WO2007021199A1

WO2007021199A1 - A method and system for polarized signalling in an optical switched network

Info

Publication number: WO2007021199A1
Application number: PCT/NO2006/000300
Authority: WO
Inventors: Steinar BJØRNSTAD; Marius Clemetsen
Original assignee: Telenor Asa
Priority date: 2005-08-19
Filing date: 2006-08-21
Publication date: 2007-02-22
Also published as: NO20053902L; NO20053902D0

Abstract

A method for signaling in an optical packet switched network, wherein unicast and multicast traffic is modulated with different state of polarization, incoming signals are splitted according to the state of polarization, unicast signals are guided to a unicast switch (10), multicast signals are guided to a multicast switch (15), output signals from the unicast switch (10) are combined with output signals from the multicast switch (15) in polarization beam combiners (20, 25).

Description

A METHOD AND SYSTEM FOR POLARIZED SIGNALLING IN AN OPTICAL SWITCHED NETWORK

Field of the invention

The present invention relates to optical switching in communication networks .

Technical background

With the introduction and development of an optical network, it is an object to reduce costs for transmission of data in telecommunication and data communication networks. For the reduction of costs it is important to reduce the number of signal conversions between optical and electronic form. Such a reduction will reduce the number of components in the network elements and reduce the need for electronic signal processing.

A particular problem in present optical networks is the separation of multicast and unicast traffic. It is difficult to implement efficient switching of traffic and copying of multicast traffic in optical networks.

Unicast traffic includes normal point-to-point data traffic, e.g. Internet traffic. Multicast traffic includes TV-broadcasting and Video-on-Demand over cable networks. The deployment of TV over IP based infrastructure has started, and may well be the preferred method of delivering TV services in the future. The network operators are working on converging their parallel network infrastructures for voice, TV and Internet into a common multiservice IP network, supporting all existing services of today.

Multicast and unicast traffic have widely different requirements to the communication channel. Packet switched networks are well suited for transferring unicast traffic, as they have high throughput efficiencies due to the statistical multiplexing. However, networks supporting TV- broadcasting is subject to very strict Quality of Service requirements. The most important parameters are packet loss

5 ratio, timing jitter and availability. Subscribing TV viewers tend to have high demands to the quality of a service, not tolerating long outage times or significant artefacts in the TV-picture. The closer to the source, the stricter the QoS requirements become. This because e.g. a o packet loss close to the broadcaster source will affect all TV-viewers fed from that source. In a packet switched network it is hard to achieve the Quality of Service required by multicast TV traffic. The continuous streaming of information precludes the use of the ordinary error s correction schemes used in packet switched networks, such as retransmission or rerouting of packets.

Circuit switched networks may support the quality requirements of TV-broadcast networks. These networks use static multiplexing, there are no queuing mechanisms and o thus packet loss and packet jitter is eliminated. However, circuit switching is also known for not supporting bursty packet arrivals, hence capacity utilization is far from optimal when transmitting video and Internet traffic that may have a bursty traffic pattern. Hence, when combining 5 services in a multiservice network, capacity utilization will be low, precluding a cost-effective network implementation.

A number of optical switching matrix designs are proposed in the literature. Some of these designs support multicast, o e.g. the broadcast and select architecture used in the EU- funded project "keys to optical packet switching" (KEOPS) . This design does, however, need a very high number of components if switches with many fibre and wavelength inputs are to be implemented. Hence, the design does not 5 scale well . Another design, based on combining Arrayed Waveguide Grating Router (AWGRs) with tunable optical wavelength sources or converters, may provide better scalability. This type of design is however known not to support multicast .

Brief summary of the invention

It is an object of the present invention to devise a method and system for optical switching which may handle both multicast and unicast traffic in a cost-effective way. This is done by separating unicast and multicast traffic in the switch, based on modulating multicast and unicast traffic with different state of polarisation, as claimed in the appended claim 1. Then, different implementations can be used allowing unicast and multicast traffic to be processed separately. Further, the invention provides for efficient utilization of wavelengths by re-using existing unicast wavelengths for multicast, and also providing many multicast wavelengths .

The invention also relates to a system using the above method, as claimed in claim 8. Embodiments of the invention appear from the dependent claims .

Brief description of the drawings

We will now describe the invention in detail, in reference to the appended drawings, in which:

Fig. 1 illustrates schematically a switch employing the inventive method for separating unicast and multicast traffic,

Fig. 2 shows an example of a node design supporting the inventive method. Detailed description of the invention

By modulating unicast and multicast traffic in different state of polarisation (SOP) in an optical fibre, it is possible to completely separate unicast and multicast in optical switches. The signals may be modulated on the same wavelengths. Employing SOP for multicast/unicast separation, a polarisation beam splitter will split the signals independent of the wavelength; hence, bands of wavelengths may be separated simultaneously. The principle is illustrated in figure 1.

At the input of the switch one Polarisation Beam splitter (PBS) 1, 5 is employed for each fibre I₁, I_N. Unicast traffic is guided from each PBS 1, 5 to the unicast sub- switch 10, while multicast traffic is guided to the multicast sub-switch 15. At the output of each sub-switch 10, 15, signals are guided to respective Polarisation Beam Combiners (PBC) 20, 25 that combine the signals for the respective output fibre, fibre 1 - fibre N, Oi, O_N. While only two fibre channels are shown on the figure, several more channels (fibres) may be used.

The multicast and unicast signals may be interleaved in time in addition to polarisation, hence polarisation time division multiplexed. Cross-talk between the polarisation multiplexed signals are then minimized or avoided. The unicast signals may alternatively be polarisation multiplexed, allowing simultaneous transmission in both states of polarisation. This principle may be preferable if the cross-talk is tolerable. The separation of two streams (orthogonal polarisation) is simple, and does not require expensive equipment, and it makes it possible to implement one-to-many multicast distribution using totally different concepts than those used for unicast. As multicast may be carried on the same wavelengths as unicast, it may not be necessary to set aside specific wavelengths for multicast or to add new wavelengths to the system. Existing optical amplifiers with low sensitivity to polarisation variations can be reused.

Alternatively each input is de-multiplexed into wavelengths, and PBSs are employed on each individual wavelength, or on wavelength bands, containing several wavelength channels .

In Fig. 2 it is illustrated a node design employing the inventive switching method described above. The node supports both optical unicast and multicast by employing segregation by optical polarization combined with multicasting through an optical star. In the figure N is the number of fibres, n is the number of wavelengths in each fibre, and m is the number of multicast sources.

Each fibre is connected to a de-multiplexer Dmux at the input. Each de-multiplexer splits the incoming signals into different wavebands X₁ - X_n. Each waveband is guided to respective tunable optical wavelength converters TWC which feeds an Arrayed Waveguid Grating Router AWGR. At the output of the AWGR, corresponding TWSs guide the signals to multiplexers that multiplex the signals into fibres 1 - N. A passive optical star is used to connect all links that are part of the multicast distribution tree . The optical star will split the incoming signals in a passive way, and duplicate it to all of its output ports .

The multicast property of the optical star network architecture is well known, and reservation-based protocols are proposed in the literature ["A Reservation-Based Multicast Protocol for WDM Optical Star Networks" , Naik, K.; Wei, D. S. L.; Krizanc, D.; Kuo, S. -Y.; Selected Areas in Communications, IEEE Journal on , Volume: 22 , Issue: 9 , Nov. 2004 Pages: 1670 - 1680] . The star architecture should be considered as only one example among several possible multicast architectures that can be employed in the multicast section of the switch. Furthermore, the switch may be a hybrid design, combining electronics and optics. Electronics and optics may (or may not) be combined in each of the switch sections, the multicast part and the unicast part. The switch may also be implemented employing e.g. only optics in the multicast part, and electronics in the unicast part, or vice versa. The choice of implementation method is given from the technology found most suitable and cost-effective for the given purpose.

Nodes at the edge of the network may use tuneable lasers to be able to choose a specific multicast wavelength. In this scenario, an option is to use other bit-rates and/or modulation/data formats for multicast traffic than for unicast, as long as the end nodes are compatible and agree on the encoding. Hence, the principle is optically transparent . A likely scenario may be to have a set of multicast wavelengths with low bit-rate encoding for reception by cheap low-end interfaces, whereas other wavelengths carry high bit-rate traffic from more expensive high bit-rate equipment.

This simple multicast implementation may give many possible wavelengths for multicast, and coarse-grained control may be sufficient for highly realistic scenarios. An example of this would be the distribution of the most popular TV- channels to all the nodes at the edge of the optical network (that connects to residents) . It could also be used in cases where all multicast traffic from one or a few nodes should be received by just a few nodes. If, for instance, we carry the traffic of another ISP between a few edge nodes, we could aggregate multicast traffic in the edge that we know all the other edge nodes are interested in, and send it on the same wavelength as the unicast traffic for that ISP, but with different polarization.

It would also be possible to use much more complicated designs where traffic for different multicast groups (with different receivers) are multiplexed on the same wavelength, and signal the time-slot in the time-division multiplexing scheme to edge nodes, so that they know when to receive at the specific wavelengths.

However, in such a system there is a need for some form of admission control to the shared media.

Some possible examples of admission control schemes :

• Only specific servers are able to inject multicast traffic - and only one for each wavelength.

• There is a media controller that explicitly gives access to the media for a limited time. The identity of the end node that should be able to send next is encoded in a message that is appended to the end of each transmission (by the controller) . The media controller is addressed by edge systems using unicast to signal interest to send. The signalling is carried out independently for each wavelength.

• There is a dedicated signalling wavelength (multicast) used to set up multicast transmissions on the other wavelengths .

Claims

C l a i m s

1. A method for signalling in an optical switched network, characterized in that unicast and multicast traffic is modulated with different state of polarisation.

2. A method as claimed in claim 1 , wherein the unicast and multicast traffic is modulated with orthogonal states of polarisation.

3. A method as claimed in claim 1, wherein unicast and multicast traffic is modulated on the same wavelengths.

4. A method as claimed in claim 1, wherein incoming signals are split according to the state of polarisation, unicast signals are guided to a unicast switch, multicast signals are guided to a multicast switch, output signals from the unicast switch are combined with output signals from the multicast switch in polarisation beam combiners .

5. A method as claimed in claim 1 , wherein incoming signals are de-multiplexed according to wavelengths, the signals in each wavelength band being split according to the state of polarisation, unicast signals are guided to a unicast switch, multicast signals are guided to a multicast switch, signals from the unicast switch are combined with signals from the multicast switch and the combined signals being multiplexed onto optical fibres.

6. A method as claimed in claim 1, wherein unicast signals and multicast signals are interleaved in time, in order to prevent cross-talk between the signals.

7. A method as claimed in claim 1, wherein unicast and multicast signals are polarisation multiplexed.

8. A system for switching signals in an optical switched network, wherein unicast and multicast traffic is polarisation modulated according to the method claimed in claim 1, c h a r a c t e r i z e d i n that said system includes a number of polarisation beam splitters (1, 5), each with an input port (2, 6), a first output port (3, 7) and a second output port (4, 8) , the first and second output ports having different polarisations, the input port (2, 6) of each polarisation beam splitter (1, 5) being connected to an input optical fibre (I₁, I_N) , the first output port (3, 7) of each polarisation beam splitter (1, 5) being connected to an input port (11, 12) on an unicast switch (10) , said unicast switch (10) having a number of input ports (11, 12) for receiving signals and a number of output ports (13, 14) for delivering signals, the second output port (4, 8) on each polarisation beam splitter (1, 5) being connected to an input port (16, 17) on a multicast switch (15) , said multicast switch (15) having a number of input ports (16, 17) for receiving signals and a number of output ports (18, 19) for delivering signals, each output port (13, 14) on the unicast switch (10) being connected to a first input port (21, 26) on a polarisation beam combiner (20, 25) , each polarisation beam combiner (20, 25) having a first input port (21, 26) and a second input port (22, 27) as well as an output port (23, 28) , the first and second input ports having different polarisations , each output port (18, 19) on the multicast switch (15) being connected to the second input port (22, 27) on each of said polarisation beam combiners (20, 25) , the output port (23, 28) of each polarisation beam combiner (20, 25) being connected to an output optical fibre (O₁, O_N) .

9. A system as claimed in claim 8, wherein the multicast switch is a passive optical star.