JPH0545930B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] This invention has a plurality of input ports and output ports, selects optical input signals supplied to the input ports in a matrix, and outputs optical signals from the output ports. This relates to an optical matrix switch that outputs signals.

[Conventional technology]

Recently, with the introduction of optical fibers into transmission systems, research has been actively conducted on the use of waveguide type optical switches in switching equipment as well.

FIG. 2 is a perspective view showing the structure of an optical matrix switch using such a waveguide type optical switch. On a substrate 9 made of LiNbO ₃ having an electro-optic effect, sixteen 2×2 optical switches 91 are arranged in a square lattice of 4 rows and 4 columns. 4 input ports 1
a, 1b, 1c, 1d have 2×2 lines that make up the row.
One input and output of each optical switch 91 are connected in series through a waveguide 93, and the output of the 2×2 optical switch 91 in the last row is connected to four output ports 2a, 2b, 2c, and 2d. It is connected to the. In addition, the output of each 2×2 optical switch 91 that does not form a row is connected to the other input that does not form a row of the adjacent 2×2 optical switch 91 in the subsequent column via a waveguide 93. connected, each 2
The ×2 optical switches 91 are connected to each other diagonally. According to the above conventional technology,
Information input from any of the input ports 1a, 1b, 1c, and 1d is stored in a 2x2 matrix placed at the intersection of the matrix.
An optical switch 91 switches the direction of the outgoing lines into a so-called direct state or a crossing state. and,
Compared to the most basic optical matrix switch, which is constructed by arranging optical switches in rows and columns in a square lattice and orthogonally connecting each optical switch, it is difficult to understand what kind of input is required even though it is composed of the same number of optical switches. The optical switch means that passes through the port and output port combinations is always the same, and the number of optical switches is 1/2 compared to the worst case with a basic optical matrix switch, so light loss and cross-linking are reduced. It has the feature of reducing talk.

However, even with the above conventional technology,
Information input from the input port 1b as shown by the solid line arrow in FIG. It had the disadvantage of causing

On the other hand, a perspective view of FIG. 3 shows the structure of an optical matrix switch that has been improved so as not to cause crosstalk. According to the improved prior art,
On the board 9 are input ports 1a, 1b, 1c,
1d, 16 2×2 optical switches 94 arranged in a square grid of 4 rows and 4 columns, and output port 2.
Similar 2x2 connected to a, 2b, 2c, 2d
An optical switch 95 is arranged. The 2×2 optical switches 94 and 95 located at each grid intersection on the input port side and the output port side are arranged so that each row on the output port side is positioned one row at a time for each column on the input port side. and are mutually connected by a waveguide 93.

According to this, since there is no case where different information is input to one 2x2 optical switch, the information input to the input port is matrix-selected and sent to the output port without causing almost any crosstalk. It can be output.

[Problem that the invention seeks to solve]

However, with the above-mentioned conventional technology, although crosstalk can be significantly reduced compared to an optical matrix switch consisting of 2x2 optical switches arranged in a simple square grid, the input port side Optical switches arranged in a square lattice are required depending on the number of ports, and the same number of optical switches arranged in a square lattice are also required on the input port side. For this reason, twice as many optical switches are required as compared to an optical matrix switch that uses a simple square grid arrangement, so the overall size of the optical matrix switch becomes larger. The number of light switch means to do so also increases. Another problem was that light loss could not be reduced.

Therefore, the present invention can be configured with the same number of optical switches as in an optical matrix switch having a simple square lattice arrangement, and can be formed compactly with fewer stages of optical switches from an input port to an output port. Another object of the present invention is to provide an optical matrix switch with excellent crosstalk characteristics.

[Means for solving problems]

In order to achieve the above object, the optical matrix switch according to the present invention branches optical input signals inputted from 2 ⁱ⁺¹ (i is a natural number) input ports into 2 ⁽²ⁱ⁺¹⁾ branch signals. An input branching tree having a tree configuration consisting of 1×2 branching means of stages, and an input branching tree connected in series to the input branching tree to form one stage of a redundant tree configuration, and branched and supplied from different input ports. 2 2i ^2x2 optical switches that input two types of branch signals, select one of the two outputs, and output a switch output signal;
A tree-configured output multiplexing tree consisting of i-stage 2×1 multiplexing means that inputs the switch output signal of the ×2 optical switch and outputs the optical output signal to the 2 ⁱ⁺¹ output ports. It is characterized by having

[Effect]

According to the present invention, since the optical matrix switch is configured as described above, the optical input signals supplied to the 2 ^{i +1} input ports are divided into 1×2 demultiplexers of i stages constituting the input demultiplexing tree. The signal is branched into 2 ⁽²ⁱ⁺¹⁾ branch signals by a means, and is supplied to two inputs of 2 ²ⁱ 2×2 optical switches which together with the input branch tree form one stage of an overlapping tree. Either of the branch signals from the different input ports is selected and the two
2 ⁱ⁺¹ is output as a switch output signal to one of the outputs, and is generated by i-stage 2×1 multiplexing means constituting an output multiplexing tree into which the switch output signal is input.
The output signal is branched and supplied to the outside from the output port.

〔Example〕

Embodiments of the present invention will be described below with reference to FIG. 1 and FIGS. 4 to 7 of the accompanying drawings.

FIG. 1 is a functional circuit showing the configuration of one embodiment. 2 ⁱ⁺¹ (i is a natural number) input port 1 (1
Optical input signals are input from ₍₁₎ , 1 ₍₂₎ , 1 ₍₃₎ , ...1 (2 ⁱ⁺¹ )). Input port 1 has i-stage 1×
An input branching tree 2 constituted by a tree consisting of two branching means 2 ₁₍₁₎ , 2 _{1(2) ,...2 i(1)} , 2 _i(2) ,... 2 i(2 ²ⁱ ) is connected. , i-th stage 1×2 demultiplexing means 2 _i(1
) , 2 _i(2) , . . . 2 i (2 ²ⁱ ), 2 ⁽²ⁱ⁺¹⁾ types of branch signals are obtained. The i-th stage 1×2 demultiplexer 2 _i further includes 2 ²ⁱ 2×2 optical switches 3.
(3 ₍₁₎ , 3 ₍₂₎ , . . . 3 (2 ²ⁱ )) are connected as the i+1 stage of the tree, forming a composite tree together with the input branch tree 2. Here, 2 ²ⁱ 2×2
Two inputs of each optical switch 3 are connected to receive branch signals branched from different input ports 1. Furthermore, two outputs are output from each of the 2×2 optical switches 3, and the total number of outputs from the 2×2 optical switches 3 is 2 ^(2i+1). It is connected to the input of 2 ²ⁱ 2×1 multiplexing means 4 ₁₍₁₎ , 4 ₁₍₂₎ , ... 4 1(2 ²ⁱ ), and sequentially passes the i-th stage multiplexing tree through the i-stage multiplexing tree. 2×1 multiplexing means 4 _i(1) , 4 _i(2
) , ...4 i (2 ⁱ⁺¹ ), and the output of the i-th stage multiplexing means 4 _i is connected to the input of 2 ^{i +1} output ports 5 (5 ₍₁₎ ,
5 ₍₂₎ ,...5 (2 ⁱ⁺¹ )). and,
Above input port 1 and input branch tree 2 and 2×
The connection between the two-way optical switch 3, the output multiplexing tree 4, and the output port 5 is formed by a waveguide 6, and the optical input signal input to any input port (for example, 1a) is connected to the input branching tree 2. and 2×2
It is configured to be output from any of all output ports 5 (5 ₍₁₎ , 5 ₍₂₎ ,...5(2i ⁺¹ )) via an optical switch 3 and an output multiplexing tree 4. .

Next, the operation of the above embodiment will be explained. The optical input signals input to 2 ⁱ⁺¹ input ports 1 (1 ₍₁₎ , 1 ₍₂₎ , ...1 (2 ⁱ⁺¹ )) are sent to i-stage 1×2 demultiplexing means (2
_i(1) , 2 _i(2) , ...2 i(2 ²ⁱ )) is branched into 2 ⁽²ⁱ⁺¹⁾ types of branch signals by the input branch tree 2, which is 2 ²ⁱ pieces. 2×2 optical switch 3 (3 ₍₁₎ , 3
₍₂₎ ,...3 (2 ²ⁱ )) input. Each 2×2 optical switch 3 receives branch signals branched from different input ports 1 into two inputs, selects one of the two outputs, and outputs a switch output signal. First-stage 2×1 multiplexing means connected to the output of the 2×2 optical switch 3, (4 ₁₍₁₎ , 4 ₁₍₂₎ ,...4 1(2 ²ⁱ ))
combines the input switch output signals and generates the second stage 2×1 multiplexing means (4 ₂₍₁₎ , 4 ₂₍₂₎ ,...4 2(2 ^2i-1 ))
The i-th stage 2×1 multiplexing means (4 _i(1) ,
4 _i(2) ^{, .}

Next, an embodiment in which the present invention is applied to a 4×4 optical matrix switch will be described with reference to FIG.

A is a circuit diagram showing the configuration of this embodiment, and B is an explanatory diagram of the operation of a 2×2 optical switch forming a redundant tree. The 4.times.4 optical matrix switch according to this embodiment corresponds to the case of i=1 in the ^{2.sup.i+ 1.times.2.sup.i+1} optical matrix switch shown in FIG.
The optical input signals input to the four input ports 1a, 1b, 1c, and 1d are sent to the 2×2 demultiplexing means 2a, 2b, and 2 of the input branch tree consisting of only one stage.
four 2×2 optical switches 3a,
It is connected to eight types of inputs: 3b, 3c, and 3d. And 2×2 optical switches 3a, 3b, 3
The eight types of outputs c and 3d are output from 2×1 multiplexing means 4a, 4b,
4 output ports 5a, 5 via 4c, 4d
b, 5c, and 5d. The four 2×2 optical switches 3a, 3b, 3c, and 3d are
It constitutes the second stage of the tree configuration for the input branching tree, and at the same time serves as the previous stage tree element for the tree 2×1 multiplexing means 4a, 4b, 4c, and 4d for the output multiplexing. This constitutes the elements of a so-called duplicate tree. Then, the connection between the output of the input branch tree and the input of the 2×2 optical switch is made via the waveguide 6, and
The connection between the 2×2 optical switch and the input of the output multiplexing tree is as follows. That is, the input of the 2x2 optical switch 3a is connected to one output of the 1x2 demultiplexing means 2a and one output of the 1x2 demultiplexing means 2b of the input demultiplexing tree, and the input of the 2x2 optical switch 3a is The input of the 2x2 optical switch 3c is connected to the other output of the 1x2 demultiplexer 2a and the other input of the 1x2 demultiplexer 2b, and the input of the 2x2 optical switch 3c is connected to one output of the 1x2 demultiplexer 2c. and one output of the 1×2 demultiplexing means 2a,
The input of the 2x2 optical switch 3d is the 1x2 demultiplexing means 2.
c and the other output of the 1×2 demultiplexer 2d. Also, 2x2 optical switch 3
The output of a is the 2×1 multiplexing means 4 of the output multiplexing tree.
The output of the 2x2 optical switch 3b is applied to one input of the 2x1 multiplexing means 4c and one of the 2x1 multiplexing means 4d. At the input, the output of the 2x2 optical switch 3c is 2x
The output of the 2x2 optical switch 3d is connected to the other input of the 2x1 multiplexing means 4a and the other input of the 2x1 multiplexing means 4b, and the output of the 2x2 optical switch 3d is connected to the other input of the 2x1 multiplexing means 4c.
They are respectively connected to the other input of the multiplexing means 4d.

Next, the operation of the above embodiment will be explained. Optical input signals input from input ports 1a, 1b, 1c, and 1d are transferred to 1×2 branching means 2a, 2b, 3c, and 3d.
The signals are branched by a branching tree consisting of 2 x 2 optical switches 3a, 3b, 3c, and 3d. At this time, 2x2 optical switches 3a, 3b, 3c, 3d
Each of the two inputs has a different input port 1.
Branch signals branched from a, 1b, 1c, and 1d are supplied. And 2x2 light switch 3
The switch output signals outputted from a, 3b, 3c, and 3d are multiplexed by an output multiplexing tree, and output as optical output signals to output ports 5a, 5b, 5c, and 5d. The operation of the 2.times.2 optical switch will now be described in more detail with reference to FIG. 4B. 2
The x2 optical switch 3 has the function of inputting two inputs 31a and 31b and selecting and outputting two outputs 32a and 32b, and is functionally equivalent to the tree shown on the right side of the figure. That is, this duplicate tree is input 31a
A 1×2 optical switch 33a that demultiplexes the signal, optical switches 34a and 34b that are tree-connected, and an input 31
A 1×2 optical switch 33b that demultiplexes the
It consists of optical switches 34a and 34b that receive and combine the outputs of x2 optical switches 33a and 33b. The 2×2 optical switch on the left side of the figure is thus a demultiplexing means, which forms part of the demultiplexing tree, and a multiplexing means, which forms part of the demultiplexing tree. ing. This is why it is called a duplicate tree.

As a result, the 4×4 optical matrix switch of FIG.
The function of supplying the optical input signal input to 1d to all output ports 5a, 5b, 5c, and 5d via the 2×2 optical switch and output multiplexing tree that constitute the input branching tree and the redundant tree. It functions in the same way as a so-called non-blocking 4×4 optical matrix switch. The input branching tree and the output multiplexing tree perform the branching or multiplexing of information with almost no crosstalk, and the 2×2 optical switch 3a, which is one stage in the duplication tree,
In the case of 3b, 3c, and 3d, since the tree is replaced with a single optical switch, crosstalk increases slightly, but the structure is simplified and miniaturization is facilitated.

The circuit configuration shown in FIG. 8 is equivalent to that shown in FIG. 4A. That is, the output multiplexing means 4a', 4 in FIG.
b', 4c' and 4d' are 4a and 4a in Figure 4A, respectively.
4c, 4b and 4d, output port 5
a', 5b', 5c' and 5d' correspond to output ports 5a, 5c, 5b and 5d, respectively.

Next, an example of an 8×8 optical matrix switch will be described. FIG. 5 is a circuit diagram showing the configuration of the embodiment,
i= in the optical matrix switch shown in Figure 1
This corresponds to case 2. The difference from the embodiment shown in FIG. 4 is that the input ports 1a, 1b, . . .
1d is composed of 8 ports, and the input branch tree is 2
2 _1a , 2 _1b ...2 ₂ in the 1×2 demultiplexing means consisting of stages
_a ...2 _2p , and the resulting 32 types of branch signals are sent to 16 2x2 optical switches 3a, 3b,
... 3p, and further inputs 32 types of switch output signals, an output multiplexing tree is formed by a 2×1 multiplexing means 4 _1a , 4 _1b , 4 _1b , 4 1p , 4 _2a , 4 _2h consisting of two stages. Output port 5 consisting of 8 ports
The point is that output signals are connected to a, 5b, . . ., 5h.

In the configuration shown in FIG. 5, unit matrices 7a, 7b, 7c, and 7d are formed by using the 4×4 optical matrix switch according to the configuration shown in FIG. ,1b
...2 groups 1a to 1d, 1e to 1h with 4 ports each
It can be understood that the input signal is divided into 1 h and allocated to each unit matrix 7 by the 1×2 demultiplexing means 2 _1a , 2 _1b . . . 2 _{2 h} that constitute the first stage of the input demultiplexing tree. can.

FIG. 6 shows the configuration of an optical matrix switch that has been generalized by such unitization. The input signals input to 2 ⁱ⁺¹ input ports 1 ₍₁₎ , ... 1 (2 ⁱ⁺¹ ) are divided into two parts of 2 ⁱ each, and then sent to 1×2 demultiplexing means 2 ₍₁₎ , ...2 (2 ⁱ ) or 2 (2 ⁱ +1), ...2 (2 ⁱ⁺¹ ) are supplied to separate unit matrices 7a, 7b or 7c, 7d, respectively. Then, the output from the unit matrix is again transferred to the 2×1 multiplexing means 4 ₍₁₎ ,...4
(2 ⁱ ) or 4 (2 ⁱ +1), ...4 (2 ⁱ⁺¹ ), and the output signal is output from the output port 5 at 2 ^{i +1} locations.
Output from ₍₁₎ ,...5 (2 ⁱ⁺¹ ).

Each of the 2 ⁱ × 2 ⁱ unit matrices 7a to 7d is
It has a configuration in which i is replaced with i+1 in FIG.
However, when this is repeated until i=1, the unit matrix is a 2×2 optical switch shown in FIG. 4B.

As mentioned above, by introducing the concept of a unit matrix, the 4×4 matrix switch shown in FIG. 4A can also be generalized by considering a 2×2 optical switch as a unit matrix. This can be understood as the configuration shown in FIG.

The circuit configuration shown in FIG. 9 is equivalent to that shown in FIG. Similar to the relationship between FIG. 8 and FIG. 4A, it can be seen that the arrangement of the multiplexing means on the output side and the output port is simply changed. however,
Such a difference affects the dimensions of each stage of the optical switch, demultiplexing means, etc., so it is necessary to take this into consideration when choosing which circuit configuration to adopt.

In the above embodiment, the 1×2 demultiplexing means is constituted by a 1×2 optical switch, and the 2×
Although it is possible to configure one multiplexing means with a 2×1 optical switch, it is also possible to configure only the 1×2 demultiplexing means with a 1×1 optical switch.
It is also possible to configure it with a two-light switch.

In particular, when the demultiplexing means is configured with a passive demultiplexing means, the seventh
As shown in the figure, ON/OFF switches 35a, 35
b can be provided to configure an optical matrix switch. According to this, if the input branching tree and the output multiplexing tree are configured by passive branching and multiplexing, the optical matrix switch will have two stages of optical switches, so the entire optical matrix switch will be The configuration of the device can be reduced in size, and especially when the number of input ports is 8 or more, the size can be significantly reduced.

In addition, in FIG. 4A, FIG. 5, FIG. 6, FIG. 8, and FIG. 9, it is also possible to adopt a matrix configuration in which the input ports and output ports are exchanged.

FIGS. 10A to 10C show still another embodiment of the present invention. In these embodiments, the connection of the waveguide itself is the same as the embodiments shown in FIG. 8, FIG. 4A, and FIG. 9, respectively. The difference is the light switch,
This is a point where at least a part of the waveguide connecting the demultiplexing means and the multiplexing means is formed by a curved line. That is, when the direction of a straight line connecting the same number of input ports and output ports is called the longitudinal direction, and the direction perpendicular to this is called the lateral direction, optical switches, demultiplexing means, At least a portion, for example, the entire waveguide 61 connecting the multiplexing means is formed by a curved line. In the illustrated example, the curve consists of a combination of circular arcs. Using curved waveguides allows the longitudinal dimensions of the optical matrix to be reduced.

This point will be explained below with reference to FIG. The figure shows two curved waveguides W that intersect with each other.
1 and W2, each curved portion is made up of a combination of circular arcs centered on points R1 to R4.

As shown in the figure, it is assumed that each element E1 to E4 such as an optical switch, a demultiplexer, a multiplexer, etc. are aligned in the longitudinal direction and the lateral direction, and the distance between the elements in the lateral direction is Y, and the distance between the elements in the lateral direction is Y, and the distance between the elements in the lateral direction is Let the distance between the elements be X in the case of a curved waveguide and X' in the case of a straight waveguide. The dimensions X and X' are constrained by the branch angle 2α of the branch portion and the crossing angle 2θ of the intersection portion. That is, the branching angle 2α is usually about 2° or less, and even if a special one is used, it can only be increased to about 4°. On the other hand, the crossing angle 2θ needs to be approximately 4 to 7 degrees or more to avoid crosstalk.

If the branching angle 2α is the same for a straight waveguide and a curved waveguide, the curved waveguide can have a larger crossing angle 2θ than the straight waveguide, and the distance between the elements can be increased. You can shorten the X. In addition, in the case of a straight waveguide, the crossing angle and branching angle are the same, and a special optical switch is required to obtain the required crossing angle (4 to 7°), but in the case of a curved waveguide, Since the branching angle can be made smaller than the crossing angle, no special optical switch is required.

The effect of using such a curved waveguide is
The problem becomes more noticeable as the distance Y between elements to be connected to each other increases. This will be explained below with reference to FIG. For simplicity, the branching angle 2α is set to 0.
shall be. Further, 2θ, X, and Y refer to the same ones as in FIG. 11, and R indicates the radius of curvature of the curved waveguide.

Considering the branch point E1 as the origin of the xy coordinates, from the circle equation, (X/2) ² + {(Y/2) - R} ² = R ² ... (1) Therefore, X = 2 (YR −Y ² /4) ^1/2 ...(2) If Y is sufficiently small compared to R, X = (4YR) ^1/2 ...(3) As can be seen from equation (3), Y increases Even if X is Y
It only increases in proportion to the (1/2) power of
On the other hand, in the case of a straight waveguide, X increases in proportion to Y. Therefore, the larger the distance Y between elements to be connected to each other, the more significant the effect of using a curved waveguide (X is relatively small) becomes. As the number of ports increases, it becomes necessary to connect elements with larger lateral distances, and a larger value of Y exists. It can be said that the effect is large.

Examples of trial calculation results for the total length of the optical matrix switch are shown in FIGS. 13 and 14. Figure 13 is a plot corresponding to the number of ports.
is a case where the demultiplexing means and the multiplexing means are configured passively without using any optical switches (the length of the element part ls =
0), B in the same figure is a complete type, that is, when all the demultiplexing means and multiplexing means are configured with optical switches (element length ls = 2), and C in the same figure is a simple type, i.e., when the demultiplexing means and multiplexing means are configured with optical switches. A case is shown in which one of the wave means is entirely composed of optical switches (element length ls = 2) and the other is entirely composed of passive means (element length ls = 0). The radius of curvature R is shown as 10, 20, and 30 mm. In the case of A in the same figure, for R = 10, 20, and 30 mm, the distance between the elements of the 2 × 2 optical switch Yu
(Y=direction) and Xu (X direction) are as follows, respectively.

When R=10mm, Yu=0.04mm, Xu=1.2mm When R=20mm, Yu=0.074mm, Xu=2.4mm When R=30mm, Yu=0.112mm, Xu=3.6mm It is assumed that the examination angle 2θ is 7°.

FIG. 14 is a plot of the total length of a 16×16 type optical matrix switch corresponding to the radius of curvature. Figure A shows a complete type, and Figure B shows a simplified type.

〔Effect of the invention〕

As described above, according to the present invention, an input signal inputted from an input port is inputted to an input branching tree consisting of a 1×2 branching means, and a switch is performed by a 2×2 optical switch that is connected to the input branching tree to form a redundant tree. The output signal is obtained, inputted to the output multiplexing tree consisting of 2x1 multiplexing means, and the output signal is obtained at the output port, so the number of stages of optical switches through which the information input to the input port passes is 2log. ₂ 2 ⁱ⁺¹ −1, and light loss can be reduced. Furthermore, according to the present invention, an optical matrix switch can be configured with 3(i+1) ² /4 optical switches compared to a normal tree configuration without increasing crosstalk. Furthermore, compared to an optical matrix switch consisting of a simple square grid arrangement, an optical matrix switch can be configured with about 5/4 times as many optical switches.
This has the effect of providing a compact optical matrix switch that has excellent selection functions and little crosstalk.

Furthermore, if a curved waveguide is used, the length of the optical matrix can be further reduced.

[Brief explanation of drawings]

FIG. 1 is a circuit diagram showing the configuration of one embodiment, FIGS. 2 and 3 are explanatory diagrams of the prior art, FIG. 4 is a circuit diagram showing the configuration of another embodiment and its operation explanatory diagram, and FIG. The figure is a circuit diagram showing the configuration of still another embodiment.
The figure is a circuit diagram showing the structure of a generalized optical matrix switch according to the present invention, FIG. 7 is a circuit diagram showing another structure of the optical switch applicable to the embodiment of the invention, and FIGS. Figures 11 and 12 are diagrams showing other embodiments of the invention, and Figures 13 and 1 are diagrams showing changes in dimensions when a curved waveguide is used.
FIG. 4 is a diagram showing the trial calculation result of the total length of the optical matrix switch. 1...Input port, 2...Input branch tree, 3
...2x2 optical switch, 4...output multiplexing tree,
5...Output port.

Claims (1)

[Claims] An input port at 1 2 ⁱ⁺¹ (i is a natural number), and 2 ⁱ⁺¹
In the optical matrix switch, which has ^{2 output ports and selects the optical input signals supplied to the input ports in a matrix and outputs the optical input signals from the output ports as optical output signals, 1)} An input branching tree formed into a tree by i-stage 1×2 branching means that branches into different types of branching signals, and an input branching tree connected in series to the input branching tree to form one stage of an overlapping tree structure. 22i 2x2 optical switches which input two types of branched signals branched and supplied from different input ports, select one of the ^two outputs, and output a switch output signal; An output tree-configured by i-stage 2×1 multiplexing means, which inputs the switch output signal of the 2×2 optical switch and outputs the optical output signal to the 2 ⁱ⁺¹ output port. An optical matrix switch characterized by comprising a multiplexing tree.