SIGNAL POWER BASED ROUTING IN WDM ALL-OPTICAL NETWORKS

Abstract

In a Wavelength Division Multiplexed (WDM) all-optical network having optical nodes linked together by optical fibers carrying communication channels of separate wave-lengths, a method is applied in order to extend the size of the all-optical network. According to the method, a model is set up for the all-optical network in which a wavelength graph containing vertices and edges represents the physical effects of the nodes and links. After identifying a set of demand, an objective function is calculated expressing the total routing cost and total signal power demand of the channels. The minimum of the objective function returns the signal powers for each channel to be tuned.

Description

TECHNICAL FIELD

The present invention relates to optical networks. In particular, the present invention is directed to a method and a node for defining signal power in a Wavelength Division Multiplexed (WDM) all-optical network.

BACKGROUND

In fiber-optic communications, WDM is a technology which multiplexes multiple optical carrier signals on a single optical fiber by using different wavelengths (colours) of laser light to carry different signals. This allows for a multiplication in capacity, in addition to enabling bidirectional communications over one strand of fiber. A WDM system uses a multiplexer at the transmitter to join the signals together and a de-multiplexer at the receiver to split them apart. With the right type of fiber it is possible to have a device that does both simultaneously, and can function as a so called optical add-drop multiplexer (OADM). Older, or fixed, OADMs cannot configure capacity at a node. The technology called Reconfigurable Optical Add/Drop Multiplexer (ROADM) represents a real breakthrough for WDM networks providing the flexibility and functionality required in today's complex networking environment. In contrast to OADMs, ROADMs a low service providers to reconfigure add and drop capacity at a no node remotely, e.g. to define channels branching-off, reducing operating expenses by eliminating the time and complexity involved in manual reconfiguration. Furthermore, ROADMs allow the service provider to design an optical network once, and then never to have to worry about it again, regardless of how the network grows. The design phase in a ring topology ensures that any node can communicate with any other node using any wavelength. ROADM functionality is part of the solution for a deployable system, but in order to exploit the benefits of ROADM in metro WDM networks increased data management capabilities on individual wavelengths are also needed. (Metro network is WDM network. with a range of 40-500 kms). For instance, ROADM rings are very sensitive to topology changes, and need close monitoring and control of wavelength power to keep the system in balance. The real innovation lies in the system engineering related to the ROADM function, addressing per-wavelength power measurement and management, and per-wavelength, fault isolation. Almost every optical system vendor has ROADM with wavelength monitoring functions in commerce.

The next step for fully reconfigurable WDM optical networks are the tunable Small Form-factor Pluggable (SFP), where the wavelength allocation is changed as the needs of the network change. The other innovations are the tunable dispersion compensation elements. These network elements are ready-made products and can be purchased nowadays. The evolution of the optical networks seems to tend towards a fully reconfigurable network where the control and management plane will have new functions such as determining the signal quality, tuning the wavelength frequency, setting dispersion compensation units and by using Variable Optical Attenuators (VOA) setting the channel powers. Of course the traditional functions such as routing will remain its main function.

In an all-optical network the problem of routing and wavelength assignment (RWA) is critically important for increasing the efficiency of wavelength-routed. Given the physical network structure and the required connections, the RWA problem is to select a suitable path and wavelength among the many possible choices for each connection so that no two paths sharing a link are assigned to the same wavelength.

Solutions are known for this type of technology like a non-disruptive lightpath routing described in patent application U.S. 2004/0109683 A1, that presents the possibility of modifying the routing lightpaths without interrupting service.

Another solution is described in patent publication WO 2006/000510 A1 relating to optical path feasibility in an optical communication network presenting a signal quality calculation method based on physical impairments.

In both publications mentioned above, the power of certain channels within a fiber is set uniformly to equal levels. This is one of the remaining effects of the point-point optical networks. Of course there is technical simplification using this kind of channel power allocation: Using the same channel powers the nonlinear effects will have the smallest impact onto the signal quality.

This allocation schemes leads to the problem that in many cases where a channel is dedicated to a connection for a short distance due to the same channel powers the signal quality will be unnecessarily good i.e. the channel is over engineered

Additionally, in patent specification U.S. Pat. No. 7,068,932 a method and system for automatic initialization of an optical network are provided. A network management system (NMS) performs remote determination of span losses and sets the operating points of network components. The initialization method comprises remotely and automatically setting target gains of optical amplifiers and signal power levels at transmitters and receivers to required operating values. The methods for initialization of the optical network of the embodiments include gain excursion minimization (GEM) for individual channels passing through amplifiers and/or pre-emphasis of the optical link, where channel powers at the transmitters are biased to compensate for the effects of optical amplifiers gain ripple. U.S. Pat. No. 7,068,932 does not describe a network-wide optimum taking into account of constrains of links and nodes in the network.

As the size of the network that can be reached all-optically is a function of the signal power, it would be possible to use higher signal powers that would increase the size of the all optical networks, i.e. would decrease the numbers of optical-electrical-optical conversions without increased physical impairments.

It is therefore seen to be desirable to construct a method for an all-optical network where it is allowed to tune the signal power of individual channels in order to achieve an increased size of the whole network in which the quality of signals remains within an acceptable level.

SUMMARY

The present invention involves a method, node and an all-optical network, which solve the aforementioned problem that will become apparent from an understanding of the following description.

According to the invention, the objective set was achieved by a method applied to a WDM all-optical network in which optical nodes are linked together by optical fibers carrying communication channels of separate wavelengths. In the first step a model is set up for the all-optical network. In the model a wavelength graph contains vertices and edges, a part of which represents switching function inside a node while other edges represent wavelengths of a physical link. Then an objective function is calculated expressing the total routing cost and total signal power demand of the channels. Looking for the minimum of the objective function, the signal powers for each channel are determined and tuned.

In a possible embodiment, the objective function may be defined to return physical effects of both optical fibers and nodes as the function between maximum allowable distance and the signal power. The physical effect of a node may be defined to be equal to a certain length of an optical fiber, e.g. 90 kms.

The objective function may be defined to use constants as input parameters, e.g. an upper limit of the total signal power in physical links, or linear relation between the input power of a demand and the maximum distance the signal is allowed to reach, or a parameter expressing the tradeoff between minimal total routing cost and minimal total signal power of the channels, or a constant for expressing a maximum allowable signal power for one channel.

In another aspect, the present invention is directed to a node for signal power based routing in a WDM all-optical network. The node comprises VOAs which are adapted to tune individual signal powers of channels so that total routing cost and total signal power of the channels are minimized. The node may be an OXC or a ROADM.

In yet another aspect, the present invention is directed to an all-optical network in which optical nodes are linked together by optical fibers carrying communication channels. At least a part of the optical nodes are adapted to tune signal power individually so that total routing cost and total signal power of the channels are minimized.

The method of the invention may also be embodied in a computer program product and a memory coupled to the processor, wherein the memory is encoded with one or more programs that may perform the method based on the model.

The most important advantage of the invention is a global optimum that provides a larger all-optical network with the same quality requirements of the traditional RWA routing allocation scheme.

It is also advantageous that increasing the number of wavelengths the performance will improve.

Another advantage is that it works reasonable faster, and it can also be considered as an Impairment Constraint Based Routing (ICBR) technology where the physical effects are taken into account while routing decisions are taken.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following, the essential features of the invention will be described in detail by an example for an embodiment illustrated by the figures of the attached drawing. In the drawing,

FIG. 1A, 1B, 1C and 1D are illustrations for the basic concept of the invention in an all-optical network.

FIG. 2 is a flowchart for illustrating method steps.

FIG. 3 schematically illustrates a model for the method.

FIG. 4 is a block diagram illustrating an OXC.

FIG. 5 is a block diagram of a ROADM.

DETAILED DESCRIPTION

FIG. 1A, 1B, 1C and 1D illustrate how the proposed algorithm gives obviously better results than the traditional RWA algorithms in an all-optical network. This is due to the additional degree of freedom namely, the possibility to tune the signal powers. For simplicity, let us consider a network with all-optical nodes “A”, “B”, “C”, “D”, “E” carrying two channels with wavelengths λ₁, λ₂. Assume, that in FIG. 1A the node “A” can only reach node “C” (solid lines) in the traditional routing schemes. Consider two demands between node “A” and “B”, and node “A” and “D”. In case of traditional routing schemes the demand between nodes “A” and “D” can not be established due to physical layer constraints, e.g. due to amplified spontaneous noise (dotted lines between node “C” and “D”). The demand between node “A” and “B” can be established but the signal quality would be unnecessary good, i.e. the connection is over-engineered. Using the proposed routing schemes according to FIG. 1B, we can increase the signal power of channel λ₂ and in this way it is possible to establish a connection between node “A” and “D” (solid line with wavelength λ₂). Of course while increasing the signal power of channel λ₂ the signal power of channel λ₁ must be decreased, but this signal power is still enough to establish the connection between node “A” and “B”.

In FIG. 1C and 1D a typical routing problem is depicted. We would like to establish a connection between node “A” and “C” (dashed lines). Consider that in the route “A”-“B”-“C” all the wavelengths are in use for other demands. The only solution remains using longer routes like “A”-“D”-“E”-“C”. Due to physical constraints this connection cannot be done (dotted lines). However, using the proposed routing schemes in FIG. 1D, we can establish this connection (solid line with wavelength λ₂) by increasing the signal power of λ₂ between nodes “A”-“D”-“E”-“C” and decreasing the signal power of λ₁ between nodes “A”-“D”- “E”- “C” at the same time.

In FIG. 2, method steps of the invention are illustrated. In the first step S201 a model is set up in which a wavelength graph contains vertices and edges representing switching function inside a physical node and other edges represent wavelengths of a physical link. (The model will be described in detail in FIG. 3).

In the second step S202 a calculation is carried out to set up an objective function expressing the total routing cost and total signal power demand of the channels. The following Integer Linear Programming (ILP) formulation to route the demands was used.

Constants:

P _pl ^max=4−20 dBm typically 14 dBm   (1)

Constant P_pl^max means the upper limit of the total power in physical link pl expressed in dBm.

len_ij   (2)

Constant len_ij is the length of the physical link used by the considered λ channel.

len_PhyNode=90 km typically   (3)

Constant len_PhyNode corresponds to the length of a fiber that has the same impairments as the switching device induces along the path of the demand.

L_c=1200   (4)

Constant L_c is the factor of the linear relation between the input power of a demand (in mW) and the maximum distance the signal is allowed to reach.

α=0, 5   (5)

Constant α expresses the tradeoff between optimization objectives: minimal routing cost or minimal power.

s°, t°   (6)

Symbols s° and t° represent source and target of demand o.

$\begin{array}{cc}\beta =\frac{n}{W}·P_{\mathrm{pl}\mathrm{lin}}^{\max }& \left(7\right)\end{array}$

Constant β express the maximum allowable signal power for one channel in mW. Where n is between 1 and W, and W is the number of wavelengths in a fiber. For example in case n=2 the maximum allowed channel power would be the double of the cannel power obtained from the traditional allocation schemes.

P_pl^max_lin   (8)

Constant P_pl^max_lin means the upper limit of total power in physical link pl in mW.

Variables:

$\begin{array}{cc}{p}^o\in \left[0,\frac{\beta }{P_{\mathrm{pl}\mathrm{lin}}^{\max }}\right],\forall o\in O& \left(9\right)\end{array}$

Variable p° denotes the input power of demand o divided by P_pl^max_lin.

$\begin{array}{cc}{p}_{\mathrm{ij}}^o\in \left[0,\frac{\beta }{P_{\mathrm{pl}\mathrm{lin}}^{\max }}\right],\forall \left(i,j\right)\in A,\forall o\in O& \left(10\right)\end{array}$

Variable p_ij^o means the power of demand o on edge (i, j) divided by P_pl^maxlin.

y _ij ^o∈{0, 1}, ∀(i, j)∈A, ∀o∈O   (11)

Variable y_ij^o indicates whether demand o uses edge (i, j) or not.

Constraints:

$\begin{array}{cc}\sum _{\forall o\in O}\sum _{\forall {\left(i,j\right)}^{\mathrm{pl}}}p_{\mathrm{ij}}^o\le 1,\forall \mathrm{pl}\in \mathrm{PhyLinks}& \left(12\right)\\ p_{\mathrm{ij}}^o\le y_{\mathrm{ij}}^o,\forall \left(i,j\right)\in A,\forall o\in O& \left(13\right)\\ \sum _{\forall j\in V^{\to i}}p_{\mathrm{ji}}^o-\sum _{\forall k\in V^{i\to }}p_{\mathrm{ik}}^o=\{\begin{array}{cc}\hspace{0.17em}-p^o& \mathrm{if}i=s^o\\ 0& \mathrm{if}i\notin \left\{s^o,t^o\right\},\forall i\in V,o\in O\\ +p^o& \mathrm{if}i=t^o\end{array}& \left(14\right)\\ \sum _{\forall j\in V^{\to i}}y_{\mathrm{ji}}^o-\sum _{\forall k\in V^{i\to }}y_{\mathrm{ik}}^o=\{\begin{array}{cc}\hspace{0.17em}-1& \mathrm{if}i=s^o\\ 0& \mathrm{if}i\notin \left\{s^o,t^o\right\},\forall i\in V,o\in O\\ +1& \mathrm{if}i=t^o\end{array}& \left(15\right)\\ \sum _{\forall o\in O}y_{\mathrm{ij}}^o\le 1,\forall \left(i,j\right)\in A& \left(16\right)\\ \sum _{\forall {\left(i,j\right)}^{\mathrm{sw}}}y_{\mathrm{ij}}^o·{\mathrm{len}}_{\mathrm{PhyNode}}+\sum _{\forall {\left(i,j\right)}^{\mathrm{pl}}}y_{\mathrm{ij}}^o·{\mathrm{len}}_{\mathrm{ij}}\le L\left(p^o\right)=L_c·p^o·P_{\mathrm{pl}\mathrm{lin}}^{\max },\forall o\in O& \left(17\right)\end{array}$

Objective function:

Minimize

$\alpha ·\left[\sum _{\forall o\in O}\sum _{\left(i,j\right)\in A/A_{\mathrm{sw}}}y_{\mathrm{ij}}^o\right]+\left(1-\alpha \right)·\left[\sum _{\forall o\in O}p_o\right]$

Constraint (12) explains that the sum power of demands traversing a physical link (fiber) cannot exceed the maximum allowed power of that link.Constraint (13) indicates that if the power of demand o in edge (i,j) is larger then zero, then edge (i,j) is used by demand o.Constraints (14) and (15) express the flow-conservation constraint of the power and of the y decision variables, respectively, for every demand.Constraint (16) indicates that a given edge can be used by only one demand.Constraint (17) indicates that the total length of demand o should be less than the distance allowed by the input power of demand o.The objective function expresses that the sum of the used edges should be minimized together with the sum of input powers of demands. If we want to minimize the total cost of routing, constant cost factors should be assigned to each edge.Constant α decides whether optimization emphasis is on minimal routing cost (α is close to 1) or on minimal input power (α is close to zero).

In the third step S203, the signal power of each channel is individually tuned so that the objective function is minimized.

In FIG. 3, a part of the model of the network consisting of physical links 301 and physical devices (nodes) 302 is represented by a wavelength graph 300. The wavelength graph 300 contains vertices V (only one is indicated) and edges A depicted by solid, dashed and dotted lines for each wavelength. (Three wavelengths are shown in this example). Symbol A_sw (only one is indicated) denotes the set of edges representing switching function inside the node 302; other edges A represent wavelengths of a physical link 301. Arrows in some nodes 302 represent branch-off for certain channels. As it is clearly seen, the effect of physical nodes 302 and links 301 is represented by edges A, A_sw and vertices V in the model.

FIG. 4 illustrates an OXC as a possible embodiment of an optical node 400 comprising de-multiplexers 411, 412, . . . , 41k wavelength switches 421, 422, . . . , 42k, Variable Optical Attenuators 431, 432, . . . , 43k and multiplexers 441, 442, . . . , 44k. The input of the de-multiplexers 411, 412, . . . , 41k are optical fibers carrying channels of wavelengths λ₁, λ₂, . . . λ_k. The output of the de-multiplexers 411, 412, . . . , 41k are connected to the wavelength switches 421, 422, . . . , 42k the outputs of which are linked to the Variable Optical Attenuators 431, 432, . . . , 43k. The Variable Optical Attenuators 431, 432, . . . , 43k are adapted to tune individual signal powers to meet minimal total routing cost and minimal total signal power of the channels according to the method described in FIG. 2. The tuned channels are multiplexed in multiplexers 441, 442, . . . , 44k for the output of the node 400.

FIG. 5 illustrates an ROADM representing another possible embodiment of an optical node 500 comprising a de-multiplexer 510, wavelength switches 521, 522, . . . , 52k, Variable Optical Attenuators 531, 532, . . . , 53k and a multiplexer 540. The input of the de-multiplexer 510 are optical fibers carrying channels of wavelengths λ₁, λ₂, . . . λ_k. The outputs of the de-multiplexer 510 are connected to the wavelength switches 521, 522, . . . , 52k the outputs of which are linked to the Variable Optical Attenuators 531, 532, . . . , 53k. The Variable Optical Attenuators 531, 532, . . . , 53k are adapted to tune individual signal powers to meet minimal total routing cost and minimal total signal power of the channels according to the method described in FIG. 2. A part of the tuned channels are multiplexed in the multiplexer 540 for the output of the node 500. There are added and dropped channels branching-off from the node 500. The added channels are connected to the inputs of the wavelength switches 521, 522, . . . , 52k, while the other part of the tuned channels constitutes the dropped channels.

Those skilled in the art will appreciate that the method and model explained above may be implemented using software functioning in conjunction with a programmed microprocessor or a general-purpose computer. It will also be appreciated that while the current invention is primarily described in the form of a method, the invention may also be embodied in a computer program product and a memory coupled to the processor, wherein the memory is encoded with one or more programs that may perform the method disclosed in connection with FIG. 2 and based on the model disclosed in connection with FIG. 3.

Although the preferred embodiments of the present invention have been illustrated for OXC and ROADM, it is understood that the invention is not limited to these embodiments disclosed, but is capable of other type of optical nodes as set forth and defined by the following claims.

Claims

1. A method of signal power based routing in a Wavelength Division Multiplexed all-optical network in which optical nodes are linked together by optical fibers carrying communication channels of separate wavelengths (λ1, λ2, . . . λk), comprising the steps of: setting up a model for the all-optical network in which a wavelength graph contains vertices (V) and directed edges (A), a part of the edges (Asw)represents switching function inside a node while other edges represent wavelengths of a physical link,calculating an objective function expressing the total routing cost and total signal power demand of the channels, andindividually tuning the signal power of the channels so that the objective function is minimized.

2. The method of claim 1, in which the objective function is defined to return physical effects of both optical fibers and nodes as the function between maximum allowable distance and the signal power.

3. The method of claim 2, in which the physical effects of a node are defined to be equal to a certain length of an optical fiber.

4. The method of claim 1, in which the objective function is defined to use a constant (Pplmax) for an upper limit of the total signal power in physical links as an input parameter.

5. The method of claim 1, in which the objective function is defined to use a constant (Lc) as an input parameter for linear relation between the input power of a demand and the maximum distance the signal is allowed to reach.

6. The method of claim 1, in which the objective function is defined to use a constant (α) for expressing the tradeoff between minimal total routing cost and minimal total signal power of the channels.

7. The method of claim 1, in which the objective function is defined to use a constant (β) for expressing a maximum allowable signal power for one channel.

8. A node for signal power based routing in a Wavelength Division Multiplexed all-optical network, comprising Variable Optical Attenuators in which Variable Optical Attenuators are adapted to tune individual signal powers of channels so that total routing cost and total signal power of the channels are minimized.

9. The node of claim 8, in which the node is implemented as an Optical Cross Connect.

10. The node of claim 9, in which Optical Cross Connect comprises de-multiplexers wavelength switches, Variable Optical Attenuators and multiplexers, the input of the de-multiplexers are made of optical fibers carrying channels of a set of wavelengths (λ1, λ2, . . . λk),the output of the de-multiplexers are connected to the wavelength switches the outputs of which are linked to the Variable Optical Attenuators, andthe tuned channels are multiplexed in multiplexers for the output of the node.

11. The node of claim 8, in which the node is implemented as a Reconfigurable Optical Add Drop Multiplexer.

12. The node of claim 11, in which a Reconfigurable Optical Add Drop Multiplexer comprises a de-multiplexer, wavelength switches, Variable Optical Attenuators and a multiplexer, the input of the de-multiplexer are made of optical fibers carrying channels of a set of wavelengths (λ1, λ2, . . . λk),the outputs of the de-multiplexer are connected to the wavelength switches the outputs of which are linked to the Variable Optical Attenuators,a part of the tuned channels are multiplexed in the multiplexer for the output of the node,added and dropped channels are prepared to branching-off from the node, andthe added channels are connected to the inputs of the wavelength switches, while the other part of the tuned channels constitutes dropped channels.

13. A Wavelength Division Multiplexed all-optical network, in which optical nodes are linked together by optical fibers carrying communication channels, in which at least a part of the optical nodes are adapted to tune signal power individually so that total routing cost and total signal power of the channels are minimized.

14. The all-optical network of claim 13, in which a part of the nodes are Optical Cross Connects.

15. The all-optical network of claim 13, in which a part of the nodes are Reconfigurable Optical Add Drop Multiplexers.

16. (canceled)