the network. Generating time aligned with GNSS but that doesn't interfere with the autonomy of the ePRTC allays GNSS vulnerabilities that can come with this combination of technologies.
How Can We Plan for OTN Buffer Delays?
Buffer-controlled OTN mapping can also be used to carry PTP signals while maintaining accuracy of time. Technology is now available that constantly monitors and averages buffer fill levels, unlike other conventional OTN mapping applications. This means slave clocks can recover the intended level of accuracy because dynamic node delays and latency asymmetries have been minimized.
Optical network operators can offer mobile operators OTN-based wholesale services carrying both backhaul traffic and timing information for precise synchronization of base stations. They are able to use time-sensitive OTN to offer Gigabit Ethernet and 10 Gigabit Ethernet wholesale services with minimal fluctuation. Such an approach has become a popular alternative to packet-switched mobile backhaul. It creates a cost-efficient wholesale service and a more transparent decoupling of mobile and fixed network operator domain.
What Are the Challenges for Optical Network Infrastructure?
More industries than ever are relying on time synchronization to ensure they use resources effectively, boost performance levels, and meet respective regulations. With today's most advanced timing solutions, this level of precise synchronization can be provided efficiently and consistently also over optical transport networks. Now, many applications – including RANs, algorithmic and high-frequency trading platforms, and smart national power grids – can be timed precisely over the network.
Michael Ritter is vice president of technical marketing and analyst relations for ADVA Optical Networking, where he drives the innovation of their Optical + Ethernet technology portfolio into the market while recognizing current technology architectures and keeping an eye on network trends. He helps the company to best position their technology and solutions for its global service provider and enterprise customers.