OPTICALLY MODULATED LASER BEAM TRANSCEIVER SYSTEM
Field of the Invention
The field of the invention relates to lasers and in particular to optical modulation of lasers used for communication systems.
Background of the Invention
The use of lasers for the transmission of information in communication systems is generally known. Such use has typically been limited to amplitude or phase modulated systems that in use often attain a speed of several megabytes .
Methods have also been developed for multiplexing a number of channels of information onto a single laser beam using a process referred to as subcarrier multiplexing (SCM) . For instance, one prior art reference frequency translates a number of information sources through the use of an equal number of microwave oscillators and combines the frequency translated signals in a microwave combiner. A continuous-wave laser may be used to produce a laser carrier that is modulated by the combined microwave signals in an external modulator of lithium niobate.
A set of Fabry-Perot (FFP) filters may be used to reecover a laser frequency of each information signal. Each FFP filter may be followed by a photodetector to detect the individual information signals.
While SCM is effective, it suffers from a number of operational problems. For example, such systems
often experience an unacceptable error rate and must rely upon error detection and correction.
In order to reduce errors , the prior art has taught that carrier and intermediate frequencies of SCM sources must be suppressed. Suppression of such signals is believed necessary to improve a signal-to- noise (S/N) ratio of any transmitted information side bands and for efficient channel usage.
Because of the importance of high speed communications, a need exists for more effective laser- based communication systems. Such systems should have an inherent noise immunity that is based upon the modulation techniques used at the transmission source instead of error correction and detection.
Summary A method and apparatus are provided for distributing a plurality of information signals among a plurality of transceivers through an optical communication system. The method includes the steps of introducing an optical splitter into an optical waveguide of the optical communication system and transceiving' a refractively synchronized laser beam between the waveguide and a local transceiver. The refractively synchronized laser beam is modulated with an information signal of the plurality of information signals and a subcarrier. The subcarrier has .a predetermined spectral and amplitude relationship with the information signal.
Brief Description of the Drawings
FIG. 1 is a block diagram of a communication system under an illustrated embodiment of the invention;
FIG. 2 is a block diagram of a transceiver of FIG. 1;
FIG. 3 is a simplified signal flow diagram of the system of FIG. 1 ;
FIG. 4 is another simplified signal flow diagram of the communication system of FIG. 1; FIG. 5 is an optical router & switch of the communication syst4em of FIG. 1;
FIG. 6 is a signal processing diagram of the communication system of FIG. 1 ;
FIG. 7 is a block diagram of the signal divider of FIG. 6;
FIG. 8 is a simplified signal flow diagram of the communication system of FIG. 1;
FIG. 9 is a simplified signal flow diagram of a bi-directional communication system of FIG. 1; and FIG. 10 is another simplified signal flow diagram of a bi-directional communication system of FIG. 1.
Detailed Description of a Preferred Embodiment FIG. 1 is a block diagram of an optical communication system 10, generally, in accordance with an illustrated embodiment of the invention. Under the embodiment, one or more transceivers 14, 16, 18 may be
interconnected via an optical waveguide 12 (e.g., a fiber optic cable) or via free space and may engage in the exchange of data and information with other transceivers 14, 16, 18 using the medium of the waveguide 12.
The transceivers 14, 16, 18 may be one-way or two- way devices capable of either transmitting and/or receiving information through the waveguide or fiber 12 or free space. For example, one transceiver 14 may be a cable channel receiver receiving blocks of television signals (e.g., 148 channels per block) for distribution to local subscribers. A second transceiver 16 may transmit the blocks to the first transceiver 14.
Alternatively, the transceivers 14, 16, 18 may engage in the simultaneous exchange of video, voice, data, or any combination of information formats over the waveguide 12 or fiber or free space. The information may be compressed or uncompressed. Further, any transceiver 14, 16, 18 may transceive information with any other transceiver 14, 16, 18. FIG. 2 depicts an exemplary block diagram of a transceiver 14, 16, 18. As shown, each transceiver 14, 16, 18 may include a transmitter section 22 and receiver section 24. An optical splitter 20 may be provided as a means for interfacing with the waveguide 12.
The splitter 20 may be any optical splitting or coupling device capable of being placed in series with a waveguide 12 or fiber, without substantially interfering with transmission of optical signals traveling in either direction along an axis of transmission of the waveguide 12 (e.g., 90% of any
incident light passes through and 10% is diverted) . The splitter 20 may function to divert a portion of any light propagating along the axis of transmission of the waveguide 12 into the transceiver 14, 16, 18, and visa versa. Under one preferred embodiment, a 90:10 optical beam splitter has been found to function advantageously. In general, a number of such devices are known in the art and, as a consequence, the details of the splitter 20 will not be discussed any further. The transmitter 22 may be constructed as generally described and shown in copending U.S. Patent Application Serial Number 09/106,881, incorporated herein by reference. The receiver 24 may be constructed as generally described and shown in copending U.S. Patent Application Serial Number 09/106,881, also incorporated herein by reference.
FIG. 3 is a simplified block diagram of a portion of an optical communication system 10, where signal progression is shown occurring in a single direction (i.e., from left to right) . A receiver 24 may receive many frequencies, but may select only a few (e.g, one) frequency for processing. FIG. 3 shows a structure by which a number of information signals may be distributed among a number of users, where the transmission of any one signal is substantially transparent to the transmission of any other signal.
Information is shown in FIG. 3 as being - distributed to the user through a router & switch 42, 44. Unless clearly indicated otherwise, the router & switch 42 will be assumed to be a channel router (i.e., a switch) instead of a packet router.
Transmitter 30 may be a component of one or more upstream transceivers 14, 16, 18, transmitting signals for the benefit of users 48. The transmitted signals may each be a modulated block of TV signals Sm (e.g., 148 channels) that have been refraction multiplexed onto a subchannel of a transmitted signal using refractive synchronization technology (RST) and sampled, or upshifted, at a sampling clock frequency Cfn, as shown schematically in FIG. 4. Modulation and/or upshifting may be accomplished within a electro- optical differential multiplexer (e.g., an External Modulator by UTP SITU APE, Marconi LC 1000, etc.) .
Optical refractive synchronization is a process whereby a controlling clock subcarrier and information signal are used to modulate an optical signal.
Fundamentally, the process of refraction is the change of energy, direction or speed of a light beam which is propagating through a medium. In a first case, the change in direction may be a continuous bending of the light beam and of the subsequent change of the speed of light in the medium which is referred to as the index of refraction of the material. In a second case, there is an abrupt change in the index, polarity, or phase of the medium which directs the energy out of the medium, or changes the ability of light to pass through the medium, thereby absorbing the energy. This second case is the technique employed in optical refractiye synchronization. The light is passed through a crystal which has the refractive index changed abruptly, by imposition of an electric field, by passing a controlling signal through the crystal, which then causes the crystal to allow the light to pass through,
or be absorbed in the crystal at the frequency of the controlling signal. This optical modulation is performed, therefore, through control of the refraction and polarization signal of the optical crystal. By combining the frequencies of various sources and optically modulating them onto a respective clock frequency with an appropriate (i.e., predetermined) sampling frequency and amplitude ratio, it is possible to synchronize all of the information signals to their respective optical light beams. It then also be comes possible to combine them with other light beams with similar optical modulation at singularly different clock frequencies and to nest these optically modulated beams onto yet another yet higher clock frequency by using optical refractive synchronization.
Although refractive synchronization technology (RST) , as outlined herein, is superficially similar to the well-known concept of subcarrier multiplexing commonly used in the microwave transmission arena, it differs from traditional microwave subcarrier multiplexing in a number of aspects. First, the prior art accomplishes subcarrier multplexing under an rf format. More specifically, the prior art has not provided a mechanism through which optical signals could be multiplexed directly and coherently. Without a mechanism for coherently mixing optical signals, subcarrier multiplexing could not be extended^ to optical communications systems, without the step of first mixing the information signals under an rf format and then transmitting the previously mixed signals under an optical forma .
Further, refractive synchronization of optical signals involves substantially different processes than microwave subcarrier multiplexing. For example, RST is a significant advancement whose efficacy is dependent upon two fundamental relationships :
The first, relates the substantially exacting choice of the clock (subcarrier) amplitude with respect to the amplitude of the maximum spectral amplitude present in the information signal spectrum. [Note: In the conventional subcarrier multiplexing literature there were only heuristic rules like make the subcarrier amplitude at least 10 dB more than the signal spectrum maximum amplitude as opposed to the exacting mathematical analysis shown in Dr. Palmer's work. The second, relates the substantially exact required frequency of the clock (subcarrier) to the amplitude ratio determined by the first relationship. this frequency can be adjusted to account for baseband noise present in the information signal. [Note: this function is completely described in the paper (attached as Appendix I) by J. R. Palmer, Frequency Sampling Ratio as a Function of the Ratio of Amplitudes Between the Clock Frequency and the Modulated Frequency for Verisimilitude-Homodyne Laser Transmitter-Receiver.] In summary, the use of RST implies the adherence to well-defined mathematical schemes to determine the ampli tude ratios employed in the clock (subcarrier) and well-defined mathematical determination of the required frequency of the clock (subcarrier) as a function of the previously determined amplitude ratios and/or possible baseband noise in the system.
Once the signal from the laser 166 is modulated with the information signal by refractive synchronization, the signal may be upshifted within a second electro- optical differential multiplexer 170 (e.g., an External Modulator by UTP SITU APE, Marconi LC 1000, etc.) . Upshifting may be accomplished at any clock frequency with an appropriate sampling rate as determined from the Palmer Sampling Modulation Transfer Function equation as follows :
/ Osti ple
where A is the modulated signal amplitude, Ao is a clock frequency amplitude.
For the case where a noise signal is included as part of the baseband signal, the equation can be modified as follows:
where M is 8πδA/A0 and δ is baseband noise amplitude/2A0 and Jι(M) is a Bessel Function of the First Kind of Order 1. Further, a required minimum clock sampling frequency for A0 = (1/fosampie) fma where fmax is the highest frequency represented in the modulated signal A. From a practical point of view the best results have been found to occur where A/Ao<1.0 (i.e., the clock
amplitude AG is 2.0 to 7.0 times larger than the modulated signal amplitude A) .
As shown in FIG. 4, each block of signals Sm may be modulated onto a laser beam either by modulation of a junction current of a source DFB laser or by external refraction synchronization. The signals Sm may be provided under any number of formats (e.g., a first block of signals Si may be 148 TV channels under a FDM format, a second block of signal S2 may be a TI telecommunications trunk signal, a third signal S3 may be provided under the SONET format used in a wide area network (WAN) , etc) .
The modulated laser beam may then be sampled or upshifted using a subcarrier sampling clock Cfn having a magnitude 2 to 7 times larger than the maximum value of the modulation information spectrum of an underlying block of the blocks of signals. Sampling, or upshifting, of the blocks of information may be used to form each of the modulated blocks of information signals superimposed on a laser beam which may be, in turn, propagated among transceivers 14, 16, 18 using the appropriate medium.
Page 7 of Appendix I is a graph which relates sampling frequency, amplitude ratio and noise levels. The ordinate of the graph is the inverse of the sampling frequency. The abscissa is the signal/clock amplitude ratio.
To use the graph of page 7 of Appendix I, a user may enter the graph using a sampling frequency and amplitude ratio to recover a noise level that may be experienced by a particular system design. Alternatively, a user may select a maximum acceptable
noise level and sampling frequency and enter the graph with the noise level and sampling frequency to recover an amplitude ratio necessary to limit noise to that noise level. FIG. 11 depicts laser frequencies of modulated blocks of information signals 302, 306, 310, 318 that have been superimposed onto the laser beam which propagates among transceivers 14, 16, 18 along the waveguide 12. Also shown with each information block (e.g., 302) is a corresponding sampling clock Cfn
(e.g., 304), lying between the information blocks.
Sampling the modulated laser beam at non- coincidental clock frequencies Cfn, may be used to move each modulated signal block to a non-interfering (i.e., non-overlapping) subchannel spectral location with regard to every other subchannel through use of predetermined set of sampling or upshifting frequencies Cfn (e.g., Cfι=2.5 GHz, Cf2=3.6 GHz, Cf3=4.7 GHz, etc.). Each source of data Sm may be assigned one of the subchannel clock (subcarrier) frequencies Cfn. For example a first signal S]_ may be assigned a first clock Cf]_. A second signal S2 may be assigned a second clock Cf2 and so on.
The subchannels may then be optically combined, detected and modulated onto another laser beam to nest a block of subchannels. The detected block of subchannels may be modulated onto the other laser beam using external (diode junction) modulation or RST. The modulated block of subchannels may then be sampled at (upshifted to) a higher system sampling clock frequency Cfv (e.g., 100 GHz, 200 GHz, etc.) using optical
refractive synchronization to form a nested modulated block of subchannels superimposed on a laser beam, where each subchannel is defined by the information block and its associated subcarrier (clock) . Further, a first nested block of subchannels sampled at a first system clock frequency Cfx may be combined with a second nested block of subchannels at a second system clock frequency Cfx+ι in a splitter 20 of a downstream transceiver 34 (FIG. 3). Such combined blocks of subchannels (depicted by arrows 38 in FIG. 3) may propagate through the waveguide 12 and be decoded in downstream receivers (e.g., receiver 36).
Within any particular transceiver 14, 16, 18 of the system 10 (e.g., receiver 24 of transceiver 34 of FIG. 4) the nested block of subchannels may be recovered by downshifted using a mixer in the receiver 24 with the system clock Cfx for a predetermined block of subchannels, which are then returned to baseband. Similarly, a block of signals may be recovered for each subchannel of a block of subchannels by downshifting using a mixer in the receiver 24 with the subchannel clock Cfn. Synchronization to the system subcarrier or subchannel clock may be accomplished within the receiver through the use of a phase-locked-loop (PLL) , operating at the subcarrier frequency, and a downshifting mixer.
FIG. 5 depicts an optical channel routing structure 50 that may be disposed between the splitter 20 and a detector 80, 82 of the receiver 24 for routing subchannels of a block of subchannels. The routing structure 50 provides switching fabric for routing the
subchannels to selected channel decoders within the receivers 24.
Depicted within the channel routing structure 50 are two optical routing channels 100 102. The routing channels 100, 102 may be constructed using 8x8 photonic switches (e.g., UTP Model YB-150-180) 74, 76 or 16x16 photonic switches 70, 72. The use of 16x16 switches 70, 72 may be used to reduce the size of the switching structure where size is at a premium, or where the number of decoded channels is large. Where size is less critical, 8x8 switching devices 74, 76 may be used.
Included within the routing structure 50 are optical inline amplifiers 52, 54, 56, 66, 68 which function to increase the amplitude of the received optical signal. The inline amplifiers may be any appropriate erbium doped fiber amplifier (EDFA) device. Inline 50:50 beam splitters 60, 62 may be provided where two or more optical routing channels 100, 102 are provided. Mirrors 58, 64 may also be provided, where necessary to direct the optical signals to appropriate switching elements 70, 74.
Decoding of a block of signals may be accomplished by detecting a subchannel signal or nested block of subchannels in a corresponding photonic detector (photo sensor) 80, 82. The detected signals may then be frequency translated to a baseband frequency by first homodyning it with predetermined system clock frequency Cfx and/or a subchannel clock Cfn assigned to the channel.
FIG. 6 shows a block diagram of an optical decoder. As shown, a photonic detector (e.g., a
photodetector) 80, 82 may be followed by a signal divider (block selector) 84 and signal amplifier 86. The detector subsystem (FIG. 6) receives the optical signal from the optical router & switch (FIG. 5), converts the optical signal into an electrical signal for further processing by the receiver block selectors, separates the signal into the required multiple output signals, amplifies the separated signal and sends it to the corresponding receiver block selectors . FIG. 7 is an exemplary block diagram of a block selector of the signal divider 84 of FIG. 6. One or more block selectors of the signal divider 84 may be used to recover signals Sm from modulated blocks of signals or to recover modulated blocks of signals from a nested modulated block of subchannels. A block selector of the signal divider 84 may be provided for each nested block of subchannels and for each signal (e.g., Sι_, S2 • • • Sm) of a block of signals to be decoded. Different sampling frequencies may be assigned to each block selector.
To ensure that the sampling clock frequencies of the transmitter and receiver are locked to one another, a homodyning circuit (FIG. 7) is provided. The first input filter 102 is a low-pass filter set to extract one of the sampling frequencies carried on the laser beam. Specifically, this filter 102 eliminates signals in the frequency range corresponding to the upper sideband in the sampling receiver subsystem. The input signal is then split in the first splitter 104. The outputs are fed into the first mixer 108 and the second mixer 110. These signals contain the lower sideband and transmitted sampling clock frequency.
A sampling frequency generator 116 may be provided in each receiver block selector. This sampling frequency generator 116 is intentionally offset in its operating frequency from the sampling frequency generator in the respective transmitter. This offset frequency varies between 200 and 500 MHz (e.g., by 200, 300, 400 or 500 MHz) . The offset allows for the homodyning of the two signal sources (transmitter and receiver) and provides for an accurate regeneration of the transmitted sampling frequency. The purpose of the homodyning circuit is to compensate for any drift in the transmitted sampling frequency.
The output of the sampling frequency generator 116 is fed into the second splitter 118. The output of this splitter 118 feeds the second mixer 110 and the third mixer 120. The output from the second mixer 110 is the difference between the original transmitted sampling frequency and the receiver's sampling frequency. The difference is approximately 200, 300, 400 or 500 MHz. The difference may be the exact offset frequency between the transmitter and receiver's sampling frequencies, provided the transmitted sampling frequency has not drifted up or down in frequency.
The output of the second mixer 110 is fed through the third filter 112 to the other input to the third mixer 120. The third filter 112 is a narrow bandpass filter set to eliminate all signals except the difference between the two sampling clock frequencies (i.e., transmitter and receiver). The output of the third mixer 120 is an exact replication of the received transmitted (i.e., the transmitter) sampling frequency. The output of the
third mixer 120 is filtered in the second filter 114 and fed into the other input port of the first mixer 108. The second filter 114 is also a narrow bandpass filter set to eliminate all frequencies except the required sampling frequency used to demodulate the received data block.
The inputs to the first mixer 108 consist of the original received data block with its corresponding sampling frequency and the regenerated sampling frequency from the receiver block selector. Here the received transmitted sampling frequency and the receiver's sampling frequency cancel one another and all that is left is the required signal data (e.g., Sι_,
≤2 , S3 . . . or Sm) or modulated blocks of signals. This data is sent from the output port of the first mixer 108 to the output driver 106 for further signal processing. If the required output of the receiver block selector is RF, then the signal is sent directly to an output connector. If the required output is to be converted into another optical signal, then the signal is sent to a laser driver and converted back into its original optical format.
Amplifiers 86 are provided in the homodyning circuit as needed to keep the signals at the required levels for further processing. These amplifiers 86 may be either of a variable gain variety or of a fixed gain type.
Once a nested optical block of subchannels has been decoded, the individual radio frequency signals (i.e., Si, S2 • • • Sm) may be routed from an output of a receiver 24 (FIG. 3) through use of a router S_ switch 42 and associated controller 46 to a user 48. Control
of the router & switch 42 may be accomplished from external sources, or, from the decoded signal itself.
For example, where a signal (e.g., Si) is a telecommunications trunk line (e.g., a Tl line), a repeating data structure of interspersed data and control frames may be provided (e.g., 28 data channels, 2 control channels) . The control channels may be used to provide assignment and routing information for each of the other 28 channels. Further, it would be clear to a person of skill in the art that the Tl channel is bidirectional, allowing the flow of control or data in either direction over the Tl link.
Alternatively, where a channel Sm is structured for support of a WAN, packet switching may be used for the transmission control. Where packet switching is used each packet may contain a packet header including identifying information identifying a target of each packet .
Such control information received from the user 48 (for information to be transmitted over the system 10) as channel associated signaling or intercepted by the router & switch 42 (for information received from the receiver 24) and forwarded to the controller 46. The controller 46, in turn, may use such information in conjunction with a look-up table to identify a pair of switch ports of the router &. switch 42 for successfully routing of the information to the desired destination.
Turning now to the system 10, as a whole, a series of examples will be offered as a means of understanding the operation of the system 10. In a first example
(FIG. 8), it will be assumed that optical signals flow in only a single direction (i.e., left to right).
Further, it will be assumed that at least some channels Sm may be dedicated to specific purposes.
For example, channel Si of information channels S-, may be dedicated for the transmission of information from transceiver 14 to transceiver 16 of FIG. 1 using subchannel sampling frequency Cf ___ . Similarly, channel
S2 may be dedicated for the transmission of information from transceiver 16 to transceiver 18 using subchannel sampling frequency Cf2 • Channels S3 and S4 may be dedicated for transmitting information from transceiver 14 to transceiver 18 using sampling frequencies Cf3 and
Cf4. A system sampling frequency of Cfx may be used as a nesting sampling frequency.
Under these conditions, the nested block of information received by the second transceiver 16 (FIG. 8) would include signals Si, S3 and S . The nested block of subchannels may be recovered using the homodyning circuit (FIG. 7) and the system sampling frequency Cfx. The transceiver 16 may recover the first signal block Si using its assigned subchannel sampling frequency Cf]_. The signal S_ may be forwarded to its assigned destination through an associated router & switch 42.
The transceiver 16 may also function to forward signals S3 and S4. The transceiver may sample S2 at its sampling frequency Cf2 and optically combine it with S3 and S4. The nested block of S2 , S3 and S4 may be sampled at a new system frequency of Cfx+ι_ and forwarded through the splitter of the transceiver 16 to transceiver 18.
As an alternative, nesting is not necessary and may not be used for implementation of the above example. A first laser may be modulated with the first information block Sι_ and sampled at the first blocks sampling frequency Cfι_ to provide a first modulated laser beam. A second and third laser may similarly be used for transmission of the third and fourth information signals S3, S4 _ The first, second and third modulated laser beams may be optically combined and transmitted to the second transceiver 16. The second transceiver 16 may recover the first information signal Sι_ and transmit the second information signal
S2 in its place to the third transceiver 18.
After leaving the transceiver 16, the beam contains the original set of sampled and modulated signals S3 and S4, plus the newly sampled and modulated signal S2 • This new set of signals can be accessed anywhere on the network 10 using a standard transceiver 14, 16, 18 which has been designed to demodulate the clock frequencies Cfn and provide any of the signals that were added subsequent to the initial transmission. As discussed, the system of FIG. 8 is a linear system, transmitting in one direction.
For bi-directional networking, schematically represented in FIG. 9, the concepts discussed above still apply. To make the linear network of FIG. 8 be- directional, the choice of clock frequencies is an important aspect of system performance. The basic components of the bi-directional transceivers 200, 204, 206 of FIG. 9 may be found in the transceivers 14, 16, 18 of FIG. 1. For example, the transceiver 200 may
include two transmitters 22. Other transceivers (e.g, 206) may contain one or more receivers 24. Signal from routers 204, 206 may be modulated using RST and combined in a Cassegrain optical combiner 202. In the network of FIG. 9, the transceiver 200 may be configured with two ports for receiving information. One port may be connected to a router & switch 208 which can add data which is shown as being transmitted westbound and another port may be connected with a router & switch 210 which supply data which may be transmitted eastbound. The transceiver 206 in the bidirectional system can receive data that is traveling either east or west and provide that data to routers 212. Where full bi-directional operation is required each transceiver 214 (FIG. 10) may be provided with two transmitters 22 and two receivers 24. Two optical routers 50 may also be provided for routing purposes. With the transceiver 214 of FIG. 10, data may be transmitted in both directions without mutual interference even using the same clock. The same laser frequency may also be used where there is a phase offset among the lasers. Further, the transceivers 214 may also using bi-directional time division multiplexing as a means of further sharing channel capacity.
As a further method of operation of the system 10, the transceivers 14, 16, 18, 200, 214 may be operated under an ethernet format. Ethernet-like operation is achieved by having a bi-directional system 10 with a laser at each end of the fiber 12 to create the "ether" for signal transmission. This actually creates
multiple "ethers", one for each allowable frequency band. There are two distinctly different implementations possible, a fully asynchronous method and a simple master-clock method. In the master-clock method, one or both ends nodes (e.g., transceivers 14, 18) transmit the various reference frequencies continuously, and all the other nodes (e.g., transceiver 16) acquire and use these frequencies. The advantage of having only one of the end nodes transmit the reference frequencies is that it greatly simplifies the construction of the intermediate nodes because the clock frequency is the same for both directions of data transmission. When a node has data to be transmitted, it first determines if there is any incoming data present in the ether. If the ether is clear, the node proceeds to transmit the data but still monitors the incoming data paths for data from another node. If a collision occurs, the node waits a random interval and tries again. Note that each node never sees its own data transmitted in the ether - all other nodes are "downstream" in the relevant direction of data transmission, so collisions are detected by data being received at the same time that data is being transmitted (plus a small guard-time dictated by the size of the system) .
In the fully asynchronous method, each node transmits its own reference frequency whenever it has data to send. The node first determines if there is a similar reference frequency being received from another node. If not, the node sends out its own reference frequency. After a preset interval, during which the destination node acquires the reference frequency, the
transmitting node proceeds to send the data, but still monitors the incoming data paths for a similar reference frequency from another node. After the data has been sent, the node also stops transmitting its local reference frequency. If a collision occurs, the node waits a random interval and then tries again. Note that as in the master clock method, each node never sees its own reference frequency transmitted in the ether, so collisions are detected by a similar reference frequency being received at the same time that the local reference frequency is being transmitted.
The system 10 may also operate under a token ring format. To achieve a true token-ring like system, each node must be able to remove data from the laser beam
(instead of just extracting it) , thereby restoring the original unmodulated beam. This can only be done in the optical domain, for example by pinholing the laser beam to strip off the modulation sidebands. After the beam has been stripped, the data may be examined for destination and used locally, if so intended. Any data needing to be passed through the node then may be remodulated back onto the outgoing laser beam. In practice, this is achieved by receiving the complete spectrum into the RF domain from the incoming beam, filtering out the frequency bands to be removed, and then transmitting the remaining spectrum along with any new data to be transmitted. One node in the ring is the "master" which generates both the raw laser beam and the reference frequencies. Each node in the ring looks at the received data stream to detect the "token" being passed around the ring.
Until the token is received each node that wishes to transmit just passes through everything it receives, that is not addressed to it. If the node has no data to send, it transmits the received token to the next node in the ring. Otherwise, it transmits the data waiting to be sent and then transmits the token. During data transmission, any data received in the appropriate frequency band is removed. Note that each frequency band is a separate ring within the fiber. If stripping the data results in too much attenuation of the laser beam, it may be necessary to use optical amplifiers at strategic points around the ring, or even to regenerate the beam using multiple lasers distributed round the ring. In the extreme case, the network degenerates into a circular sequence of point-to-point links, each with its own laser. An alternative approach is to use an unmodulated beam in the reverse direction and splitting it at each node to use part of it to boost the power of the stripped beam; this technique could be extended to eliminate stripping altogether.
A variant token-ring like system can be achieved by double-frequency allocation. Each "ring" within the fiber is assigned two reference frequencies. The first reference frequency is used for passing the token around the ring and for sending data from an originating node around to the master node.
The second reference frequency is used for sending the data from the master node (no matter where it originated) around the whole ring. This removes the necessity for nodes to strip data from the laser beam. Effectively, the network is similar to an orbital
satellite system with one frequency used for the uplink and a different frequency used for the downlink. Since the "token" transmitted is now send by all downstream nodes, the transmitter must encode its own identifier into the taken, and each node has to know which other node is immediately upstream so that it can discard inappropriate tokens .
An alternative variation is to utilize the direction of transmission around the ring instead of using two reference frequencies. One direction may be used for transmission to the master node, and the other direction is used for distribution out from the master node. With this two-directional method of operation, the functions of the "master" node can be split between two separate sites provided the tail-end master can transmit the token back to the head-end master on the reverse channel. The "ring" is then from the head-end node to the tail-end node in one direction and then from the tail-end node back to the head-end node in the other direction. Note that this is still substantially different from a circular ethernet by virtue of the tokens-passing mechanism versus the asynchronous collision-detecting mechanism, and also by the data reception being limited to only one direction of data transmission.
In another embodiment, a multi-drop system 10 can be achieved in a manner similar to token-ring,, like system described above. In this variation, the token- passing control mechanism is replaced by a poll/response system. The master node sends out a
"poll" token to each node in turn and gets back either a data block or a no-data token. The data block itself
can be a request for channel bandwidth, and then the system transforms into a CDMA-like system. Forcing the data blocks and poll/response tokens to be fixed-size entities allows the network to be transformed into a TDMA-like system. For a network 10 with a small number of nodes, each node may be allocated its own reference frequency, and then the network transforms into an FDMA-like system. Combinations of these various methods can also be implemented to suit specific network requirements. In particular, a single master node can have multiple fiber links attached, resulting in a star-like network (or even a concentrator) .
A specific embodiment of a method and apparatus of distributing a plurality of information signals among a plurality of transceivers through an optical communications system according to the present invention has been described for the purpose of illustrating the manner in which the invention is made and used. It should be understood that the implementation of other variations and modifications of the invention and its various aspects will be apparent to one skilled in the art, and that the invention is not limited by the specific embodiments described. Therefore, it is contemplated to cover the present invention any and all modifications, variations, or equivalents that fall within the true spirit and scope of the basic underlying principles disclosed and claimed herein.