Optical switching system and routing method thereof, optical switching chip

Abstract

The application provides an optical switching network, a routing method thereof and an optical switching chip. The optical switching network comprises a plurality of input ports and output ports, the number of the output ports is the same as that of the input ports; a plurality of switching stages are arranged between the input ports and the output ports and extend along the direction of optical propagation; each switching stage comprises a plurality of 2*2 optical switch units, each 2*2 optical switch unit is connected with two adjacent optical waveguide channels; the 2*2 optical switch units of adjacent switching stages are arranged in a staggered manner in the channel direction, so that all the 2*2 optical switch units form a planar adjacent switching topology. By using the application, a wide-sense non-blocking optical switching network can be realized, and rearrangement of existing connections during online routing is avoided. The optical switching network also has the characteristics of recursive expansion, controllable number of switch units, good path length matching and easy realization of photonic integration.

Description

Technical Field

[0001] This invention relates to the fields of optical communication and photonic integration technology, and in particular to an optical switching system and its routing method. Background Technology

[0002] As data centers, artificial intelligence clusters, and high-performance computing systems place increasingly stringent demands on bandwidth, latency, and power consumption, N×N optical switching networks based on photonic integrated chips have become an important technological approach for realizing high-capacity optical interconnects. These networks typically consist of multiple 2×2 optical switching units, such as Mach-Zehnder interferometers, cascaded to form a multi-stage optical switching matrix.

[0003] Among various topologies, planar adjacent switching topologies have attracted much attention due to their intuitive layout, absence of optical waveguide crossovers, and good process compatibility. However, while traditional N-level planar optical switching networks can achieve deterministic port replacement, in scenarios with continuously arriving online connection requests, if it is required not to rearrange established connections, their limited path diversity will lead to situations where new legitimate connection requests cannot be accessed, i.e., they cannot meet the Widely Defined Non-Blocking (WSNB) characteristic under dynamic services. To avoid this blocking problem, existing improvement schemes often tend to introduce more complex topologies, increase the number of crossover points, or adopt non-uniform path length designs. However, this inevitably brings costs such as increased insertion loss, aggravated phase drift, deteriorated polarization-dependent loss (PDL), and a significant increase in control logic complexity. Summary of the Invention

[0004] This invention provides an optical switching system and its routing method to solve the problems of online routing congestion caused by insufficient path diversity in existing planar optical switching networks, as well as the increased losses and layout implementation costs caused by the complex structure of non-planar networks.

[0005] On one hand, embodiments of the present invention provide an optical switching system, the system including an optical switching network; the optical switching network includes: multiple input ports and output ports, and multiple switching stages arranged sequentially extending along the optical propagation direction are disposed between the input ports and the output ports; the number of output ports is the same as the number of input ports; Each of the switching stages includes several 2×2 optical switching units, and each of the 2×2 optical switching units is connected to two adjacent optical waveguide channels; The 2×2 optical switch units in adjacent switching stages are offset in the channel direction so that all the 2×2 optical switch units form a planar adjacent switching topology.

[0006] Optionally, the number of input ports and the number of output ports are N, where N is an integer not less than 4, and the number of switching stages is S, where S is an integer not less than 2N-4.

[0007] Alternatively, S equals 2N-4; The optical switching network with N=4 serves as a 4-level 4×4 planar basic example network. When N>4, each additional input and output port creates an (N+1)×(N+1) network, which adds two effective switching stages for recursive order reduction compared to the current N×N network.

[0008] Optionally, the 4-level 4×4 planar basic network uses planar basic routing to complete the replacement connection of all ports.

[0009] Optionally, the offset setting includes: The 2×2 optical switch units of the odd-numbered levels connect the 1st and 2nd adjacent channel pairs, the 3rd and 4th adjacent channel pairs, and so on; the 2×2 optical switch units of the even-numbered levels connect the 2nd and 3rd adjacent channel pairs, the 4th and 5th adjacent channel pairs, and so on; or, The 2×2 optical switch unit of the even-numbered level connects the first and second adjacent channel pairs, the third and fourth adjacent channel pairs, and so on; the 2×2 optical switch unit of the odd-numbered level connects the second and third adjacent channel pairs, the fourth and fifth adjacent channel pairs, and so on.

[0010] Optionally, the 2×2 optical switch unit can be any of the following: Mach-Zehnder interferometer optical switch, micro-ring optical switch, MEMS optical switch, liquid crystal optical switch, phase change material optical switch, or directional coupling type optical switch.

[0011] Optionally, when S is greater than 2N-4, routing is performed by any 2N-4 switching stages arranged alternately with selected odd-numbered and even-numbered stages, and the unselected switching stages are configured as at least one of the following: redundancy stage, calibration stage, protection stage, wavelength / polarization compensation stage, or backup routing stage.

[0012] Optionally, the system further includes: a control module, configured to receive a connection request, and when it is determined that there is no port conflict between the connection request and an established connection, determine a routing path for the connection request according to a predetermined routing rule, and determine, set and lock the state of the 2×2 optical switch units involved in the routing path, so that the optical switching network can realize optical path switching based on the routing path.

[0013] Optionally, the control module is implemented by an off-chip control board, an on-chip control circuit, or a combination of both.

[0014] Optionally, the control module includes: a request management unit, a sub-network labeling unit, a state storage unit, and a switch driving unit; The request management unit is used to receive connection requests and determine whether there is a port conflict between the connection request and an established connection. The subnetwork labeling unit is used to determine the routing path by performing a recursive degradation process according to the set routing rules when the request management unit confirms that there is no port conflict; and to maintain the port number and switch level number of the current subnetwork during the recursive degradation process. The state storage unit is used to record the switching state of the established 2×2 optical switch units and mark the switching state as a locked state. The switch driving unit is used to generate a driving control signal according to the routing path and output the driving signal to the corresponding 2×2 optical switch unit to perform optical path switching.

[0015] Optionally, the predetermined routing rule is a recursive two-level descending routing rule, including: For the current k×k subnetwork, a recursive order reduction process is performed, where 4≤k≤N; the recursive order reduction process includes: Based on the location of the input and output ports corresponding to the connection request, a priority edge waveguide is determined and an attempt is made to establish a boundary path. If the establishment fails, another edge waveguide is used to establish the boundary path. Once the boundary path is established, lock the state of all 2×2 optical switch units on the boundary path; Remove the input and output ports occupied by the connection from the current k×k subnetwork, and perform logical translation on the 2×2 optical switching units of the remaining optical switching network and / or set the 2×2 optical switching units not occupied by the boundary path in the two boundary switching stages occupied by the boundary path to pass-through state and lock them, so as to effectively remove the two boundary switching stages from the topology of the current subnetwork to form a (k-1)×(k-1) recursive subnetwork; Repeat the recursive order reduction process until the recursive subnetwork is reduced to a level 4 4×4 planar basic instance network; In the 4-level 4×4 planar baseline network, routing is completed using baseline routing rules based on the 4-level planar topology. The baseline routing rules refer to routing rules based on non-incremental planar topology. If it is necessary to exchange the path order of adjacent inputs, at least one of the 2×2 optical switch units in the first switching stage is set to the opposite state of the non-incremental planar topology routing rules.

[0016] On the other hand, embodiments of the present invention also provide a routing method for the optical switching system, characterized in that the method includes: Receive a connection request, the connection request including the input port number and output port number to be connected; Determine whether the connection request conflicts with an established connection; In the absence of port conflicts, the connection request is treated as the current request to be routed, and a recursive descent process is performed according to the set routing rules to determine the routing path; during the recursive descent process, the port number and switch level number of the current sub-network are maintained.

[0017] Optionally, the recursive reduction process according to the set routing rules to determine the routing path includes: For the current k×k subnetwork, a recursive order reduction process is performed, where 4≤k≤N; the recursive order reduction process includes: Based on the location of the input and output ports corresponding to the connection request, a priority edge waveguide is determined and an attempt is made to establish a boundary path. If the establishment fails, another edge waveguide is used to establish the boundary path. Once the boundary path is established, lock the state of all 2×2 optical switch units on the boundary path; Remove the input and output ports occupied by the connection from the current k×k subnetwork, and perform logical translation on the 2×2 optical switching units of the remaining optical switching network and / or set the 2×2 optical switching units not occupied by the boundary path in the two boundary switching stages occupied by the boundary path to pass-through state and lock them, so as to effectively remove the two boundary switching stages from the topology of the current subnetwork to form a (k-1)×(k-1) recursive subnetwork; Repeat the recursive order reduction process until the recursive subnetwork is reduced to a level 4 4×4 planar basic instance network; In the 4-level 4×4 planar baseline network, routing is completed using baseline routing rules based on the 4-level planar topology. The baseline routing rules refer to routing rules based on non-incremental planar topology. If it is necessary to exchange the path order of adjacent inputs, at least one of the 2×2 optical switch units in the first switching stage is set to the opposite state of the non-incremental planar topology routing rules.

[0018] Optionally, the method further includes: when there are multiple connection requests, selecting each connection request in a predetermined order to complete the routing.

[0019] On the other hand, embodiments of the present invention also provide an optical switching chip, including a substrate, an optical waveguide layer formed on the substrate, a plurality of 2×2 optical switching units, driving devices corresponding to each optical switching unit, and a control interface; the plurality of 2×2 optical switching units form an optical switching network in the optical switching system as described in any one of claims 1 to 11; The control interface is used to receive routing results from the optical switching system, generate drive signals based on the routing results, and output them to the corresponding drive devices. The driving device is used to control the state of the optical switch unit according to the driving signal, so as to form a route to realize signal transmission.

[0020] Optionally, the substrate is any one of the following: silicon, silicon nitride, lithium niobate, or indium phosphide.

[0021] On the other hand, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to perform the steps of the routing method.

[0022] Compared with the prior art, the technical solution of the embodiments of the present invention has the following beneficial effects: (1) While maintaining the plane adjacent switching topology, wide-sense non-blocking is achieved through 2N-4 effective switching stages, avoiding the rearrangement of existing connections during online routing.

[0023] (2) The preferred implementation has a number of 2×2 optical switch units of (2N-4)(N-1) / 2, which is on the order of N^2, making it easy to expand with the number of ports.

[0024] (3) The switching unit only connects to adjacent channels, which helps to reduce layout complexity and parasitic loss.

[0025] (4) The recursive two-level reduction algorithm is deterministic, which makes it easy for the electronic controller to quickly calculate and convert it into the MZI driving state.

[0026] (5) This planar topology is advantageous for providing a better matching path length from adjacent inputs to adjacent outputs, and is suitable for polarization diversity optical switching. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the topology of an optical switching network in an optical switching system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the straight-through and cross-through states of the 2×2 optical switch unit in an embodiment of the present invention; Figure 3 This is a schematic diagram of a structure of an optical switching system provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the control module in one embodiment of the present invention; Figure 5 This is a schematic diagram of a recursive two-level descending route in an embodiment of the present invention; Figure 6 This is a schematic diagram of the complete routing process in a 5×5 enhanced planar optical switching subnetwork in an embodiment of the present invention; Figure 7 This is a flowchart of a routing method for an optical switching network provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of a network structure for achieving polarization diversity using an optical switching network in an embodiment of the present invention; Figure 9 This is a schematic diagram of another network structure for achieving polarization diversity using the optical switching network in this embodiment of the invention. Detailed Implementation

[0028] To make the above-mentioned objectives, features and beneficial effects of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0029] This invention provides an optical switching system and its routing method, which, while maintaining the characteristics of plane adjacent switching, achieves a broadly defined non-blocking optical switching network by appropriately increasing the number of switching stages and cooperating with deterministic routing rules.

[0030] like Figure 1 The diagram shown is a topology diagram of an optical switching network in an optical switching system provided in an embodiment of the present invention.

[0031] The optical switching network 100 includes multiple input ports 101 and multiple output ports 102. The number of input ports and the number of output ports are the same. For ease of description, the number of input ports and the number of output ports are set as N, where N is an integer not less than 4.

[0032] Multiple switching stages 104 are arranged sequentially along the direction of light propagation between the input port 101 and the output port 102. N optical waveguide channels 103 extend along the direction of light propagation.

[0033] Each switching stage 104 includes several 2×2 optical switching units 140, and each 2×2 optical switching unit 140 is connected to two adjacent optical waveguide channels 103. The number S of switching stages is an integer not less than 2N-4.

[0034] In some embodiments, the number of switching stages S = 2N - 4. For example, when N = 4, S = 4; when N = 5, S = 6; when N = 6, S = 8; and when N = 8, S = 12. This can be understood as adding N-4 switching stages to a traditional N-level planar topology (with N switching stages), ensuring that during the recursive downscaling from an N×N scale to a 4×4 baseline, each downscaling operation (i.e., stripping one input port, one output port, and two boundary switching stages) provides a corresponding effective switching stage for stripping. In other words, the optical switching network with N = 4 serves as a 4-level 4×4 planar baseline network; when N > 4, each increase in input and output port size, i.e., expanding from the current N×N level network to an (N+1)×(N+1) level network, adds two effective switching stages to the current N×N level network, thus providing a pair of strippable boundary switching stages for each round of downscaling in the recursive routing. The 4-level 4×4 planar basic network uses planar basic routing rules to complete the replacement connection of all ports. These planar basic routing rules are based on traditional non-incremental planar topology routing rules, but do not mandate that inputs must be edge inputs. In this planar basic routing, if an input is not an edge input, the path order of adjacent inputs needs to be swapped. The corresponding 2×2 optical switch unit in the first switching stage is set to a cross-connect or pass-through state, the opposite of the non-incremental planar topology routing rules. The remaining switch units are arranged according to the target output and set to either pass-through or cross-connect states accordingly.

[0035] In this embodiment of the invention, the 2×2 optical switching units 140 of adjacent switching stages are offset in the channel direction so that all 2×2 optical switching units 140 form a planar adjacent switching topology. The offset setting includes: The 2×2 optical switch units of the odd-numbered levels are connected to the first and second adjacent channel pairs, the third and fourth adjacent channel pairs, and so on; the 2×2 optical switch units of the even-numbered levels are connected to the second and third adjacent channel pairs, the fourth and fifth adjacent channel pairs, and so on.

[0036] If N is even, the number of 2×2 optical switch units 140 in the odd-numbered and even-numbered stages are N / 2 and (N-2) / 2, respectively; If N is odd, then the number of 2×2 optical switch units 140 in each stage is (N-1) / 2.

[0037] For the case of S=2N-4, the total number of all 2×2 optical switch units 140 can be expressed as (2N-4)(N-1) / 2.

[0038] In some embodiments, the number of switching stages S is greater than 2N-4. In this case, routing can be performed by any 2N-4 switching stages arranged alternately with selected odd and even stages. Unselected switching stages can be configured as at least one of redundant stages, calibration stages, protection stages, wavelength / polarization compensation stages, or backup routing stages.

[0039] It should be noted that the above arrangement is only a numbering convention. In actual chip layouts, the odd / even order can be reversed, the input / output direction can be reversed, or a different arrangement can be used. Figure 1 An equivalent structure with left-right mirroring, as long as each switching stage maintains the planar characteristics of adjacent channel switching and progressive offset, falls within the scope of this invention. For example, the 2×2 optical switching unit of the even-numbered stage connects the 1st and 2nd adjacent channel pairs, the 3rd and 4th adjacent channel pairs, and so on; the 2×2 optical switching unit of the odd-numbered stage connects the 2nd and 3rd adjacent channel pairs, the 4th and 5th adjacent channel pairs, and so on.

[0040] In this embodiment of the invention, the 2×2 optical switch unit 140 can be a Mach-Zehnder interferometer optical switch, a micro-ring optical switch, a MEMS optical switch, a liquid crystal optical switch, a phase change material optical switch, a directional coupling type optical switch, or other optical switch structures capable of switching between through-state and cross-state between two inputs and two outputs. This embodiment of the invention does not limit the specific type of optical switch. Taking a Mach-Zehnder optical switch as an example, each Mach-Zehnder interferometer optical switch includes two couplers and at least one phase modulation arm, switching between through-state and cross-state through thermo-optic modulation, electro-optic modulation, carrier injection, or carrier depletion.

[0041] like Figure 2 The diagram shown is a schematic diagram of the straight-through state and cross-through state of the 2×2 optical switch unit in an embodiment of the present invention.

[0042] Figure 2 In the middle (a) state, it means that the two input terminals are connected to the output terminals on the same side respectively; Figure 2 In the middle (b), the crossover state is indicated, which means that the two input terminals are swapped to the output terminals on the opposite side.

[0043] The optical switching network 100 in this embodiment of the invention also has the characteristics of recursive expansion, controllable number of switching units, good path length matching, and easy photonic integration, which will be described in detail below.

[0044] like Figure 3As shown, in some embodiments, the optical switching system 30 provided by the present invention not only includes the optical switching network 100 described above, but may further include a control module 200, which is used to receive connection requests, determine a routing path for the connection request according to a predetermined routing rule when it is determined that there is no port conflict between the connection request and the established connection, and determine, set and lock the state of the 2×2 optical switch units involved in the routing path so that the optical switching network 100 can realize optical path switching based on the routing path.

[0045] The control module 200 may be implemented by an off-chip control board, an on-chip control circuit, or a combination of both, and the embodiments of the present invention do not limit this.

[0046] like Figure 4 As shown, a non-limiting embodiment of the control module 200 may include: a request management unit 201, a sub-network labeling unit 202, a state storage unit 203, and a switch driving unit 204. Wherein: The request management unit 201 is used to receive connection requests and determine whether there is a port conflict between the connection request and an established connection. The subnetwork labeling unit 202 is used to determine the routing path by performing a recursive degradation process according to the set routing rules when the request management unit confirms that there is no port conflict; and to maintain the port number and switch level number of the current subnetwork during the recursive degradation process. The state storage unit 203 is used to record the switching state of the established 2×2 optical switch units and mark the switching state as a locked state. The switch driving unit 204 is used to generate a driving control signal (such as generating driving voltage, current or digital control code for MZI, micro ring, MEMS or other optical switches) according to the routing path, and output the driving signal to the corresponding 2×2 optical switch unit to perform optical path switching.

[0047] The routing rules set are recursive two-level descending routing rules, including: For the current k×k subnetwork, 4≤k≤N, determine the priority edge waveguide based on the position of the input and output ports corresponding to the connection request and attempt to establish the boundary path. If the establishment fails, use another edge waveguide to establish the boundary path. Once the boundary path is established, lock the state of all 2×2 optical switch units on the boundary path; Remove the input and output ports occupied by the connection from the current k×k subnetwork, and perform logical translation on the 2×2 optical switching units of the remaining optical switching network and / or set the 2×2 optical switching units not occupied by the boundary path in the two boundary switching stages occupied by the boundary path to pass-through state and lock them, so as to effectively remove the two boundary switching stages from the topology of the current subnetwork to form a (k-1)×(k-1) recursive subnetwork; Repeat the above operation until the recursive subnetwork is reduced to a level 4 4×4 planar basic network; In the 4-level 4×4 planar basic network, the basic routing rules of the 4-level planar topology (i.e., the routing rules based on the traditional non-incremental planar topology, but without restricting the input end to be an edge input end) are used to complete the routing. If the path order of adjacent inputs needs to be swapped because the input end is not an edge input end, then at least one of the 2×2 optical switch units in the first switching stage is set to the opposite state of the non-incremental planar topology routing rules.

[0048] In this embodiment of the invention, the edge waveguide refers to the two outermost optical waveguide channels arranged along the channel direction in the current sub-network. The edge waveguide is relative to the current sub-network. As the sub-network shrinks in size through recursive order reduction (deleting ports and switching stages), the position of the edge waveguide (referring to its logical number position) will dynamically change. The boundary path refers to an end-to-end optical path in which, during routing, the optical signal is guided to the edge waveguide using a 2×2 optical switch unit starting from the input port, propagates along the edge waveguide for as long as possible, and only leaves the edge waveguide before reaching the target output port, where the optical signal is guided to the target output port using a 2×2 optical switch.

[0049] It should be noted that when multiple connection requests are pending in the current subnetwork, the control module 200 can select one of the connection requests for priority routing according to a predetermined order. When there is only one connection request pending, it is directly routed. The predetermined order may be a combination of input port number or a function thereof, output port number or a function thereof, connection request arrival order and port parity priority, or other deterministic order fixed during system initialization. This embodiment of the invention does not limit this.

[0050] The control module 200 determines the boundary path for the selected connection request. The boundary path is formed using only the 2×2 optical switch units of the current subnetwork, so that the connection can be established without changing the locked path.

[0051] In this embodiment of the invention, the control module 200 determines the routing path through a recursive order reduction process, the specific process of which is as follows: According to the upgraded planar topology provided in this embodiment of the invention, the current subnetwork has 2k-4 effective switching stages.

[0052] For the current k×k subnetwork, 4≤k≤N, first determine the priority edge waveguide based on the position of the input port and output port corresponding to the connection request and attempt to establish the boundary path. If the establishment fails, use another edge waveguide to establish the boundary path.

[0053] If both the input and output ports of the selected connection request are located in the upper half of the current subnetwork, the boundary path preferentially uses the upper edge waveguide; if both the input and output ports of the selected connection request are located in the lower half of the current subnetwork, the boundary path preferentially uses the lower edge waveguide; if one end (e.g., the input port) is located in the upper half of the current subnetwork and the other end (e.g., the output port) is located in the lower half of the current subnetwork, the boundary path can preferentially use either the upper or lower edge waveguide. The positions of the input and output ports can be determined based on the port numbers.

[0054] The boundary path requires the use of at least one edge waveguide. If the preferred path fails, the edge waveguide on the other side is used. To ensure broad-sense non-blocking operation, the boundary path needs to propagate along the uppermost or lowermost edge waveguide of the current sub-network as much as possible. Specifically, this is achieved by using cross-state optical switches as much as possible to guide light to the edge waveguides starting from the input port, and then using cross-state optical switches to guide light to the output waveguide at the latest. After completing this connection, the control module 200 sets the 2×2 optical switch units involved in the boundary path to through-state or cross-state and locks them.

[0055] The control module 200 removes one input port and one output port occupied by the connection from the current k×k subnetwork, and performs logical translation on the 2×2 optical switching units of the remaining optical switching network and / or sets the 2×2 optical switching units not occupied by the boundary path in the two boundary switching stages to a pass-through state and locks them, thereby effectively removing the two boundary switching stages from the topology of the current subnetwork. The remaining parts are renumbered in relative order to obtain a recursive subnetwork of size (k-1)×(k-1). Since 2k-4-2=2(k-1)-4, the recursive subnetwork still satisfies the incremental planar topology condition of the present invention.

[0056] Therefore, the above process can continue until k=4.

[0057] The above process ultimately reduces the recursive subnetwork to a level 4 4×4 planar basic network through logical translation. The logical translation refers to the equivalent removal of the locked optical switches without changing the actual positions of the physical input ports, physical output ports, physical waveguides, and optical switches. That is, treating them as fixed connections and deleting the switch units from the equivalent subnetwork, and then logically renumbering the ports, waveguide channels, and optical switches in the remaining subnetwork.

[0058] Reference Figure 5 , Figure 5 This is a schematic diagram of a recursive two-level descending routing in an embodiment of the present invention.

[0059] like Figure 5 As shown in (a), the input port corresponding to the current connection request is I3, and the output port is O3. In the current sub-network 100, the boundary path 106 is determined for this connection request, and the state of all 2×2 optical switch units on the boundary path 106 is locked. The input port I3 and output port O3 occupied by the connection are deleted from the current sub-network 100. After the waveguide occupied by the boundary path 106 is equivalently removed, some 2×2 optical switch units in the boundary switching stage need to be set to the through state and locked through logical translation, thereby equivalently removing two boundary switching stages, resulting in the recursive sub-network as shown in (a). Figure 5 The recursive subnetwork 100-1 is shown in (b). Therefore, it can be observed that output port O1 is directly connected to the upper port of optical switch unit 15 via optical switch 6 locked in a cross state; output port O2 is directly connected to the upper port of optical switch unit 7 via optical switch 9 locked in a cross state; the upper port of optical switch unit 1 is directly connected to the upper port of optical switch unit 14 via optical switch 13 locked in a cross state; and the lower port of optical switch unit 1 is directly connected to the upper port of optical switch unit 12 via optical switch unit 4 locked in a cross state. Therefore, from a topological perspective, [the network structure can be described as follows]. Figure 5 The logic translation of the optical switch unit 1 in (a) is as follows: Figure 5 The position of the optical switch unit 4 in the recursive subnetwork 100-1 shown in (b) is determined, and the output ports O1 and O2 are logically shifted to... Figure 5 The positions of the two output ports in the recursive subnetwork 100-1 shown in (b) are shown, but this does not indicate any physical movement of any optical switch unit or input / output ports. After logical translation, the positions are visible... Figure 5 Comparison of optical switch unit 3 and optical switch unit 10 in (a) Figure 5 The recursive subnetwork 100-1 shown in (b) is a redundant optical switch unit, which can be effectively removed after being set to the through state, thereby eliminating the need for two boundary switching stages ( Figure 5In (a) the boundary switching levels 51 and 52) are removed from the current subnetwork 100 topology, and the 6×6 subnetwork is recursively downgraded to a 5×5 subnetwork. Figure 5 In the recursive subnetwork 100-1 shown in (b), the physical input ports I2 and I3 and each optical switch unit have not changed their positions.

[0060] When k≤4, the control module 200 uses a baseline routing rule based on a 4-level planar topology to process the remaining connections. The baseline routing rule refers to the routing rule based on a non-incremental planar topology, which rearranges the connection requests, starting from an edge input request, using edge waveguide routing as much as possible, and then using an optical switch unit set to cross-state to route the input to the output.

[0061] Compared to the routing rules of traditional non-incremental planar topologies, because broad-sense non-blocking routing requires routing without affecting existing paths when a connection request arrives, it is impossible to rearrange connection requests and start from edge inputs. Therefore, in this embodiment of the invention, the baseline routing rules of the 4-level planar topology may need to swap the path order of adjacent inputs. If necessary, at least one 2×2 optical switch unit in the first switching stage is set to the opposite state of the non-incremental planar topology routing rules; the remaining 2×2 optical switch units are arranged in either a through-state or a cross-state according to the target output.

[0062] like Figure 6 The diagram illustrates the complete routing process in a 5×5 extended planar optical switching subnetwork according to an embodiment of the present invention.

[0063] In this example, the subnetwork performs a two-level descending route from 5×5 to 4×4, a process similar to... Figure 5The description and rules are consistent, so they will not be repeated. The subnetworks recursively progress from 4×4 to 3×3, and then from 3×3 to 2×2; a basic routing rule based on traditional non-incrementing plane topology routing rules is adopted. In the traditional non-incrementing plane topology routing rules, the input must be at the edge of the subnetwork, i.e., I1 or I5; if the connection request is not at the edge of the subnetwork, when a new connection request arrives, all connection requests, including those that have already been routed, can be rearranged and routed to ensure that each connection request is at the edge of the subnetwork when it is routed. However, to ensure a broad non-blocking condition, the optical switching network cannot rearrange and route previous routing requests, so it is necessary to ensure that routes for connection requests whose input is not at the edge of the subnetwork can be satisfied. As shown in the 4×4 subnetwork, the basic routing rule of the traditional non-incrementing plane topology routing rules can satisfy the routing of I5-O3, and its path is represented by the blue line. When the input port of the connection request is I4, a 2×2 optical switch unit in the first switching stage can be configured with the opposite routing rule of the traditional non-incrementing planar topology, and the path indicated by the red line completes the I4-O3 route. In the 4×4 to 3×3 example, when the input port of the connection request is I2, the blue path with I1 as the input is used as a reference, and a 2×2 optical switch unit is configured with the opposite routing rule of the traditional non-incrementing planar topology to achieve the route. Because k≤4, there are at most two non-edge input ports, and both can be routed using this routing rule. Therefore, for the case of k≤4, it is only necessary to configure one 2×2 optical switch unit in the first switching stage with the opposite routing rule of the traditional non-incrementing planar topology to achieve the same first-level recursive descending routing as the traditional non-incrementing planar topology.

[0064] Since the 4×4 basic example can be completed in this planar topology, combined with the above recursive order reduction process, a broad-sense non-blocking route for any N≥4 can be obtained.

[0065] The term "broad non-blocking" means that as long as all connection requests are established according to the same predetermined routing rules, for any subsequent new request that does not conflict with existing connections on the input and output ports, the control module 200 can establish a connection for the new request in the remaining unlocked resources without changing the switch state on the existing connection path.

[0066] In relation to the optical switching network provided in the embodiments of the present invention, the embodiments of the present invention also provide a routing method for the optical switching network, such as... Figure 7 The diagram shown is a flowchart of a routing method for an optical switching network provided in an embodiment of the present invention.

[0067] In step 701, a connection request is received, which includes the input port number and output port number to be connected.

[0068] In step 702, it is determined whether there is a port conflict between the connection request and an established connection.

[0069] Specifically, it determines whether the input and output ports to be connected in the connection request are already occupied by other connections. If they are occupied, a port conflict is determined; otherwise, no port conflict exists.

[0070] In step 703, if there is no port conflict, the connection request is taken as the current request to be routed, and a recursive descent process is performed according to the set routing rules to determine the routing path; during the recursive descent process, the port number and switch level number of the current sub-network are maintained.

[0071] Specifically, the current k×k subnetwork is determined, and the unlocked 2×2 optical switch units in the subnetwork are obtained.

[0072] In the current k×k subnetwork, a boundary path is determined. The method for determining the boundary path can be found in the description in the previous embodiments, and will not be repeated here.

[0073] After the boundary path is established, the state of all 2×2 optical switch units on the boundary path is locked. The state can be a through state or a cross state. For example, these switch states are written to the state storage module.

[0074] Keeping the locked switch state unchanged, delete the input and output ports occupied by the connection from the current k×k subnetwork, and perform logical translation on the remaining 2×2 optical switch units of the optical switching network and / or set the 2×2 optical switch units not occupied by the boundary path in the two boundary switching stages occupied by the boundary path to the pass-through state and lock them, so as to remove them equivalently from the topology of the current subnetwork.

[0075] Then, the remaining network is logically renumbered to form a (k-1)×(k-1) recursive subnetwork. The units that need to be renumbered include: input ports, output ports, and switching stages.

[0076] Repeat the above process until the recursive subnetwork is reduced to a level 4 4×4 planar basic network.

[0077] Then, in the 4-level 4×4 plane baseline network, the baseline routing rules of the 4-level plane topology are used to complete the routing.

[0078] It should be noted that if it is necessary to swap the path order of adjacent inputs, at least one of the 2×2 optical switch units in the first switching stage shall be set to the opposite state of the non-incremental plane topology routing rule.

[0079] Accordingly, this embodiment of the invention also provides an optical switching chip, including a substrate, an optical waveguide layer formed on the substrate, a plurality of 2×2 optical switching units, driving devices corresponding to each optical switching unit, and a control interface; the plurality of 2×2 optical switching units form an optical switching network in the optical switching system of this embodiment of the invention.

[0080] In some embodiments, the substrate is any one of the following: silicon, silicon nitride, lithium niobate, indium phosphide, or a hybrid integrated platform comprising multiple materials.

[0081] The control interface is used to receive routing results from the optical switching system, generate drive signals based on the routing results, and output them to the corresponding drive devices. The driving device is used to control the state of the optical switch unit according to the driving signal, so as to form a route to realize signal transmission.

[0082] It should be noted that for channel positions in a certain switching stage that do not have 2×2 optical switch units, straight waveguide sections, equivalent delay sections, or loss compensation sections can be set to make it easier to match different paths in terms of optical path length and loss.

[0083] When routing from adjacent input ports to adjacent output ports, the planar adjacent topology in this embodiment of the invention can usually allow the corresponding optical paths to pass through similar waveguide lengths, thus making it applicable to polarization diversity structures.

[0084] for example Figure 8 As shown, in one polarization diversity embodiment, the input light is separated into TE polarization components and TM polarization components by a polarization beamsplitter 81. The TE polarization component enters the first N×N extended planar optical switching subnetwork 801, and the TM polarization component enters the second N×N extended planar optical switching subnetwork 802. The two subnetworks have the same or mirror topology and are synchronously controlled by the same control module according to the corresponding routing results. Finally, they are combined into output light by a polarization beam combiner 82.

[0085] For example Figure 9 As shown, in another polarization diversity embodiment, a physical 2N-channel augmented planar optical switching network can be used to implement polarization diversity switching of N logical input ports and N logical output ports. The augmented planar optical switching network has at least 4N-4 switching stages. Each logical input port corresponds to two adjacent physical channels, such as... Figure 9 The logic input port I91 corresponds to two adjacent physical channels 901 and 902, which respectively carry the TE polarization component and TM polarization component of the same input light; each logic output port also corresponds to two adjacent physical channels, such as... Figure 9The logical output port O91 in the optical switching system corresponds to the adjacent physical channels 901 and 902. The control module in the optical switching system performs associated routing on paired channels, so that two polarization components of the same logical connection pass through the same or similar number of switching stages and waveguide lengths.

[0086] The optical switching system and routing method provided in this invention are not limited to single-wavelength optical switching applications. For wavelength division multiplexing (WDM) scenarios, multiple wavelengths can be carried simultaneously on each waveguide channel. The control module can uniformly configure the spatial switching path according to the port connection relationship, and can be expanded by combining a wavelength selector or power equalizer if necessary.

[0087] It should be noted that, unless otherwise stated, the optical switch unit mentioned in the preceding embodiments refers to a 2×2 optical switch unit.

[0088] Accordingly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program is executed by a processor. Figure 7 The steps of the routing method are described.

[0089] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article indicates that the preceding and following related objects have an "or" relationship.

[0090] In this application's embodiments, "multiple" refers to two or more. The descriptions of "first," "second," etc., appearing in this application's embodiments are merely illustrative and for distinguishing the described objects; they have no order and do not indicate a specific limitation on the number of devices in this application's embodiments, nor do they constitute any limitation on the embodiments of this application.

[0091] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. An optical switching system, characterized in that, The system includes an optical switching network; the optical switching network includes: multiple input ports and output ports, and multiple switching stages arranged sequentially along the optical propagation direction are provided between the input ports and the output ports; the number of output ports is the same as the number of input ports; Each of the switching stages includes several 2×2 optical switching units, and each of the 2×2 optical switching units is connected to two adjacent optical waveguide channels; The 2×2 optical switch units in adjacent switching stages are offset in the channel direction so that all the 2×2 optical switch units form a planar adjacent switching topology.

2. The optical switching system according to claim 1, characterized in that, The number of input ports and the number of output ports are N, where N is an integer not less than 4, and the number of switching stages is S, where S is an integer not less than 2N-4.

3. The optical switching system according to claim 2, characterized in that, S equals 2N-4; The optical switching network with N=4 serves as a 4-level 4×4 planar basic example network. When N>4, each additional input and output port creates an (N+1)×(N+1) network, which adds two effective switching stages for recursive order reduction compared to the current N×N network.

4. The optical switching system according to claim 3, characterized in that, The 4-level 4×4 planar basic network uses planar basic routing to complete the replacement connection of all ports.

5. The optical switching system according to claim 1, characterized in that, The offset setting includes: The 2×2 optical switch units of the odd-numbered levels connect the 1st and 2nd adjacent channel pairs, the 3rd and 4th adjacent channel pairs, and so on; the 2×2 optical switch units of the even-numbered levels connect the 2nd and 3rd adjacent channel pairs, the 4th and 5th adjacent channel pairs, and so on; or, The 2×2 optical switch unit of the even-numbered level connects the first and second adjacent channel pairs, the third and fourth adjacent channel pairs, and so on; the 2×2 optical switch unit of the odd-numbered level connects the second and third adjacent channel pairs, the fourth and fifth adjacent channel pairs, and so on.

6. The optical switching system according to claim 1, characterized in that, The 2×2 optical switch unit can be any one of the following: Mach-Zehnder interferometer optical switch, micro-ring optical switch, MEMS optical switch, liquid crystal optical switch, phase change material optical switch, or directional coupling type optical switch.

7. The optical switching system according to claim 2, characterized in that, When S is greater than 2N-4, routing is performed by any 2N-4 switching stages arranged alternately with odd and even selected stages. Unselected switching stages are configured as at least one of the following: redundancy stage, calibration stage, protection stage, wavelength / polarization compensation stage, or backup routing stage.

8. The optical switching system according to any one of claims 1 to 7, characterized in that, The system also includes: The control module is used to receive connection requests, and when it is determined that there is no port conflict between the connection request and the established connection, it determines the routing path for the connection request according to the predetermined routing rules, and determines, sets and locks the state of the 2×2 optical switch units involved in the routing path, so that the optical switching network can realize optical path switching based on the routing path.

9. The optical switching system according to claim 8, characterized in that, The control module is implemented by an off-chip control board, an on-chip control circuit, or a combination of both.

10. The optical switching system according to claim 8, characterized in that, The control module includes: a request management unit, a sub-network labeling unit, a state storage unit, and a switch driving unit; The request management unit is used to receive connection requests and determine whether there is a port conflict between the connection request and an established connection. The subnetwork labeling unit is used to determine the routing path by performing a recursive degradation process according to the set routing rules when the request management unit confirms that there is no port conflict; and to maintain the port number and switch level number of the current subnetwork during the recursive degradation process. The state storage unit is used to record the switching state of the established 2×2 optical switch units and mark the switching state as a locked state. The switch driving unit is used to generate a driving control signal according to the routing path and output the driving signal to the corresponding 2×2 optical switch unit to perform optical path switching.

11. The optical switching system according to claim 10, characterized in that, The predetermined routing rules are recursive two-level descending routing rules, including: For the current k×k subnetwork, a recursive order reduction process is performed, where 4≤k≤N; the recursive order reduction process includes: Based on the location of the input and output ports corresponding to the connection request, a priority edge waveguide is determined and an attempt is made to establish a boundary path. If the establishment fails, another edge waveguide is used to establish the boundary path. Once the boundary path is established, lock the state of all 2×2 optical switch units on the boundary path; Remove the input and output ports occupied by the connection from the current k×k subnetwork, and perform logical translation on the 2×2 optical switching units of the remaining optical switching network and / or set the 2×2 optical switching units not occupied by the boundary path in the two boundary switching stages occupied by the boundary path to pass-through state and lock them, so as to effectively remove the two boundary switching stages from the topology of the current subnetwork to form a (k-1)×(k-1) recursive subnetwork; Repeat the recursive order reduction process until the recursive subnetwork is reduced to a level 4 4×4 planar basic instance network; In the 4-level 4×4 planar baseline network, routing is completed using baseline routing rules based on the 4-level planar topology. The baseline routing rules refer to routing rules based on non-incremental planar topology. If it is necessary to exchange the path order of adjacent inputs, at least one of the 2×2 optical switch units in the first switching stage is set to the opposite state of the non-incremental planar topology routing rules.

12. A routing method for an optical switching system as described in any one of claims 1 to 11, characterized in that, The method includes: Receive a connection request, the connection request including the input port number and output port number to be connected; Determine whether the connection request conflicts with an established connection; In the absence of port conflicts, the connection request is treated as the current request to be routed, and a recursive descent process is performed according to the set routing rules to determine the routing path; during the recursive descent process, the port number and switch level number of the current sub-network are maintained.

13. The routing method according to claim 12, characterized in that, The recursive descent process according to the set routing rules, which determines the routing path, includes: For the current k×k subnetwork, a recursive order reduction process is performed, where 4≤k≤N; the recursive order reduction process includes: Based on the location of the input and output ports corresponding to the connection request, a priority edge waveguide is determined and an attempt is made to establish a boundary path. If the establishment fails, another edge waveguide is used to establish the boundary path. Once the boundary path is established, lock the state of all 2×2 optical switch units on the boundary path; Remove the input and output ports occupied by the connection from the current k×k subnetwork, and perform logical translation on the 2×2 optical switching units of the remaining optical switching network and / or set the 2×2 optical switching units not occupied by the boundary path in the two boundary switching stages occupied by the boundary path to pass-through state and lock them, so as to effectively remove the two boundary switching stages from the topology of the current subnetwork to form a (k-1)×(k-1) recursive subnetwork; Repeat the recursive order reduction process until the recursive subnetwork is reduced to a level 4 4×4 planar basic instance network; In the 4-level 4×4 planar baseline network, routing is completed using baseline routing rules based on the 4-level planar topology. The baseline routing rules refer to routing rules based on non-incremental planar topology. If it is necessary to exchange the path order of adjacent inputs, at least one of the 2×2 optical switch units in the first switching stage is set to the opposite state of the non-incremental planar topology routing rules.

14. The routing method according to claim 12 or 13, characterized in that, The method further includes: When multiple connection requests exist, the connection requests are selected sequentially according to a predetermined order to complete the routing.

15. An optical switching chip, characterized in that, The system includes a substrate, an optical waveguide layer formed on the substrate, a plurality of 2×2 optical switch units, a driving device corresponding to each optical switch unit, and a control interface; the plurality of 2×2 optical switch units form an optical switching network in the optical switching system as described in any one of claims 1 to 11. The control interface is used to receive routing results from the optical switching system, generate drive signals based on the routing results, and output them to the corresponding drive devices. The driving device is used to control the state of the optical switch unit according to the driving signal, so as to form a route to realize signal transmission.

16. The optical switching chip according to claim 15, characterized in that, The substrate is any one of the following: silicon, silicon nitride, lithium niobate, or indium phosphide.

17. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when executed by a processor, performs the steps of the routing method according to any one of claims 12 to 14.

Images