WO2013091688A1

WO2013091688A1 - Multilayer topology representation for network

Info

Publication number: WO2013091688A1
Application number: PCT/EP2011/073581
Authority: WO
Inventors: Giovanni Fiaschi; Christian SCHENONE; Giulio Bottari
Original assignee: Telefonaktiebolaget L M Ericsson (Publ)
Priority date: 2011-12-21
Filing date: 2011-12-21
Publication date: 2013-06-27

Abstract

Planning a communications network involves generating (100) a multilayer topology representation. Nodes (30) of the network have configurable electrical switches (10), and a preconfigured static optical layer of optical links between the electrical switches of neighbouring nodes. Optical bypasses for the multilayer topology representation between non neighbouring ones of the nodes, via others of the nodes, are selected (120) for a subset of all the possible routes. The selected optical bypasses are added (130) to the static optical layer of the multilayer topology representation. As the topology representation is enriched by a selected subset of optical bypasses, more traffic can be routed to avoid the electrical switches, hence the network is more scalable before being limited by the high costs of electrical switching capacity and by the high complexity of routing if there are unlimited bypass paths.

Description

MULTILAYER TOPOLOGY REPRESENTATION FOR NETWORK Field

The present invention relates to methods of planning a communications network by generating a multilayer topology representation, to methods of operating a network using the multilayer topology representation, to corresponding computer programs for such methods, and to corresponding path computation entities for such networks. Background

It is well established that both optical and electrical technologies must be combined to build efficient communications networks, particularly for backbone or transport parts of networks, as opposed to access parts of networks. At the beginning, optical technologies only offered means to provide large bandwidth capacity on fibre and long haul point-to-point transmission, while electrical technologies provided switching and allocation flexibility for a better resource usage and for dynamic network reconfiguration. Although recent advances add flexibility and meshing in optical and scalability in electrical, both technologies still preserve their differences. To obtain the most from both, in place of opaque (electrical) or transparent (all optical) networks, it is preferable to use translucent (hybrid) networks.

Some advantages and drawbacks of state-of-the-art optical and electrical telecommunications can be summarized as follows. Optical lightpaths are difficult to configure, needing to consider physical impairments and feasibility constraints as well as the wavelength assignment problem, due to the cost of wavelength conversion. Optical switching, with the current technology, is also much slower than electrical switching.

Electrical switches, on the other hand, suffer from higher cost and lower scalability, hence it's desirable, wherever possible, to reduce their size, encouraging the use of optical bypasses. This justifies the usage of hybrid transport networks also known as translucent networks, having both electrical and optical switching technologies. This leaves an open question of how best to set up and configure such a translucent network in terms of relative resource usage.

While there is an extensive literature about operation and design of translucent networks, the proposed solutions generally fail to recognize really distinct roles to the two network components. The common approach is to add electrical switching and regeneration when optical is insufficient, but still considering the network as an integrated whole, where both components perform where possible the same functions at the same level in the network.

As a result, several drawbacks remain in such translucent networks. In particular, hybrid path provisioning is complex and slow, especially if compared with the same operation in a full electrical network (e.g. , SDH). This is partly because routing across all the electrical and optical switches, treating them as a single layer with a full mesh of paths, provides so many possible routes and so is not scalable. Also, by providing dynamic switching at the optical switches, setting up chosen routes is slow.

Summary

Embodiments of the invention provide improved methods and apparatus. According to a first aspect of the invention, there is provided a method of planning a communications network by generating a multilayer topology representation of the network, nodes of the network having configurable electrical switches, and the network having a preconfigured static optical layer of optical links between the electrical switches of neighbouring ones of the nodes, by selecting optical bypasses for the multilayer topology representation between non neighbouring ones of the nodes, via others of the nodes so as to bypass the configurable electrical switch of at least one of these other nodes, for a subset of all the possible routes between all the pairs of non neighbouring nodes. The selected optical bypasses are then added to the static optical layer of the multilayer topology representation. By having a topology representation of a static optical layer enriched by a selected subset of optical bypasses, more traffic can be routed to avoid the electrical switches, hence the network is more scalable before being limited by the high costs of electrical switching capacity and by the high complexity of routing if there are unlimited bypass paths. See figures 1 and 2 for example.

Any additional features can be added, some such additional features are set out below.

The method of planning can have the further step of using the multilayer topology representation for determining capacities of the electrical switches and the optical links and optical bypasses, to provision the network. This can provide better matching of the provisioning to the expected traffic. See figure 3 for example.

There can be a step of adding to the topology representation indications (S 12, S 1 3, S24, S35, S46, S56) of which of the optical links and optical bypasses share components with a risk of failure. This can enable more reliable restoration, since such indications can make it easier to ensure the restoration paths avoid such shared components. See figures 4 and 5 for example.

The selecting step can comprise selecting at least one route between all the pairs of non neighbouring nodes. This is one way of making a representative selection automatically. See figs 6 and 8 for example.

The selecting step can comprise selecting one of the nodes as a central node and selecting a limited number of routes sufficient to connect the central node to all other nodes which are not neighbours of the central node. This way enables a maximum distance of the bypasses to be reduced. See figs 9 and 10 for example. The selecting step can comprise selecting two or more nodes as primary nodes, providing optical bypasses to connect every different pair of the primary nodes and providing a limited number of optical bypasses sufficient to connect all other nodes which are not neighbours of the primary nodes to any one of the primary nodes. This way can enable multiple domains and gives more choice of maximum distance of the optical bypasses. See figs 1 1 and 12 for example.

The selecting of the optical bypasses can comprises selecting according to an indication of anticipated traffic levels on the optical bypasses or through the configurable electrical switches. The method can comprise using the topology representation, and a demand matrix of simulated connection requests based on anticipated connection requests or on records of actual connection requests, to carry out a simulated routing of the simulated connection requests. See fig 3 for example.

The method can be for reconfiguring an existing network and have the preliminary step of generating a multilayer topology representation of the optical links between neighbouring nodes of the existing network. This m ay be commercially valuable to exploit existing networks. See figure 13 for example.

The method can be a method of planning and operating a communications network in which case there can be the planning steps as set out above and subsequent steps of operating the network by routing new connection requests through the network according to the multilayer topology representation, and configuring the electrical switches along the selected route to set up the requested connection. See fig 2 for example.

Another aspect of the invention provides a method of operating a communications network having nodes, and having the step of routing new connection requests through the network according to a multilayer topology representation of the network, the nodes having configurable electrical switches, and the network having a preconfigured static optical layer of optical links between the electrical switches of neighbouring ones of the nodes. The multilayer topology representation has optical bypasses between non neighbouring ones of the nodes, via others of the nodes so as to bypass the configurable electrical switch of at least one of these other nodes, and the optical bypasses comprise a subset of all the possible routes between all the pairs of non neighbouring nodes. There is a further step of configuring the electrical switches along the selected route to set up the requested connection. This may enable more traffic to be routed to avoid the electrical switches and thus be more scalable and efficient. An additional feature is the step of altering a bandwidth allocation of the optical links or optical bypasses according to the connection requests in the operating phase, or according to the simulated connection requests in the planning phase. This may provide some flexibility in optical layer without needing complexity of dynamic optical switching. See figure 2 or 3 for example. Another such additional feature is the topology representation having indications of which of the optical links and optical bypasses share components with a risk of fai lure, and the method having the step of determ ining a restoration path for a connection according to the indications of shared components. This may provide more reliable restoration since the risk of the same fault causing failure of the restoration path can be reduced. See figure 14 for example.

Another aspect of the invention provides a computer program on a computer readable medium, which when executed by a processor, causes the processor to carry out a method of planning as set out above, or a method of operating a network as set out above. Another aspect of the invention provides a path computation element for a node of a communications network, the element having a store for storing a multilayer topology representation of the network, and being arranged to carry out the above methods of operating a network. This can for example involve routing of new connection requests through the network according to the multilayer topology representation. Nodes of the network have configurable electrical switches, and a preconfigured static optical layer of optical links between the electrical switches of neighbouring ones of the nodes. The multilayer topology representation has optical bypasses between non neighbouring ones of the nodes, via others of the nodes without using their electrical switches, and the optical bypasses comprise a subset of all the possible routes between all the pairs of non neighbouring nodes. The path computation element is arranged to set up the requested connection by dynamically configuring the electrical switches along the selected route. Any of the additional features can be combined together and combined with any of the aspects. Other effects and consequences will be apparent to those skilled in the art, especially over com pared to other prior art. Numerous variations and modifications can be made without departing from the claims of the present invention. Therefore, it should be clearly understood that the form of the present invention is illustrative only and is not intended to limit the scope of the present invention.

Brief Description of the Drawings:

How the present invention may be put into effect will now be described by way of example with reference to the appended drawings, in which:

Fig 1 shows a schematic view of a network used in a first embodiment, Fig 2 shows steps in a method according to a first embodiment,

Fig 3 shows an embodiment with simulated routing and provisioning, Fig 4 shows an example of a representation of a topology before enhancement,

Fig 5 shows a topology with optical bypasses and SRLGs,

Fig 6 shows a topology with four different bypass routes between nodes, Fig 7 shows steps for selecting optical bypasses by an all pairs strategy,

Fig 8 shows a topology example for the all pairs strategy,

Fig 9 shows steps for selecting by central node and star selection strategy,

Fig 10 shows a topology example for the central node and star strategy,

Fig 1 1 shows steps for selecting by multiple central nodes strategy,

Fig 12 shows a topology example for the multiple central nodes strategy,

Fig 13 shows steps for upgrading an existing network,

Fig 14 shows steps for an embodiment using the SLRGs for restoration, and

Fig 1 5 shows an exam ple of a node having a PC E according to an embodiment.

Detailed Description:

The present invention wi l l be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims. The drawings described are only schematic and are non-limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn to scale for illustrative purposes.

Abbreviations: GMPLS Generalized Multiprotocol Label Switching

IP Internet Protocol

NE Network Element

NMS Network Management System.

OTN Optical Transport Network

PCE Path Computation Element

ROADM Reconfigurable optical add drop multiplexer/demultiplexer

RSVP-TE Resource Reservation Protocol - Traffic Engineering

RWA Routing and Wavelength Assignment SDH Synchronous Digital Hierarchy

SNMP Simple Network Management Protocol

SRLG Shared Risk Link Group

WSS Wavelength Selective Switch

Definitions:

Where the term "comprising" is used in the present description and claims, it does not exclude other elements or steps and should not be interpreted as being restricted to the means listed thereafter. Where an indefinite or definite article is used when referring to a singular noun e.g. "a" or "an", "the", this includes a plural of that noun unless something else is specifically stated.

Elements or parts of the described nodes or networks may comprise logic encoded in media for performing any kind of information processing. Logic may comprise software encoded in a disk or other computer-readable medium and/or instructions encoded in an application specific integrated circuit (ASIC), field programmable gate array (FPGA), or other processor or hardware.

References to nodes can encompass any kind of switching node, not limited to the types described, not lim ited to any level of integration, or size or bandwidth or bit rate and so on.

References to switches can encompass switches or switch matrices or cross connects of any type, whether or not the switch is capable of processing or dividing or combining the data being switched.

References to software can encompass any type of programs in any language executable directly or indirectly on processing hardware.

References to processors, hardware, processing hardware or circuitry can encompass any kind of logic or analog circuitry, integrated to any degree, and not limited to general purpose processors, digital signal processors, ASICs, FPGAs, discrete components or logic and so on. References to a processor are intended to encompass implementations using multiple processors which may be integrated together, or co-located in the same node or distributed at different locations for example. References to optical paths can refer to spatially separate paths or to different wavelength paths multiplexed together with no spatial separation for example.

References to connection can encompass any kind of connection or circuit, or communications service between nodes of the network, and so on,

References to topology representations are intended to encom pass representations in any form, as a list of nodes and links, as any kind of data structure, as multiple lists or data structures and so on, held in distributed or centralized form, with or without additional information about node locations, link capacities, switch capacities and so on.

References to electrical layer or electrical switch are intended to encompass any electrical standard, particular examples include SDH/SONET, OTN, MPLS, IP, and Ethernet.

References to optical layer or optical switch are intended to encompass any optical standard, particular examples include WDM 40, 80, 48, 96 channels, 2.5G, 10G, 40G, 100G or flexigrid.

Introduction

By way of introduction to the embodiments, how they address some issues with conventional designs will be explained.

The known hybrid path provisioning is complex and slow, partly because routing across all the electrical and optical switches, treating them as a single layer with a full mesh of paths, provides so many possible routes and so is not scalable. Also, by providing dynamic switching at the optical switches, setting up chosen routes is slow. To address both these issues, embodiments described below involve finding and dynamically reconfiguring flexible optical bypasses, selected from all the possible optical bypasses. In some embodiments, there is a first phase during network planning, in which "eligible bypasses" are calculated, and in a second phase during operation, in which connectivity requests are dynamically served using existing or reconfigured bypasses.

A method to operate a translucent (electrical and optical) transport network is presented and schematized below. The electrical layer implements the final service that the transport network is going to offer and may use several technologies, e.g. , SDH, OTN or packet transport. The presented method is invariant with respect to the technology implemented at the electrical layer. The optical layer is treated as a separate layer and is preconfigured by setting up the optical paths through the optical switches before operation. Hence this is a static layer, for switching purposes, so that the delay of validating optical paths and configuring optical switches dynamically is not an issue. However the bandwidth of the optical links and bypasses can still be configured dynamically as this need not cause such delays.

Figures 1 and 2, a first embodiment.

Figure 1 shows a schematic view of a network used in a first embodiment. Three nodes 30 are shown for the sake of clarity, typically there would be many more, perhaps tens or hundreds. Each node has an electrical switch 10 and an optical switch 20. A transponder TP15 is used to couple the electrical switch and the optical switch and to convert signals between optical and electrical domains. A control plane 40 is provided spread across the nodes, with a PCE for routing and for dynamic configuration of the electrical switch.

Optical fibers link the optical switches of neighbouring nodes. The optical switches are configured either to pass an incoming optical signal through to another node, or to pass it to the electrical switch of its node. Where it is configured to pass the signal through, this forms an optical bypass path.

An off line planning tool 50 is provided to carry out planning operations, and to pass resulting configuration information to the nodes using the control plane, to set up the optical switches for example. The control plane is used to set up connections by selecting and using the preconfigured optical links and optical bypass paths and by configuring dynam ically the electrical switches. Two examples of connections are shown by dotted lines. Connection A passes from a client interface (not shown) through the electrical switch of the left-most of the three nodes, to the optical switch of that node. From there it is passed to an optical path which is preconfigured to pass through the optical switch of the middle node, to the optical switch of the right most of the nodes shown. Connection B is shown passing through the electrical switch of the right-most of the three nodes, to the optical switch of that node. From there it is passed to an optical path which is preconfigured to pass to the optical switch of the middle node, where does not pass through to the left-most of the nodes, but instead it is switched up to the electrical switch of the middle node. This electrical switch is dynamically configured to drop the signal to a client interface (not shown).

Figure 2 shows steps in a method according to a first embodiment. The method is split into two major phases: planning eligible optical bypasses and network operation. Em bodiments can have either or both of these major phases.

During the eligible bypass calculation, only the physical topology of the network is used as input in this example. Optical bypass routes are computed and these are used to enrich the topology of the network. Step 100 shows a step of planning a network by generating a multilayer topology representation, involving a number of sub steps as follows. At step 1 1 0, a representation is generated of a topology of a static optical layer of optical links between neighbouring nodes having electrical switches. At step 120 a selection of optical bypasses is made for a subset of all possible routes between non-neighbouring nodes via intervening nodes. The selected optical bypasses are added to the static optical layer of the multilayer topology representation at step 130.

The operation of the network operation is shown at step 140, with a number of sub steps. Connection requests are received at step 150. Considering requests towards the electrical layer, these are processed by routing using the enriched topology representation. This can be carried out by a path computation element at the ingress node or at another node for example. At step 160, the result of the routing (which may be a list of nodes along the route for example) is used to configure the electrical switches. This may involve using the control plane to pass the list of nodes along the path for example.

A link in the electrical topology is implemented by a path on the optical topology, which can be a physical link (fiber) or a bypass. If an optical path on the enriched topology already exists and has sufficient capacity to support the electrical path, it is simply used; otherwise, as shown at step 170, additional bandwidth is added if needed, by adding an additional lightpath (wavelength) for example.

The usage of the enriched topology (i.e., optical bypasses) during the network operation allows routing with a reduction in the electrical resource usage, consequently a considerable network improvement in terms of cost and scalability. At the same time, the electrical flexibility is preserved.

The total theoretical number of optical bypass routes is impractically large, unless the network is really small. For this reason, in the bypass calculation phase heuristics are used to find bypasses. The bypasses should maximize the improvement of the network topology while remaining reasonably low in number.

Figure 3, embodiment with simulated routing and provisioning

As in figure 2, step 100 shows a step of planning a network by generating a multilayer topology representation. As before, this can be for a new network, or for upgrading an existing network. Step 200 shows carrying out a simulated routing based on the multilayer topology representation, and based on a demand matrix of connections, derived from predicted or measured traffic patterns. Step 210 shows carrying out a provisioning to set the capacities or quantities of components such as electrical switches, transponders, optical switches and optical bandwidth. This can be based on the simulated routing results, or on other inputs. Subsequently the network can be operated to receive real connection requests at step 140. The demand matrix of connections, based on predicted or actual traffic patterns can also be used as an example of an indication of anticipated traffic levels on the optical bypasses or through the configurable electrical switches, to adapt the selection of the optical bypasses. This can enable the selected optical bypasses to be those which are likely to carry more traffic and thus are more useful in reducing amounts of electrical switching, or to select the optical bypasses which relieve the most heavily congested electrical switches, thus providing a more localized benefit.

Figs 4 and 5 topology before and after enhancement

Figure 4 shows an example of a representation of a topology of part of a network having nodes N1 to N6 and corresponding links between the nodes, L12 between N1 and N2, link L13 between N1 and N3, link L24 between nodes N2 and N4, link L34 between N3 and N4, link L35 between N3 and N5, link L56 between N5 and N6, and L46 between nodes N4 and N6. No optical bypasses are shown in figure 4, but an example of an enhanced topology having selected optical bypasses is shown in figure 5 for the same nodes.

The step of enriching the topology of figure 4 can have as its input:

Network topology, in terms of nodes Ni (sites) and links Lij (adjacencies between nodes Ni and Nj where fibers exist can be deployed).

The purpose of this phase is to enrich the physical topology, as received in input, with optical bypasses. Also shown in figure 5 is a further enhancement by adding well defined Shared Risk Link Groups (SRLGs) representing the fibers or other optical components such as amplifiers or filters, which are shared by the links and the optical bypasses. As input, only the physical topology is used in this example. SRLGs are associated to all the fibers of the physical topology, with SRLG Sij associated to fiber of link Lij. A bypass between nodes Ni and Nj is described by a sequence of n contiguous links Lik_1} Lk₁k_2l... Lk^.

Diversity of path for restoration is fundamental to provide a solid ground for protections and restorations. The diversity provided by the network is not changed from the input physical topology; hence physical fibers are the basis of the list of components that are subject to failures (SRLGs). An SRLG is defined for each fiber and all bypasses crossing that fiber are defined as pertaining to that SRLG. In this way, diversity schemes satisfying the correct physical topology may easily be represented.

A bypass Lik_1} Lk₁k_2l... Lk^ is then declared to pertain to all the SRLGs of the fibers it crosses, namely Siki, Sk₁k_2l... Sk .

If SRLGs that do not match with the fibers are known, fibers are replaced with the SRLG sets associated with them in the process above.

Figs 6 to 12, selection of optical bypasses A num ber of ways of im plementing the selection of bypasses can be envisaged. If all the possible routes are considered, the number of bypasses becomes rapidly unfeasible; therefore an "optimally" useful subset of all the bypasses is sought. Different heuristic strategies to select useful subsets of bypasses are discussed in this section. Heuristic is used in the sense of finding selection algorithms which prove beneficial in practice, since it is not usually practical to evaluate all the possible selections or to predict how beneficial a particular selection algorithm will be. The selection algorithms explained here can all be adapted to the previously recorded actual traffic patterns or to the forecasted traffic patterns, to select preferentially those optical bypasses which will be more heavily used, or which bypass the more heavily used electrical switches. This can be implemented by assigning a score to each of the possible optical bypasses for example, and comparing scores in making the selections, or by other techniques.

The following values can be used to give an assessment of the heuristics complexity in terms of number of bypasses to be found:

- V: number of nodes (vertices) in the graph (V=Vp+Vs)

- Vp: primary nodes

- Vs: secondary nodes

- E: number of links (edges) in the graph

- D: average meshing degree (number of outgoing edges from a vertex)

- L: average path length

- K: required redundancy (protection on K diverse paths)

A first example shown in figure 6, for reference only, is an impractical strategy of selecting all the possible routes in the network. It generates an indicative number of bypasses of order 0(D^{L χ} V²), because for each node pair (V²) at each node D alternatives have to be evaluated, and this repeats L times if the path length is L. This strategy is only theoretical and used for comparison. Figure 6 shows four different routes between non neighbouring nodes A and B.

Figs 7, 8, all pairs selection strategy Figures 7 and 8 show another example using a selection strategy of selecting at least one route between all pairs of non neighbouring nodes. Figure 7 shows steps similar to those of part of figure 2, where the selection step 120 comprises step 123 of selecting at least one route between all pairs of non neighbouring nodes. In other words, given K as the number of diverse optical paths required, K diverse optical bypasses are computed between any two nodes (relation). K is typically 1 or 2. The indicative number of generated bypasses is 0(K χ V²), as for any node pair K bypasses are computed. Figure 8 shows an example where K=2, so there are two diverse routes from node A to node B, selected from the possible four routes shown in figure 6. The choice can be made by finding the shortest routes which share no common links, or share fewest links.

Figs 9 and 10, central node and star selection strategy

In this strategy a node C in a central position is chosen as a hub, that is a node that in an ideal example minimizes its maximum distance from any other node, though less ideal examples can be useful. This choice of which node is the central node can be made automatically. The central position has the advantage of minimizing the average bypass length.

K diverse paths from each node to C are generated. This is a special case of the next strategy described below (hierarchical). It reduces the number of bypasses at the expense of creating a central electrical node with the burden of a large amount of traffic that could have been bypassed. Figure 9 shows steps similar to those of part of figure 2, where the selection step 120 comprises step 125 of selecting at least one route between the central node and every other node which is non neighbouring with the central node.

The number of bypasses is 0(K χ V), that is K bypasses from C to any other node. In Figure 10 an example is shown with a central node C and non neighbouring nodes A and B, in which K=1 . So there is one route from node A to the central node C, and one route from node B to the central node C. A bypass is not generated if there is already a fiber, since that would make the nodes effectively neighbours.

Figs 1 1 and 12, multiple central nodes and hierarchical star selection strategy

In this selection strategy there are multiple central nodes, each connected together. The V nodes are split into Vp primary nodes acting as central nodes, and Vs secondary nodes. Connect all the Vp primary nodes with K bypasses in a full mesh. Connect each secondary node with K bypasses to its closest primary. Alternatively, connect each secondary with K diverse primaries to increase node diversity.

Several ways can be devised for the choice of primary and secondary nodes. One example is to partition the network of V nodes into V V areas, and choose a central node for each area. The central nodes are primary and the others are secondary. The indicative order of bypasses is O(K χ (Vp² + Vs)).

Replacing for the above partition example, gives O(K χ ((V V)² + V - V V)), that is, it becomes linear with V, or it has the same complexity as the star, but reducing and distributing the transit traffic on the electrical nodes.

Figure 1 1 shows steps similar to those of part of figure 2, where the selection step 120 comprises step 127 of selecting multiple central nodes and selecting at least one route between any central node and every other node which is non neighbouring with the central nodes. Figure 12 shows an example topology in which a central node C and another central node C2 have been chosen, and this leaves nodes A, B, D and E as secondary nodes non neighbouring to the central nodes. A single shortest path has been chosen from E to C2, from D to C2, from A to C, and from B to C.

If the input network is complete with physical parameters of fiber and components (e.g., amplifiers, WSS, etc.), the physical feasibility of the selected bypasses can be computed. Considering that the higher cost of state-of-the-art equipment is given by electro-optical conversion, once a lightpath needs regeneration, it is usually worth adding full electrical flexibility. This is equivalent to removal of all the bypasses which are too long to carry an optical signal without regeneration. This reduces the number of bypasses, simplifying the enriched topology.

Fig 13, upgrading existing network

Figure 13 shows steps similar to those of part of figure 2. Instead of step 1 10, there is a step 1 12 which involves generating a representation of a static optical layer of optical links between neighbouring nodes having electrical switches for an existing network. This topology can then be enhanced as described above by adding the selected optical bypasses. The upgraded existing network can then be operated to receive new connection requests for routing based on the multilayer topology representation having selected optical bypasses.

Network Operation

Inputs:

- Enriched network topology, in terms of nodes, links, bypasses and

SRLG information as derived from the physical network in the above two sections

One or more connectivity requests Rk: a connectivity request includes a set of parameters, namely the node pair to be connected Ni and Nj, and the connectivity characteristics (bandwidth, protection or restoration type, etc., depending on the offered service)

In the above sections, rules have been given to define an enriched topology on which to operate the network. However, only topological information has been considered, without any calculation about the required resources.

During network operation, the actual bandwidth demands are received in the form of requests. With these, first the route calculation on the enriched topology is performed, and then the definition of the needed resources is derived. The resources are defined on both the electrical and optical layer.

In the optical layer, the links are considered one by one. If there is enough residual capacity to allocate the electrical path, its usage is simply registered. If not, an additional wavelength is added on the optical bypass; this may imply the lightpath creation. Conversely, when a path is deleted, capacity on the optical bypasses is freed. If a wavelength bypass becomes empty, it is then deleted, to free optical resources.

Wavelength Assignment

As the bypass route is calculated during the eligible bypass computation, considering the classical Routing and Wavelength Assignment problem of optical networks, only the Wavelength Assignment remains to be done. Methods for Wavelength Assignment are well known in literature, hence this phase is not described in detail other than to illustrate how it fits in the complete process.

In eligible bypass computation, optical bypasses were defined (routing) to be used in conjunction with direct links. In network operation, a number of wavelengths are required to satisfy the demand and appropriate wavelengths must be assigned to bypasses and links in such a way to

-satisfy the required capacity, and

- use the minimum number of wavelengths across the network

As noted, heuristics for Wavelength Assignment have already been proposed in literature. Here it is suggested, as an example, to use first fit. The first fit heuristics assigns the first available wavelength to the first item in the list, so that the wavelength shall not conflict with other wavelengths already assigned in the route of the item. This heuristics is simple and proven to give good results in comparison with other methods.

Fig 14, restoration

Figure 14 shows an example similar to that of part of figure 2. After step 160 of configuring the electrical switches dynamically along the route, (or possibly before this step), there is a step 180 of determining a restoration path for each route if needed. This can be carried out, according to indications in the topology of which of the optical links and bypasses share components. At step 190, in the event of failure of the working path, traffic is restored using the restoration path.

Fig 15, example of node having PCE

A node may be for example a ROADM having an input path for optical signals which can be wavelength multiplexed signals, and an output path for wavelength multiplexed optical signals. Individual wavelengths or some part of the bandwidth can be divided out optically onto optical drop paths. Individual wavelengths or some part of the bandwidth can be added from optical add paths into the wavelength multiplexed output. Converters are provided on the add paths and the drop paths, to provide conversion between electrical and optical formats. The configuration of the electrical switch can be changed dynamically based on configuration messages from the PCE, to select which of the add paths are coupled to which of the drop paths. The converters may have a fixed wavelength or in some cases, they may be tunable, to output a selectable wavelength on the optical add path, or in the case of the O/E converter, have a tunable optical filter to select which wavelength is received and converted to an electrical signal.

The node also may in some cases have circuitry for recording what configurable regeneration capacity remains available, such as how many paths, which drop wavelengths, which add wavelengths, whether they are tunable, tuning range, whether they can be switched to or from different neighboring nodes, optical parameters such as output optical powers and so on. Such information can be passed back to the PC E typical ly via messages sent to other nodes on overhead channels using existing standards. Figure 15 shows a schematic view of an example of a node (30) based on a ROADM operating in a WSON 1 .x landscape. An optical switch 20 is coupled to an electrical switch (10) such as an OTN switch via a bank of transponders 15 having O/E converters and E/O converters. The ROADM is a multi-way ROADM having bidirectional lines A, B and C carrying WDM signals to other nodes. The ROADM has an optical amplifier interface part 930 to couple the incoming and outgoing WDM signals on line A to a Wavelength Selective Switch WSS part 950. This demultiplexes the incoming wavelengths and switches them selectively to the corresponding WSSs for either line B or line C. The optical paths within the ROADM between the WSSs can be spatially separate for each wavelength or can be wavelength multiplexed together again between each WSS in principle. At the next WSS, either 920 for line B or 960 for line C, the incoming wavelengths are separated again if needed, so that a selection can be made as to which wavelengths are to be multiplexed together to go out on line B or line C respectively. Hence for each wavelength there can be two optical paths arriving at the WSS from other WSSs, and there is an optical switch com ponent or selector to select which of the two is passed to a WDM multiplexer within the WSS 920 for line B or WSS 960 for line C. This outputs a WDM signal to the bidirectional optical amplifier 910 for line B and 940 for line C. Hence each WSS has a demultiplexer and input wavelength switches for incoming paths, and has output wavelength switches followed by a wavelength multiplexer for outgoing paths.

In addition, WSS 950 is shown as having optical drop paths fed from the input wavelength switches and typically output from the WSS on optical drop paths which may be wavelength multiplexed again to reach a further wavelength demultiplexer 970. Individual wavelengths are fed to individual O/E converters of the transponder 15. These are coupled to respective electrical drop paths which are fed to electrical switch 10. This can have a client interface (not shown) for adding or dropping traffic into or out of the network. For traffic which the electrical switch is to pass to other nodes of the network, the electrical switch has outputs to electrical add paths which connect via respective E/O converters to wavelength muxer 980 of the optical switch. The corresponding optical add paths are wavelength specific if the E/O converters are not tunable. In this case the electrical switch can still configure which add wavelength is used by selecting which of the electrical add paths is used. Add multiplexer 980 multiplexes the optical add paths together to reach the WSS 960. Here they are demutiplexed again so that they can be fed as inputs to be selected or not by the output wavelength switch component within WSS 960. If selected, the optical add paths are coupled to the wavelength multiplexer within WSS 960 and become part of the outgoing WDM signal fed though optical amplifier 940 onto line C. The paths described above through optical switch 20 and electrical switch 10 have been described in one direction for clarity, but can be bi-directional. A limitation of this arrangement is that it is not colorless, there is limited selection of colors, and, the selection of which output port is used is limited to selection by the electrical switch. In another example, the E/O converters of the transponders are made tunable. This so called colorless version provides more configurability and can make it easier to add bandwidth dynamically to particular optical bypasses. The E/O converter can be made tunable by providing a tunable laser.

The node has various control parts 41 to 44 which can be part of the control plane 40 shown in figure 1 . There is a PCE 43 for routing and dynamic control of the electrical switch. This has a store 44 for the m u lti layer topology representation, typically in the form of a database. A static control part 42 is provided for controlling the optical switch. A part 41 provides dynamic control of the transponders to enable dynamic control of bandwidth of the optical bypasses for example. These parts can be implemented separately or integrated together as desired.

As can be seen, with this kind of node, the optical switch can be configured to pass any incom ing wavelength either to bypass the node or to reach the electrical switch and so the number of different possible bypass paths can be huge. Hence by selecting a subset of the possible bypass paths, the network can become much more scalable.

Control Plane and GMPLS

Path computation can be performed for example distributed -by the head NE (Ingress node) or other node- or centrally by a part of a NMS. Dynamic recovery (restoration) is usually a distributed functionality performed by the control plane.

Generalized Multiprotocol Label Switching (GMPLS) provides a control plane framework to manage arbitrary connection oriented packet or circuit switched network technologies. Two major protocol functions of GMPLS are Traffic E ngineering (TE) information synchron ization and connection management. The first function synchronizes the TE information databases of the node in the network and is implemented with either Open Shortest Path First Traffic Engineering (OSPF-TE) or Intermediate System to Intermediate System Traffic Engineering (ISIS-TE). The second function, managing the connection, is im plemented by Resource Reservation Protocol Traffic Engineering (RSVP-TE).

The Resource Reservation Protocol (RSVP) is described in IETF RFC 2205, and its extension to support Traffic Engineering driven provisioning of tunnels (RSVP-TE ) is descri bed i n I ETF RFC 3209. Relyi ng on the TE information, the GMPLS supports hop-by-hop, ingress and centralized path computation schemes. In hop-by-hop path calculation, each node determines only the next hop, according to its best knowledge. In the case of the ingress path calculation scheme, the ingress node, that is the node that requests the connection, specifies the route as well.

In a centralized path computation scheme, a function of the node requesting a connection, referred to as a Path Com putation Client (PCC) asks a Path Computation Element (PCE), to perform the path calculations, as described in IETF RFC 4655: "A Path Computation Element (PCE)-Based Architecture". In this scheme, the communication between the Path Computation Client and the Path Computation Element can be in accordance with the Path Computation Communication Protocol (PCEP), described in IETF RFC 5440. A path com putation element for routing and bandwidth assignment of planned paths and of unplanned restoration paths, may be implemented by a processor running software. A database is provided (either co-located with the processor or remotely) for storing for exam ple the topology, bandwidth availability and any regeneration capacity. This can be located remotely in principle, provided it is accessible by the processor. Configuration messages are sent from the PCE to the nodes to set up the paths in the network. The PCE may be distributed amongst the nodes, or be centralized. The nodes may return indications of availability of capacity, or the PCE may update the availability in the database as it sets up new paths. Concluding remarks:

By recognizing the split of advantages between optical and electrical network technologies, the presented network operation method can enable better efficiency in a m u lti-technology transport network: combining cost effectiveness and flexibility. Moreover, the resulting network structure is simplified, and makes use of best features of each technology. The optical part is enhanced to carry more of the traffic and to avoid com plex dynamic reconfiguration operations in response to connection requests. The electrical part implements all the flexibility, in the traditional ways these features were implemented in the past electrical networks (e.g., SDH).

From another point of view, some embodiments provide ways of operating an electro-optical network where requests in the electrical layer are automatically processed to derive the optical layer configuration, for enhancing automatically the efficiency of resource usage. This can involve adapting the selection of optical bypasses according to the predicted or simulated traffic pattern, and/or dynamically altering the bandwidth of the preconfigured optical bypasses as traffic increases.

The embodiments described are open to be extended to the incom ing "flexigrid" paradigm. In this new scenario, in which optical channels at different bit rates and bandwidth occupancy coexist on the same fiber infrastructure, the proposed optical bypasses could be adjusted as needed, with more flexibility than s i m ply chang i ng a n umber of wavelengths. In other words, the "reconfiguration" cited above could include the dynam ic modification of the bypass "dimension" to adjust its size to the evolving dimension of the served electrical traffic.

As has been described, in other words there can be: - a network of nodes with both electrical and optical capabilities, both configurable, the network being then a dual layer network (with electrical layer and optical layer) a method applicable to both network operation and network planning the method acting first on the optical layer; (during this phase the electrical layer is not yet considered), the step of selecting among all the potential optical paths the ones that are deemed to be most useful to the electrical layer then there are essentially two alternatives to go on: alternative 1 : the method uses knowledge or estimations of the electrical traffic matrix to perform capacity planning, calculate the needed capacity in optical and finalize the bypass pre-configuration on the optical layer alternative 2: there is no traffic knowledge, thus there is a step of pre-configuring bypasses with a m inimal capacity on the optical layer, then operating the electrical layer while monitoring the bypass usage. If the usage of an optical bypass crosses a predefined threshold, additional capacity is allocated on that bypass, so the optical layer is not completely pre-configured in alternative 2, but there is some adaptation or dynamicity according to the actual traffic. Other variations can be envisaged within the scope of the claims.

Claims

Claims:

1 . A method of planning a communications network having the steps of: generating a multilayer topology representation of the network, nodes of the network having configurable electrical switches, and the network having a preconfigured static optical layer of optical links between the electrical switches of neighbouring ones of the nodes, by selecting optical bypasses for the multilayer topology representation between non neighbouring ones of the nodes, via others of the nodes so as to bypass the configurable electrical switch of at least one of these other nodes , for a subset of all the possible routes between all the pairs of non neighbouring nodes, and adding the selected optical bypasses to the static optical layer of the multilayer topology representation.

2. The method of planning, having the further step of using the multilayer topology representation for determining capacities of the electrical switches and the optical links and optical bypasses, to provision the network.

3. The method of claim 1 or 2, having the step of adding to the topology representation indications of which of the optical links and optical bypasses share components with a risk of failure.

4. The method of any of the preceding claims, the selecting step comprising selecting at least one route between all the pairs of non neighbouring nodes.

5. The method of any of claims 1 to 3, the selecting step comprising selecting one of the nodes as a central node and selecting a limited number of routes sufficient to connect the central node to all other nodes which are not neighbours of the central node.

6. The method of any of claims 1 to 3, the step of selecting the optical bypasses comprising selecting two or more nodes as primary nodes, providing optical bypasses to connect every different pair of the primary nodes and providing a limited number of optical bypasses sufficient to connect all other nodes which are not neighbours of the primary nodes to any one of the primary nodes.

7. The method of any of claims 1 to 6, the step of selecting the optical bypasses comprises selecting according to an indication of anticipated traffic levels on the optical bypasses or through the configurable electrical switches.

8. The method of any of claims 1 to 7 further comprising using the topology representation, and a demand matrix of simulated connection requests, based on anticipated connection requests or on records of actual connection requests, to carry out a simulated routing of the simulated connection requests.

9. The method of claim 8 having the step of altering a bandwidth allocation of the optical links or optical bypasses according to the simulated connection requests.

10. The method of any of claims 1 to 9, for reconfiguring an existing network and having the prelim inary step of generating a multilayer topology representation of the optical links between neighbouring nodes of the existing network.

1 1 . A method of planning and operating a network, having the steps of planning as set out in the method of any of claims 1 to 10, and further comprising the subsequent steps of operating the network by routing new connection requests through the network according to the multilayer topology representation, and configuring the electrical switches along the selected route to set up the connection requests.

12. A method of operating a com m un ications network having nodes, the method having the step of: routing new connection requests through the network according to a multi layer topology representation of the network, the nodes havi ng configurable electrical switches, and the network having a preconfigured static optical layer of optical links between the electrical switches of neighbouring ones of the nodes, and the multilayer topology representation having optical bypasses between non neighbouring ones of the nodes, via others of the nodes so as to bypass the configurable electrical switch of at least one of these other nodes, the optical bypasses comprising a subset of all the possible routes between all the pairs of non neighbouring nodes, and the method having the further step of; configuring the electrical switches along the selected route to set up the requested connection.

13. The method of claim 1 1 or 12, having the step of altering a bandwidth allocation of the optical links or optical bypasses according to the connection requests.

14. The method of any of claims 1 1 to 13, the topology representation having indications of which of the optical links and optical bypasses share components with a risk of failure, and the method having the step of determining a restoration path for a connection according to the indications of shared components.

15. A computer program having instructions on a computer readable media which when executed by a processor, cause the processor to carry out a method of planning as set out in any of claims 1 to 1 1 , or a method of operating a network as set out in any of claims 12 to 14.

16. A path com putation elem ent for a node of a communications network, the path computation element having a store for storing a multilayer topology representation of the network, and the path computation element being arranged to carry out the method of operating a communications network as set out in any of claims 12 to 14.