METHOD AND ROUTING SYSTEMS FOR OPTICAL DATA TRANSMISSION
FTELD AND BACKGROUND OF THE INVENTION The present invention relates to optical data transmission networks and, more particularly, to a method for combined time-domain and wavelength-domain multiplexing of data packets, and to associated routing systems.
There are two kinds of digital communications networks in common use: electronic and optical. Optical data transmission has the advantage over electronic data transmission of relative immunity to interference and crosstalk and a significantly higher bandwidth. For several messages to share the same physical communications channel, the messages must be multiplexed. Several multiplexing schemes are commonly used in electronic networks, including time division multiplexing (TDM), frequency division multiplexing (FDM) and code division multiplexing (CDM). FDM is straightforward to implement in an optical network, where it is also commonly referred to as "wavelength division multiplexing" (WDM): each message is assigned its own carrier wavelength, and simple wavelength-sensitive optical components such as diffraction gratings are used to sort out the different messages. Several papers were presented at the 1998 Conference on Lasers and Electro-Optics (San Francisco CA, May 3-8) on optical TDM and CDM implementations.
In both electronic and optical networks, all but the shortest messages are transmitted as several discrete packets, according to a variety of well-known protocols such as TCP/IP. Figure 1 shows a typical format of a data packet 10: an address block 14, which indicates the destination of data packet 10 by having a destination address encoded therein, followed by a data block 12 wherein the portion of the message contained in data packet 10 is encoded. The convention in Figure 1, and in other Figures below, is that data packets are transmitted from left to right, so that, for example, address block 14 reaches a destination node in a network before data block 12. Both the address and the message are encoded digitally, as sequences of regularly spaced electronic or optical pulses, with, for example, the presence of a pulse indicating a binary 1 and the absence of a pulse indicating a binary 0. The first bit of address block 14 is on the right side of address block 14, followed by the rest of the bits of address block 14, and similarly for data block 12. Both address block 14 and
data block 12 have fixed and known transmission times. Typically, address block 14 and data block 12 both are formatted with a fixed number of bits, and their transmission times are the number of bits multiplied by the time interval associated with one bit. There may be a time delay between the end of address block 14 and the beginning of data block 12.
Figure 2 is a block diagram of an optical system 20, for decoding address block 14 of a data packet 10 and routing data packet 10 accordingly, as described in copending U. S. Patent Application Ser. No. 09/118,372, which is incorporated by reference for all purposes as if fully set forth herein. Single arrows indicate the flow of optical data. Double arrows indicate the flow of electronic data. System 20 enables the implementation of asynchronous transfer mode in an optical transmission network. A data packet 10 enters system 20 from an input waveguide 34. A 1 x 2 switch 22 diverts address block 14 to the left branch of Figure 2 and data block 12 to the right branch of Figure 2. Address block 14 is decoded in a decoder 24. According to the decoded address, an electronic signal is sent to a switch control 33 instructing switch control 33 to set a switch array 32 to direct data packet 10 to one of a plurality of output ports 36. Because the decoding process in decoder 24 destroys, or at least rearranges, address block 14, the decoded address is sent to an address restoration unit 26 which reconstructs address block 14. Depending on the type of optical network, the reconstructed address block 14 may or may not be identical to address block 14 that was input to decoder 24. Meanwhile, data block 12 is transmitted along the right branch of Figure 2. A delay unit 30 is provided to retard the transmission of data block 12 for the amount of time needed to decode and reconstruct address block 14, to preserve the built-in delay between address block 14 and data block 12. Finally, address block 14 and data block 12 merge to reconstruct data packet 10, which enters switch array 32 and then exits switch array 32 along the appropriate output port 36.
The duration of data packet 10 is the sum of the duration of data block 12, the duration of address block 14, and the time delay between the end of address block 14 and the beginning of data block 12. As data packet 10 traverses an optical transmission network, the network pays attention only to the contents of address block 14, while ignoring the contents of data block 12. If data block 12 and address block 14 could be overlapped in time, the bandwidth of the network could be increased accordingly.
DEFINITIONS
As used herein, the term "carrier wave" refers to an oscillatory wave of a single frequency and a single wavelength. Although the present invention is directed primarily towards communications using waves of electromagnetic radiation, and more particularly to infrared waves, the scope of the present invention includes all waves that can be modulated to carry messages, including but not limited to acoustic and elastic waves.
Two carrier waves are said to be "distinct" if they differ in at least one time- independent property. For example, two carrier waves of two different wavelengths are distinct, and two carrier waves of the same wavelength but of two different polarizations are distinct. Indeed, WDM is based on the use of different carrier waves that are distinguished by frequency to carry different messages. In principle, two carrier waves of identical wavelengths and polarizations but mutually shifted in phase by 90° also are distinct; but because the timing of the modulation of such carrier waves in optical communications must be accurate to within several femtoseconds, which is attainable with current technology only with great difficulty, such "phase multiplexing" is not used in optical communications.
As used herein, the term "spatial channel" refers to a spatially contiguous portion of a medium that supports the propagation of one or more carrier waves. For example, a single optical waveguide is a single spatial channel for the propagation of infrared waves. The term "communication channel" refers to the spectral region occupied by one or more modulated carrier waves to transmit a data packet. For example, if data packet 10 is formed by modulating a 1550 nm carrier wave at a data transmission rate of 20 GHz, then, in principle, data packet 10 can be transmitted in a communications channel that spans the spectral region from 1549.9 nm to 1550.1 nm. (In practice, the spectral widths used are wider than this, to prevent crosstalk.)
Two message blocks (data blocks or address blocks) are said to be transmitted "substantially simultaneously" if they overlap at least partially in time.
SUMMARY OF THE INVENTION
According to the present invention there is provided a method for sending a message to a destination identified by an address, including the steps of: (a) modulating a data carrier wave with the message, thereby forming a data block; (b)
modulating an address carrier wave with the address, the address carrier wave being distinct from the data carrier wave, thereby forming an address block; and (c) transmitting the data block and the address block substantially simultaneously towards the destination along a common spatial channel. According to the present invention there is provided in an optical data transmission network, a system for routing a data packet to an output port of a node of the network, the data packet including a data block and an address block, the system including: (a) a passive switch for separating the data block and the address block; (b) a decoder for receiving the address block from the passive switch and decoding the address block to provide a decoded address; and (c) a switch array for selecting the output port according to the decoded address and transmitting the data packet at the selected output port.
According to the present invention there is provided a method for sending a message to a destination identified by an address including a plurality of address bits, including the steps of: (a) modulating a data carrier wave with the message, thereby forming a data block; (b) modulating each of a plurality of mutually distinct address carrier waves with a different at least one of the address bits, thereby producing, for each of the address carrier waves, an address block; and (c) transmitting the data block and the address blocks towards the destination, the address blocks being transmitted substantially simultaneously along a common spatial channel.
According to the present invention there is provided in an optical data transmission network, a system for routing a data packet to an output port of a node of the network, the data packet including a data block encoded in a data carrier wave and a plurality of address blocks encoded in a corresponding plurality of address carrier waves, the carrier waves being mutually distinct, the system including: (a) a passive switch for separating the data block and the address blocks from each other; (b) a decoder for receiving the address blocks substantially simultaneously from the passive switch and decoding the address blocks to provide a decoded address; and (c) a switch array for selecting the output port according to the decoded address and transmitting the data packet at the selected output port.
According to the present invention there is provided a method for sending a plurality of messages to a corresponding plurality of destinations, each of the destinations being identified by an address including a plurality of address bits, the
method including the steps of: (a) for each message, modulating a data carrier wave with the message, thereby forming a data block, the data carrier waves being mutually distinct; (b) for each address, modulating each of a plurality of address carrier waves with a different at least one of the address bits, thereby producing, for each of the address carrier waves, an address block, all of the address carrier waves being mutually distinct and distinct from all of the data carrier waves; thereby forming, for each message, a data packet including the data block and the address blocks associated with the each message; and (c) transmitting each the data packet towards the destination, two of the data packets being transmitted substantially simultaneously along a common spatial channel.
According to the present invention there is provided in an optical data transmission network, a system for routing each of a plurality of data packets to one of a plurality of output ports of a node of the network, each data packet including a data block encoded in a data carrier wave and a plurality of address blocks encoded in a corresponding plurality of address carrier waves, the data carrier waves and the address carrier waves all being mutually distinct, each data carrier wave and the associated plurality of address carrier waves defining a communication channel of the network, the system including: (a) a passive input switch for separating the data packets from each other; and (b) for each communication channel, a subsystem for receiving data packets associated with the each communication channel from the passive switch and determining to which one of the output ports each the data packet associated with the each communication channel is to be routed.
The basic idea of the present invention is to include two or more distinct carrier waves in the same communication channel. One of the carrier waves, referred to herein as the "data carrier wave", is used to carry the data block of a data packet. The destination address of the data packet is encoded in one or more address blocks, and the other carrier waves, referred to herein as "address carrier waves", are used to carry the address blocks of the data packet. Figure 3 shows a data packet 110 of the present invention that includes a data block 112 and an address super-block 114. Address super-block 114 includes three address blocks 116, 118 and 120. As in Figure 1, the horizontal axis of Figure 3 is time; but the vertical axis of Figure 3 is wavelength. Data block 112 is modulated onto a carrier wave of wavelength λ address block 116 is modulated onto a carrier wave of wavelength λ2, address block
118 is modulated onto a carrier wave of wavelength λ3 and address block 120 is modulated onto a carrier wave of wavelength λ4. Because blocks 112, 116, 118 and
120 are modulated onto distinct carrier waves, blocks 112, 116, 118 and 120 can be transmitted substantially simultaneously on the same spatial channel, i.e., overlapping in time, as shown.
In system 20, 1 x 2 switch 22 is an active switch. Because the data blocks and address blocks of the present invention are carried on distinct carrier waves, the analogous input switch in an optical decoding and routing system of the present invention is a passive switch. In embodiments of the present invention in which the carrier waves are distinguished by wavelength, such a passive switch is based on passive elements such as add-drop grating elements, array waveguide gratings or multilayer reflective dielectric films. The timing of the operation of an optical decoding and routing system of the present invention is such that the data and address blocks of any given data packet are transmitted substantially simultaneously at the appropriate output port.
The simplest way to encode an address according to the present invention is to use one address block per address bit. The duration of an address bit then is comparable to, and could be almost as long as, the duration of the entire data block, rather than being only as long as a single data bit. The decoder of a corresponding optical decoding and routing system of the present invention includes as many input ports and output ports as there are address blocks. The passive input switch of an optical decoding and routing system of the present invention directs each address block, via a separate spatial channel, to a corresponding input port of the decoder, where a transducer senses the presence (address bit=l) or absence (address bit=0) of incident radiation and either generates (address bit=l) or does not generate (address bit=0) an electronic signal accordingly. A suitable light source, such as an appropriately tuned diode laser, is used at the corresponding output port of the decoder to regenerate an address block that encodes a "1" bit.
In other address encoding schemes, each address block is formed by modulating a corresponding address carrier wave with two or more address bits, as in the prior art, and decoders similar to those described in co-pending U. S. Patent Application Ser. No. 09/118,372 are used to decode the address.
The use of multiple distinct carrier waves enables the present invention to support the simultaneous transmission of two or more data packets on the same spatial channel. A single communication channel, including the spectral regions associated with one data carrier wave and one or more address carrier waves, is assigned to each data packet. In an optical decoding and routing system of the present invention, the passive input switch separates the data packets from each other and directs the various blocks of each data packet to a decoding subsystem that is dedicated to that data packet's communication channel. In one embodiment of the optical decoding and routing system, all the subsystems direct the data packets they receive to a single switch array, similar to switch array 32, that receives input from all the communication channels. Because such a switch array can direct only one data packet at a time to a particular output port, if two subsystems determine that two simultaneously received data packets should be directed to the same output port, then the transit by the data block of one of the data packets, and the reconstruction of the address block or blocks of that data packet, are delayed while the other data packet is transmitted at that output port. Consequently, although the ability of the present invention to transmit all the blocks of a single data packet on the same spatial channel is preserved, the ability to transmit several data packets on the same spatial channel is lost. Therefore, in a more preferred embodiment of the optical decoding and routing system of the present invention, each subsystem directs the data packets it receives to a switch array that also is dedicated to the corresponding communication channel. All the dedicated switch arrays feed into all the output ports. In this way, two data packets can be transmitted simultaneously at the same output port.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention is herein described, by way of example only, with reference to the accompanying drawings, wherein:
FIG. 1 shows the format of a prior art data packet;
FIG. 2 is a block diagram of a prior art optical address decoding and routing system;
FIG. 3 shows the format of a data packet of the present invention; FIG. 4 is a block diagram of a first optical address decoding and routing system of the present invention;
FIG. 5 shows the bit pattern of a data packet of the present invention:
FIG. 6 is a high level block diagram of a second optical address decoding and routing system of the present invention;
FIG. 7 is a block diagram of a decoding subsystem of the system of FIG. 6; FIG. 8 is a partial schematic diagram of a delay unit;
FIG. 9 is a high level block diagram of a third optical address decoding and routing system of the present invention
FIG. 10 shows the bit pattern of a variant of the data packet of FIG. 5;
FIG. 11 shows the bit pattern of a hybrid of the data packet of FIG. 10 and the prior art data packet of FIG. 1.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
The present invention is of a method of optical data transmission, and associated address decoding and routing systems. Specifically, the present invention can be used to transmit data packets on a common spatial channel at a higher rate than has been possible heretofore.
The principles and operation of optical data transmission according to the present invention may be better understood with reference to the drawings and the accompanying description. To prevent crosstalk among simultaneously transmitted data and address blocks, the wavelengths of the various carrier waves of the present invention must be adequately separated. For example, infrared carrier waves with wavelengths near 1550 nm should be at least about 0.4 nm apart in wavelength. Present technology supports the simultaneous use of 64 such carrier waves. It is expected that technology that supports 256 such carrier waves will be available soon.
Referring again to the drawings, Figure 4 is a block diagram of an optical address decoding and routing system 120 of the present invention, for use with data packets of the present invention that include a data block and a single address block modulated onto two carrier waves of two different wavelengths. System 120 inherits all but one of its components from prior art system 20, and these components are numbered identically in Figures 2 and 4. As in Figure 2, single arrows indicate the flow of optical data and double arrows indicate the flow of electronic data. There are only two differences between system 20 and system 120. The first difference is that
instead of active 1 x 2 switch 22, system 120 uses a passive 1 x 2 switch 122 to separate the data block of an incoming data packet from the address block and' to direct the data block to delay unit 30 and the address block to decoder 24. As noted above, switch 122 is based on elements such as add-drop grating elements, array waveguide gratings or multilayer reflective dielectric films. Because the data block and the address block are transmitted substantially simultaneously on waveguide 34, the data block and the address block are directed substantially simultaneously to delay unit 30 and to decoder 24, respectively, by switch 122. The second difference is that the delay imposed on the data block by delay unit 30 is smaller in system 120 than in system 20, because the data block and the address block overlap in time. There being only one address block per data packet, the address is encoded as in the prior art, and decoder 24 is preferably as described in co-pending U. S. Patent Application Ser. No. 09/118,372.
Figure 5 shows the bit patterns of a data block 212 and seven address blocks 214, 216, 218, 220, 222, 224 and 226 of a data packet of the present invention wherein a seven-bit address is encoded by modulating a different carrier wave with each bit of the address. Data block 212 and each address block are labeled on their right-hand sides with the bit stream that they encode. Sixteen bits are encoded in data block 212. One bit is encoded in each of address blocks 214, 216, 218, 220, 222, 224 and 226. Note that because the convention used herein for the direction of transmission is from left to right, the bit pattern of data block 212 must be read from right to left. The address encoded in the address blocks is 1101111. Arrow 228 indicates the leading edge of data block 212. Arrow 230 indicates the trailing edge of data block 212. Note that the duration of the bits encoded in the address blocks is sixteen times the duration of each bit of data block 212, to make the durations of all the blocks equal. Note also that the leading edges of the address blocks do not coincide exactly in time with leading edge 228: the transmission of data block 212 and address blocks 214, 216, 218, 220, 222, 224 and 226 is only substantially simultaneous. Using present technology, the present invention supports the substantially simultaneous transmission of up to eight such data packets.
That the address bits and the data bits of the present invention are of different duration allows even more bandwidth than is allowed merely by the overlapping in time of data block 212 and the address blocks. Generally, the minimum allowable
duration of the address bits is determined by the electronics used to interpret the address. In prior art data packet 10, wherein address bits and data bits have the same duration, this minimum allowable duration applies to the data bits too. The present invention breaks this connection between data bit duration and address bit duration. Figure 10 shows the bit patterns of a data packet of the present invention that is identical to the data packet of Figure 5, except that the duration of the bits of the seven address blocks 214', 216', 218', 220', 222', 224' and 226' is in fact selected to be as short as is allowed by the electronics used to interpret the address (three data bit durations, in this example), rather than as long as the total duration of data block 212, as in Figure 5.
Figure 6 is a high level block diagram of another optical address decoding and routing system 240 of the present invention, for use with data packets, such as the data packets of Figure 5 and Figure 10, that are transmitted in eight distinct communication channels along a single common spatial channel. As in Figures 2 and 4, the input waveguide is designated by reference numeral 34, the output ports are designated by reference numeral 36, single arrows indicate the flow of optical data, and double arrows indicate the flow of electronic data. Incoming data packets on waveguide 34 enter a passive switch 242, which separates the data packets one from another and directs each data packet to one of eight routing subsystems 244. Most preferably, passive switch 242 is an array waveguide grating (AWG), as labeled in
Figure 6; but, like passive switch 122, passive switch 242 alternatively is based on add-drop grating elements or on multilayer reflective dielectric films. Each routing subsystem 244 is associated with a unique one of the communications channels.
Switch 242 actually separates all the incoming data and address blocks from each other, so that each of the arrows pointing from switch 242 to a subsystem 244 actually represents eight waveguides, one for each carrier wave that is modulated with one of the blocks of the data packets that are associated with the corresponding communications channel. The blocks of the data packets are recombined in subsystems 244, as described below, and are transmitted to an 8 x 16 switch 246. A suitable architecture for switch 246 is described in co-pending U. S. Patent
Application Ser. No. 09/085,369, which is incorporated by reference for all purposes as if fully set forth herein. Each subsystem 244 decodes the address of a received data packet, signals switch 246 which of the sixteen output ports 36 of switch 246 the
received data packet should be transmitted from, and directs the received data packet to switch 246. Note that each arrow pointing from a subsystem 244 to switch 246 represents only one waveguide, because the separately received data and address blocks of the incoming data packets are recombined in subsystems 244 to propagate thenceforth in a common spatial channel.
Figure 7 is a block diagram of a subsystem 244. Data blocks 212 enter from the left via waveguide 261. Address blocks 214, 216, 218, 220, 222, 224 and 226 enter from the left via waveguides 248, 250, 252, 254, 256, 258 and 260, respectively, that terminate in transducers (optical detectors) 262. Each transducer 262 converts incoming infrared energy into an electronic signal that is detected by electronic circuitry in a processor 264. Because each incoming address block of an incoming data packet represents one bit of the address, processor 264 thus is provided with an electronic representation of the data packet's address. Processor 264 then signals switch 246 to direct the data packet to be received from this subsystem 244 to output port 36 that corresponds to the thus decoded address. This signaling of switch 246 by processor 264 is symbolized in Figure 7 by the double arrow above processor 264. Processor 264 then reconstructs address blocks 214, 216, 218, 220, 222, 224 and 226 by turning on corresponding light sources 266 for the duration of the address blocks. Light sources 266 are optically coupled to waveguides 268, 270, 272, 274, 276, 278 and 280. Address block 214 is recreated by light source 266 that is coupled to waveguide 268. Address block 216 is recreated by light source 266 that is coupled to waveguide 270. Address block 218 is recreated by light source 266 that is coupled to waveguide 272. Address block 220 is recreated by light source 266 that is coupled to waveguide 274. Address block 222 is recreated by light source 266 that is coupled to waveguide 276. Address block 224 is recreated by light source 266 that is coupled to waveguide 278. Address block 226 is recreated by light source 266 that is coupled to waveguide 280. So, for example, to recreate the address of the data packet of Figure 5, light sources 266 that are coupled to waveguides 268, 270, 274, 276, 278 and 280 are turned on, while light source 266 that is coupled to waveguide 272 is not turned on. For any particular incoming data packet, while processor 264 is decoding and reconstructing the data packet's address, data block 212 is delayed in a delay unit 282. The time delay is such that data block 212 emerges from delay unit 282 along a waveguide 284 while address blocks 214, 216. 218, 220. 222, 224 and 226 are being
reconstructed by light sources 266. Finally, data block 212 and address blocks 214,
216, 218, 220, 222, 224 and 226 are merged simultaneously into an output waveguide
288 by an add-drop multiplexer switch 286.
Preferably, light sources 266 are diode lasers. The light source 266 that is coupled to waveguide 268 is tuned to the wavelength of the carrier wave of address block 214. The light source 266 that is coupled to waveguide 270 is tuned to the wavelength of the carrier wave of address block 216. The light source 266 that is coupled to waveguide 272 is tuned to the wavelength of the carrier wave of address block 218. The light source 266 that is coupled to waveguide 274 is tuned to the wavelength of the carrier wave of address block 220. The light source 266 that is coupled to waveguide 276 is tuned to the wavelength of the carrier wave of address block 222. The light source 266 that is coupled to waveguide 278 is tuned to the wavelength of the carrier wave of address block 224. The light source 266 that is coupled to waveguide 280 is tuned to the wavelength of the carrier wave of address block 226.
The degree of simultaneity required among leading edge 228 of data block 212 and the leading edges of address blocks 214, 216, 218, 220, 222, 224 and 226 is determined by the response time of transducers 262. Specifically, all the leading edges must be within this response time of each other. Switch 246 is incapable of directing data packets from two different subsystems 244 to the same output port 36. Therefore, if switch 246 is signaled by two subsystems 244 to direct the data packets therefrom to the same output port 36, switch 246 signals one of those subsystems 244 to delay reconstruction of its data packet's address and transmission of its data packet. This two-way electronic signaling is symbolized in Figures 6 and 7 by the double-headed double arrows connecting subsystems 244 with switch 246. Correspondingly, delay unit 282 differs from delay unit 30 in that the delay imposed by delay unit 282 on data block 212 is adjustable. The double arrow connecting processor 264 and delay unit 282 in Figure 7 represents the electronic signals from processor 264 to delay unit 282 that instruct delay unit 282 to retard the transmission of data block 212 long enough to free up the needed output port 36.
Figure 8 is a partial schematic diagram of delay unit 282. Delay unit 282 includes a preliminary waveguide delay loop 310 and a series of cascaded delay
blocks 320. Each delay block 320 includes an input 1x2 switch 314 and an output 2x1 switch 316. The upper branch of switch 314 is connected to the upper branch of switch 316 by a short waveguide section 318. The lower branch of switch 314 is connected to the lower branch of switch 316 by a waveguide delay loop 312. Waveguide delay loop 312 provides a propagation path from switch 314 to switch 316 that is of considerably longer duration than the propagation path provided by waveguide section 318. As in Figure 7, double arrows pointing to switches 314 and 316 represent control of switches 314 and 316 by processor 264.
Data block 212 enters waveguide delay loop 310 from the left. In each delay block 320, processor 264 sets switches 314 and 316 to either the upper branches thereof or the lower branches thereof, thereby controlling the time required for data block 212 to transit each delay block 320.
Figure 9 is a high level block diagram of a third optical address decoding and routing system 300, also for use with data packets such as the data packets of Figure 5 and Figure 10. As in Figures 2, 4 and 6, the input waveguide is designated by reference numeral 34, the output ports are designated by reference numeral 36, single arrows indicate the flow of optical data, and double arrows indicate the flow of electronic data. As in the case of system 240, incoming data packets on waveguide 34 enter a passive switch 242 (preferably an AWG, as labeled), which separates the data packets one from another and directs each data packet to one of eight routing subsystems 244'. Subsystems 244' are similar to subsystems 244, but lack the ability to vary the amount by which the reconstruction of address blocks and the transmission of data blocks are retarded. To each subsystem 244' is dedicated a 1 x 4 switch 302. Each switch 302 is able to direct an incoming data packet to any one of four add-drop multiplexers 304. Each add-drop multiplexer 304 directs incoming data packets to an associated one of the four output ports 36. Every subsystem 244' receives data packets from switch 242 along eight waveguides, and transfers decoded and reconstructed data packets to the corresponding switch 302 along a single waveguide. Every switch 302 is connected to every add-drop multiplexer 304 by a single waveguide. For illustrational simplicity, not all of the optical connections between switches 302 and multiplexers 304 are shown in Figure 9. Each subsystem 244' decodes the addresses of the data packets received from switch 242 and signals the corresponding switch 302 to direct each data packet so decoded to multiplexer 304
whose output port 36 corresponds to the decoded address. Because there is no direct connection between switches 302 and output ports 36, system 300 is capable -of retransmitting two simultaneously received data packets from the same output port 36.
Figure 11 shows a data packet that is similar to the data packet of Figure 10, but each leading bit of the seven address blocks 214", 216", 218", 220", 222", 224" and 226" is followed, after a delay of four data bit durations, by two more bits. Each of these two additional address bits has the same duration as a data bit. Thus, the data packet of Figure 11 is provided with a 21 -bit address. A data packet such as the data packet of Figure 11 is decoded and routed by a decoding and routing system that is a hybrid of the system of Figure 2 and either the system of Figure 6 or the system of Figure 9.
While the invention has been described with respect to a limited number of embodiments, it will be appreciated that many variations, modifications and other applications of the invention may be made.