This application claims priority from United Kingdom Application No.: 1610136.2, filed Jun. 10, 2016, the entire disclosure of which is incorporated herein by reference in its entirety.
The present invention relates to clock synchronisation in wireless mesh communications networks.
Such a mesh network is suitable for providing a cell 16 of a cellular wireless telecommunications network with a connection to a fibre optic network connection 15 for communication with the base network 14. The cell 16 is operable to communicate with a plurality of mobile communications devices in accordance with well-known standards and techniques. For example, the Long Term Evolution (LTE) standard defines one suitable cellular communications technique.
In the example of
In order for the cellular network to operate correctly, it is important that the cells of the network maintain a synchronous clock signal. Accordingly, it is necessary for the individual cells to maintain a clock signal that is synchronised with a master clock signal for the cellular network concerned. In addition, it is desirable for the individual network nodes to have access to a master reference clock signal in order that mobile phone cell to cell interference and signal management can be completed accurately.
One of the challenges associated with implementing a mesh network, particularly a wireless mesh network over a wide area, is that of maintaining such accurate and synchronised clock signals over the network. However, previously-considered network techniques for adjusting clock signal synchronisation are not ideally suited to wireless mesh networks, since the nature of multi-hop wireless connections mean that the clock signals can quickly become asynchronous.
Accordingly, it is desirable to provide a new technique that seeks to address the drawbacks of previously-considered clock synchronisation techniques.
According to one aspect of the present invention, there is provided a method of synchronising a local clock signal with a reference clock signal in a network node of a wireless mesh communications network which includes a plurality of such network nodes interconnected by another plurality of wireless communications links, the method comprising, at the network node, receiving a plurality of clock synchronising signals over respective wireless radio frequency signals at a beamforming steerable antenna of a transceiver having a beamforming steerable antenna for which reception parameters define a reception direction for the antenna, the clock synchronising signals being received on respective incoming beam reception directions; selecting one of the received clock synchronising signals as a reference clock signal by selecting an incoming beam reception direction of the antenna; producing a reference comparison value by comparing the reference clock signal with a local clock signal of the network node; and adjusting the local clock signal in dependence upon the reference comparison value.
According to another aspect of the present invention, there is provided a network node for a wireless mesh communications network which includes a plurality of such network nodes interconnected by another plurality of wireless communications links, the network node comprising a receiver operable to receive a plurality of clock synchronising signals; and a selection unit operable to select one of such a plurality of received clock synchronising signals as a reference clock signal, to produce a reference comparison value by comparing the reference clock signal with a local clock signal of the network node, and to adjust the local clock signal in dependence upon the reference comparison value, and a transceiver operable to receive at least one radio frequency signal, the transceiver having a beamforming steerable antenna for which reception parameters define a reception direction for the antenna wherein the transceiver is operable to receive such clock synchronising signals on respective incoming beam reception directions, and wherein the selection unit is operable to select one such received clock synchronising signal by selecting an incoming beam reception direction of the antenna.
In one example, the network node comprises a plurality of transceivers, each of which includes a beamforming steerable antenna and a processing unit, the antennas having respective communications directions, and receiving a plurality of clock signals includes receiving respective radio frequency signals at the transceivers, extracting respective clock signals form the received radio frequency signals, and selecting one of the received clock synchronising signals as a reference clock signal, producing a reference comparison value by comparing the reference clock signal with a local clock signal, and adjusting the local clock signal in dependence upon the reference comparison value are performed for each transceiver.
Such an example may also include routing any of the reference clock signals to any of the processing units for use thereby in reception and transmission of radio frequency signals received and transmitted from the associated antenna. In such an example, the network node may include a non-blocking multiplexer for the routing of the reference clock signals. In one example, such routing is performed in accordance with a timing schedule. In one example, such routing is controlled by a central controller of the network. In one example, the central controller is a software defined network controller.
In one example, the network node includes a plurality of such local clock signals, which are adjusted with reference to respective reference clock signals. In one example, adjusting the local clock signals is performed in accordance with a timing schedule. In one example, the timing schedule is provided a central controller of the network. In one example, the central controller is a software defined network controller.
According to another aspect of the present invention, there is provided a method of providing a local clock signal in a network node of a wireless mesh communications network which includes a plurality of such network nodes interconnected by another plurality of wireless communications links, the method comprising, at the network node, receiving a plurality of clock synchronising signals on respective reception beam directions of a beamforming steerable antenna device; and selecting one of the received clock synchronising signals as a local reference clock signal by selecting one of the reception beam directions.
According to another aspect of the present invention, there is provided a method of synchronising respective local clock signals of network nodes of a wireless mesh communications network which includes a plurality of such nodes interconnected by another plurality of wireless communications links, each node having at least one transceiver having a beamforming steerable antenna and operable to transmit and receive radio frequency signals in distinct signal beams having respective directions, the method comprising at a first network node: selecting a clock beam direction of the beamforming steerable antenna of the first network node; receiving a master clock signal on a radio frequency signal received by the beamforming antenna in the clock beam direction; synchronising a local clock signal of the first network node with the master clock signal; selecting a synchronisation beam direction of the beamforming antenna of the first network node; and forwarding a first synchronisation signal to a second network node on a radio frequency signal transmitted from the beamforming antenna in the synchronisation beam direction of the first network node, the first synchronisation signal being dependent upon one or both of the local clock signal of the first node and the master clock signal; at a second network node: selecting a synchronisation beam direction of the beamforming steerable antenna of the second network node; receiving the first synchronisation signal from the first network node on a radio frequency signal received by the antenna in the synchronisation beam direction of the second network node; and synchronising a local clock signal of the second network node with the first synchronisation signal, the first and second network nodes thereby forming a first clock region of the wireless mesh network.
In the example of
In a preferred example, the wireless communication links 22 are radio frequency links, using radio frequency signals in the millimeter wave range, that is in the range 20 GHz to 315 GHZ, preferably in the 60 GHZ waveband (as defined by the IEEE 802.1 lad standard and typically in the range approximately 58 GHz to 64 GHz).
The network 2 also includes a node 24 which connects the mesh network to a base network 25. This node 24 is also known as a “point of presence (POP)” node. Such a POP node 24 is typically connected with a first network node 201 by way of a wired or optical connection 23A. The POP node 24 is connected to the base network 25 by way of a wired or optical connection 23B. The base network may be provided by any suitable communications network, such as a mobile network operator's packet data network or the Internet.
The POP node 24 provides a clock signal for the mesh and cellular networks, this clock is derived from a master clock signal is also known as a Grand Master (GM) clock signal. The GM clock signal can be generated locally or provided by the mobile network operator's packet data network. The POP node 24 provides the master clock signal to the first network node 201. As will be described below, the master clock signal is used to provide a reference clock signal across the mesh network, and to a cell of destination network node 206 in particular.
In the example clock sync path 26, the POP node 24 supplies the master clock signal to a first node 201. The first node 201 synchronises its internal clock with the received master clock signal, as will be described below, and passes the master clock signal to a second node 202. This process is repeated until the master clock signal reaches a predetermined destination node. In the example shown in
The destination network node 206 provides packet data and synchronisation clock to a cell 29 for a cellular communications network device. The cell is able to communicate in a wireless manner using appropriate cellular technologies and techniques. The cell 29 makes use of the master clock signal delivered by the mesh network in order to maintain synchronicity with the cellular communications network of which it is part.
The clock sync path (or “tree”) 26 can be defined by an appropriate network resource. For example, in a software defined network (SDN), a suitably modified SDN controller may be responsible for the definition of the clock sync path. The definition and choice of the clock sync path 26 may be dynamic and respond to changes in the mesh network 2, and in the requirements for clock distribution. For example, a particular network node on a defined clock sync path may become inactive or faulty in some way. In such a case the controller may define a new clock sync path to bypass the network node concerned. This redefinition of clock sync paths is particularly suitable in a mesh network, since the very nature of the mesh enables multiple routes across the network to be defined.
One example of a modified SDN controller that controls clock synchronisation signal routing is described in a paper entitled “Extending OpenFlow for SDN-enabled Synchronous Ethernet networks” by Raúl Suárez, David Rincón, and Sebastià Salient. The paper describes one possible modification of existing SDN controller functionality to enable the propagation of an Ethernet clock synchronisation signal across a software defined network. In an example embodiment of the present invention, this modified SDN controller is further extended in order to select a particular clock synchronisation signal from a plurality of received signals, such as those received from a plurality of antenna beam directions, as described in more detail below.
Synchronisation of an internal clock of a network node 20 with a received master clock signal will now be described with reference to
The example of
With reference to
When in a receiving mode of operation, the reception characteristics of the antenna elements of the antenna 36,60 are modified according to weighting values determined by the processing unit 32 and supplied to the baseband unit 38, such that the antenna 36,60 receives RF signals from a specific direction (that is, from a specific transmitting network node).
In a transmitting mode of operation, respective drive signals are generated for the antenna elements of the antenna 36,60. The drive signals are respective modified versions of the RF modulated output signal specific to each antenna element. The output signal may be modified in phase and/or amplitude in order to produce the desired beam pattern, and hence beam direction.
The example network topologies of
The processing unit 32 will now be described. The processing unit 32 is provided with a clock generator 34. The processing unit 32 includes the baseband unit 38 which receives the baseband signal 37. The baseband unit 38 operates to synchronise to the start of a packet in the baseband signal and digitises the baseband signal into an encoded data stream 39 for further processing. This processing of the baseband signal 37 requires a clock signal 35.
The data stream 39 represents a series of data packets, a very simplified structure of which is illustrated in
The PHY header portion 73 includes information about the modulation and coding scheme used for the packet portion 74. The packet portion 74 also includes a media access control (MAC) header portion and a user data portion. The MAC header portion contains data identifying the source and destination for the user data portion. The user data portion contains at least one user data packet, and possibly associated additional control or header data, for delivery to the ultimate destination.
In one example of an aspect of the present invention, the payload portion 72 is a dedicated synchronisation payload, and so contains only the synchronisation indicators 75, and does not contain data items for transfer through the node.
In such an example, the network node may be configured to switch to the synchronisation signal at regular intervals, for example every 1 millisecond, or according to an appropriate timing schedule, in order that the local clock can remain synchronised with the master clock. Where the reception direction for the beam carrying the synchronisation signal is different to that of the current data transfer beam, the network node switches between data transfer and synchronisation beams appropriately.
In a preferred example, the network node that is responsible for transmitting the clock synchronisation signal will adhere to the predetermined timing schedule, and will adjust the transmission parameters of its beamforming steerable antenna so that the synchronisation transmission beam is transmitted in the correct direction and at the appropriate time. The adherence to the predetermined timing schedule both the transmitting network node and the receiving network node allows for the regular synchronisation of the clock signals.
For the case where the synchronising signal is a dedicated signal, the low amount of data (only the synchronisation indicators) being transferred allows for the use of a modulation and coding scheme that maximises the range of the synchronisation signal. The positions of the synchronisation indicators do not depend upon the modulation and coding scheme used.
In another example of an aspect of the present invention, the payload portion 72 includes data items to be transferred by the node, and includes the synchronisation indicators 75.
In another example, timing information may be derived from detected changes in the data modulation constellation, or by any other suitable technique.
Returning to
The payload processing unit 42 demodulates and decodes the payload portion, thereby producing a series of data packets 43 which is supplied to a media access controller (MAC) 46. The payload processing unit 42 generates a second phase signal 43a relating to the relative phase of the reference clock 35 to the encoded reference signal of the payload. The second phase signal 43a is more precise than the first phase signal 41a. The second phase signal 43a also provides a running estimate of phase changes with respect to the reference clock 35 during the payload portion of the incoming data stream. The payload processing unit 42 identifies the synchronisation indicators (75,
A detection unit 44 receives the first phase signal 41a from the preamble processor 40, and the second phase signal 43a from the payload processing unit 42. The detection unit 44 combines the first and second phase signals 41a and 43a and generates a signal 45 which relates the phase changes between the data stream and the internal reference clock. In summary, the detection unit 44 determines the difference between the reference clock and the carrier frequency (the “carrier frequency offset (CFO)”), and determines the difference between the reference clock and the sampling frequency (the “sampling frequency offset (SFO)”).
The detection unit 44 outputs the indicator signal 45 to a computational unit 48 which in turn is able to process the indicator signal 45, in order to create a long term adjustment signal 49 for phase offset adjustment of the reference clock that doesn't interfere with the baseband unit 38, the preamble processor 40, and the payload processing unit 42 data processing chain.
The MAC 46 determines routing decisions for the data packet from the header portion of the packet, and outputs each data packet appropriately, as an output data stream 47. In addition, the MAC 46 produces a signal 47a used by the computational unit 48 to indicate valid inclusion of that signal from data derived from the packet header. In such a manner, the computational unit 48 is able to use a synchronisation signal from the correct source, as determined by the MAC 46.
The long term adjustment signal 49 is output to a phase processing unit 50. The phase processing unit 50 determines how the clock generator unit 34 must be adjusted in order to reduce the phase difference between the internal clock signal 35 and the data stream 37. The phase processing unit 50 outputs a control signal 51 to the clock generator 34. The clock generator 34 adjusts the local clock signal 35 for the processing unit 32 so that the phase difference values originating from the CFO and SFO and computed by the detection unit 44, the computational unit 48 and the phase processing unit 50 tend to zero.
The local clock signal 35 is adjusted at a slower rate of change than the incoming data stream required offset adjustments, and the adjustment is controlled such that the local clock signal is in a holdover and remains within appropriate tolerance even if a synchronisation signal is not available.
In one alternative example, an additional number of clock generators 34 and associated phase processing units 50 are provided. Such an example allows multiple clock signals to be routed across the network through the network node concerned. This routing may be achieved by further multiplexing of the multiple clock signals. The routing may be governed by a timing schedule, which may be determined locally or by a central controller, such as a suitably modified SDN controller.
selecting one of the received radio frequency signals as a synchronisation signal by adjusting the reception parameters of the steerable antenna;
producing a digital data stream from the synchronisation signal using a local clock signal; extracting a reference clock signal from the digital data stream;
producing a reference comparison value by comparing the reference clock signal with the local clock signal; and
adjusting the local clock signal in dependence upon the reference comparison value.
The provision of a clock sync path across a mesh network as described above enables the dynamic adaptation of the path. In addition, multiple clock sync paths may be defined when appropriate. For example,
The second clock sync path 27 provides an alternative route for the synchronisation of the destination network node 206. However, the destination network node 206, needs only a single master clock reference, and so the destination network device 206 determines which of the received master clock signals, received via the first and second clock signal paths 26 and 27, is to be used. This decision may be made by a suitable adapted SDN controller, or locally in the network node. A switching unit in each node 20 is used to direct the reference clock signals appropriately.
The network node 3 of
The processing units 32A, 32B, 32C, 32D are interconnected by a switch unit 5 which operates to transfer data packets between the processing units 32A, 32B, 32C, 32D. In such a manner data packets can be routed through the network node 20. In addition, the switch unit 5 is connected with a local connection 6, for example a cell of a cellular communications network, or other local device.
Although the switch unit 5 is used for switching data packets through the network node for routing across the network, in the context of the present invention, it is the routing of clock synchronisation signals that is of interest. In this regard, the switching unit 5 is operable to switch such synchronisation signals between the processing units 32A, 32B, 32C, 32D.
In addition, any of the multiplexer outputs 35A′, 35B′, 35C′, 35D′ can be transmitted as a clock synchronisation signal from any of the antenna units, on any appropriate beams. The clock synchronisation signal may be part of a data transfer signal, or may be dedicated clock synchronisation signals.
The reference clock signals are communicated over dedicated clock transmission steerable directional radio frequency beams from the antenna units of the network node 20. In the exemplary case of the sync path 26 of
receiving a plurality of clock synchronising signals selecting one of the received clock synchronising signals as a reference clock signal, producing a reference comparison value by comparing the reference clock signal with a local clock signal, and adjusting the local clock signal in dependence upon the reference comparison value.
The network may have more than one POP node.
Steps in a method embodying this aspect of the present invention are illustrated in
at a first network node:
a. receiving a master clock signal;
b. synchronising a local clock signal of the first network node with the master clock signal; and
c. forwarding a first synchronisation signal to a second network node, the first synchronisation signal being dependent upon on or both of the local clock signal of the first node and the master clock signal;
at a second network node:
d. receiving the first synchronisation signal from the first network node; and
e. synchronising a local clock signal of the second network node with the first synchronisation signal, the first and second network nodes thereby forming a first clock region of the wireless mesh network.
Accordingly, embodiments of the various aspects of the present invention are able to provide improved techniques for the synchronisation of local clock signals of network nodes across a wireless mesh network having a plurality of such nodes.
In one example, an embodiment of the present invention is suitable for providing synchronising clock signals for different network slices. Network slicing is a concept in which a single hardware network is utilised by different operators using partitioning into multiple virtual networks allowing the operator to offer optimal support for different types of services for different types of customer segments. The key benefit of network slicing technology is that it enables network access as a service, which enhances operational efficiency while reducing time-to-market for new services. The network slices may be on any type of basis. For example, characteristics including latency or bandwidth requirements may be used to define different network slices. Other examples include time and location slicing. Network slices are isolated from each other in the control and user planes, and so embodiments of the present invention enable provision of the user experience of the network slice will be the same as if it was a physically separate network.
Number | Date | Country | Kind |
---|---|---|---|
1610136.2 | Jun 2016 | GB | national |
Number | Name | Date | Kind |
---|---|---|---|
20030128985 | Elbers | Jul 2003 | A1 |
20050277443 | Ozluturk | Dec 2005 | A1 |
20070226530 | Celinski | Sep 2007 | A1 |
20110280178 | Heifner | Nov 2011 | A1 |
20160095075 | Bin Sediq | Mar 2016 | A1 |
20160204507 | Karjalainen | Jul 2016 | A1 |
20170257836 | Ho | Sep 2017 | A1 |
Number | Date | Country |
---|---|---|
WO20060102558 | Sep 2006 | WO |
Entry |
---|
Suarez, R. et al., “Extending OpenFlow for SDN-enabled Synchronous Ethernet Networks,” Proceedings of the 2015 1st IEEE Conference on Network Softwarization (NetSoft), Apr. 13-17, 2015. |
PCT/GB2017/051666—International Search Report dated Sep. 11, 2017. |
Number | Date | Country | |
---|---|---|---|
20170359789 A1 | Dec 2017 | US |