JP4028557B2 - Optical network, node device, and optical cut-through link setting method - Google Patents

Description

  The present invention relates to an optical network and node device using an optical cross-connect, and an optical cut-through link setting method in the optical network.

  Internet technology advances rapidly. The traffic flowing through the communication network was mainly fixed telephone traffic in the early 1990s, but now most of it is data traffic flowing through the Internet. With the explosive spread of the Internet and broadband, traffic on the metro and trunk lines that support the Internet is also increasing rapidly.

  Various techniques have been studied in which a photonic network using an optical cross-connect as a node device is superimposed on an existing network, and a portion where traffic has increased is transferred to the photonic network. Regarding the transfer method, there are various methods, from a method of establishing a fixed optical path to the photonic network to a method of using it temporarily.

  Non-Patent Document 1 shows an example in which an optical path is temporarily used, and it is possible to reduce the initial investment in the portion of the photonic network as compared with the method of fixedly assigning. The outline will be described below.

  FIG. 9 is a conceptual diagram of a network described in Non-Patent Document 1. Each of the node devices 103-1 to 10-3 includes label switch routers 101-1 to 101-5 and photonic cross-connects (PXC) 102-1 to 102-5. The node devices are appropriately connected by optical transmission lines 104-1 to 104-6. Each optical transmission line includes a plurality of wavelength-multiplexed optical links. One of the wavelengths λ0 is set as a default wavelength, and the node devices are connected hop-by-hop. That is, λ0 is always subjected to photoelectric conversion in each node device, and electrical processing of the signal is performed. λ0 is used for transmission of network control signals and L2 / L3 layer packets. The packet passing through the L2 / L3 layer is switched by the label switch router (LSR) of each node device that has passed through until reaching the destination. In the figure, the dotted arrows connecting the LSRs in the plane indicated as the L2 / L3 layer indicate that the node devices are logically mesh-connected. The logical route that connects each node device to the mesh is called a label switched path (LSP). The mesh connection is only logical, and in practice, packets passing through this network proceed through the LSP while being switched hop-by-hop by LSR. The LSR has a function of measuring traffic of an LSP passing through its own node device. When each node device detects an increase in traffic at any LSP, it causes the traffic of that LSP to be cut through at the L1 layer. Specifically, an L1 layer optical path via PXC is set along the route through which the LSP passes, and the LSP traffic is shifted to the optical path.

  The PXC is a cross-connect having a function of connecting light of an arbitrary wavelength input to an arbitrary input port to an arbitrary output port without converting the light into an electrical signal and without performing wavelength conversion. Therefore, the optical path starts from an ingress (starting point) node device, passes through several PXCs without being converted into an electrical signal at the same wavelength, and reaches an egress (end point) node device.

For example, in the figure, when it is desired to cut through the LSP that connects LSRs 101-1 and 101-3 in the L1 layer, it is assumed that the LSP actually passes through LSRs 101-1, 101-2, and 101-3. When an optical path setting request is made to the node devices 103-1, 103-2, and 103-3, and the conditions such as the wavelength are satisfied and the request is accepted, the light passing through the PXCs 102-1, 102-2, and 102-3 A path 105 is set. Note that the node device that generates the optical path setting request is either the node device 103-1 or 103-2. When the optical path 105 is successfully set up, the node device 103-1 directs the LSP traffic for the LSR 101-3 input to the LSR 101-1, to the PXC 102-1, and to the node device 103-3 by the optical path 105. To reach. The node device 103-3 directs the traffic of the optical path 105 reaching the PXC 102-3 to the LSR 101-3. In this way, the node device 103-2 is cut through, and the packet processing load of the LSR 101-2 is reduced.

  Non-Patent Document 1 aims to reduce the cost of the system by limiting the number of optical transceivers provided in each node device. At the same time, in order to secure an opportunity to establish a fair path with all the other node devices with a number of optical transceivers that is less than the number of mesh optical paths, the time from when the optical path is established until it is released Time (light path survival time) is limited in seconds. Also, the optical path is released even when the traffic in the optical path decreases and the meaning of occupying the optical path is lost. For this reason, optical paths are frequently set and released in units of seconds during network operation.

  In the network in Non-Patent Document 1 shown in FIG. 9, one optical path is extended to one LSP that detects an increase in traffic. While the default wavelength is shared by many LSPs, the optical path is monopolized by one LSP. The purpose is to accommodate traffic deviation of default wavelength by the optical path and avoid congestion at default wavelength, so that if each traffic is all average value, no optical path is needed A default wavelength is designed. A network using such a photonic cross-connect is a network close to a trunk line having a high bit rate, and traffic flowing through the network is statistically multiplexed with traffic from a number of lower network terminals. Although an optical path is established for an LSP with increased traffic, even if the traffic of a specific LSP increases in a short period of seconds, it rarely increases in number, such as several times the average value. If the optical path setting threshold is set high, a large amount of traffic flows in the optical path while the optical path is set, but the probability of setting the optical path is significantly reduced, and the temporal usage rate of the optical path Becomes lower. On the other hand, if the optical path setting threshold is lowered, the probability that an optical path is set increases and the temporal optical path usage rate increases, but traffic that flows in the optical path when the optical path is set. Becomes an amount close to the average traffic of the original LSP, and the bandwidth utilization of the optical path is remarkably reduced. In any case, the usage rate of the optical path is lowered. We are aiming to reduce the cost of the system by reducing the number of expensive optical transceivers, but because the usage rate of the optical path is low, the optical transceiver for the optical path has only the meaning of an extra of the default wavelength. . After all, the throughput of the network is not commensurate with the number of optical transceivers used, and the cost performance is poor.

  In Patent Document 1, the node device requests setting of a wavelength path from the start node device to the end node device via the relay node device, and the relay node device converts a specific optical wavelength to be converted in response to this request. The start-point node device sets a specific optical wavelength to be inserted, requests the release of the set wavelength path when a predetermined time is reached after completion of the wavelength path setting, and responds to this request by the end-point node device Is a technology that enables large-capacity transmission by releasing a specific optical wavelength that is branched, releasing a specific optical wavelength converted by the relay node device, and releasing a specific optical wavelength inserted by the start node device It is disclosed. This technology enables large-capacity transmission by operating wavelengths efficiently.

However, the technique disclosed in Patent Document 1 does not necessarily reveal the timing conditions for making a wavelength path setting request. For example, in a non-linear network such as a mesh network, the wavelength path must be set and released correctly. The problem that cannot be done remains.
JP 2002-16950 A IEICE 2003 Society Conference B-6-140

  The present invention has been made in view of the above-described conventional problems, and an object of the present invention is to more efficiently use a limited number of optical transceivers, so that an optical network having a high throughput as a whole and its optical The object is to provide a node device for realizing a network.

  In order to solve the above problem, an optical network according to the first aspect of the present invention includes a packet switch router that switches a packet forwarded on a default link connecting adjacent node devices. An optical network in which an optical cut-through link is introduced into a communication network, wherein the optical cut-through link is smaller in number than all other node devices in the packet communication network included in each node device serving as an end point thereof. Formed by an optical cross-connect for an optical cut-through link and an optical cross-connect that defines the path of each node device cut-through by the optical cut-through link, and each node device includes an entry including a metric value Based on the routing table that describes The output destination of the packet is determined according to a protocol, and the start node device of the optical cut-through link sets the value of the metric changed by the setting of the optical cut-through link after the setting of the optical cut-through link to the adjacent node device and the optical Traffic measurement to advertise to the end node device connected by the cut-through link, and each node device measures the packet traffic switched by the packet switch router and the traffic of the optical cut-through link having its own node device as the end point And at least one of the cut-through links with a small traffic value of the optical cut-through link measured by the traffic measuring means among the optical cut-through links starting from the own node device Free The new end node device determined based on the value of the packet traffic measured by the own node device and other node devices is selected, and an optical cut-through link is newly established for the selected end node device. Perform signaling for configuration.

  That is, the optical route connecting the node devices is used as an optical link instead of an optical path. In the node device, adjacent node devices are connected by a default link. The default link can be realized by an optical transmission technique using a default wavelength, but is not necessarily limited to an optical transmission technique. For example, the default link may be realized by a telecommunication technology including a wireless communication technology.

  In addition, each node device has an optical transceiver for forming the end point of the optical cut-through link, but the number is limited, and the optical cut-through link of the mesh is formed simultaneously with all the node devices in the network. I can't do it. An optical cut-through link is an optical route that starts from a start node device, passes through an optical cross-connect at an intermediate node device, and ends at an end node device. This is an optical link connecting node devices. In the network of the present application, an autonomous distributed routing protocol operates in each node device, and the route of the packet and, in some cases, the route of the LSP is determined based on the convergence value of the route information distributed by the routing protocol. . Each node device advertises route information owned by each node device to an adjacent node device periodically or when the route information is changed. One of the features of the present invention is that when an optical cut-through link is provided, the start node device regards the optical cut-through link as a new link that directly connects the start node device and the end node device. The device route information is reconstructed and advertised to adjacent node devices. Since the optical cut-through link is merely a link, the end node device connected by the optical cut-through link is also treated as an adjacent node device, and the route information is also advertised to the end node device via the optical cut-through link.

  The contents to be included in the route information require a metric value in addition to the connection destination (end point node device). What is the metric value depends on the routing protocol. In the present application, for example, the metric value of the optical cut-through link is made smaller (identified as a close or thick easy-to-use link) than the value when all the destination node devices are assumed to have passed through the default link. Therefore, the optical cut-through link is identified as existence like a highway connecting the start node device and the end node device. As a result, the route of traffic that reaches from the node device in the vicinity of the start node device to the node device in the vicinity of the end node device is easily determined so as to pass through the optical cut-through link.

  Since each node device has only a limited number of optical transceivers for the optical cut-through link, it is necessary to select a counterpart node device that forms the optical cut-through link. Packets traveling on the default link are switched in a store-and-forward manner by the packet switch router of the node device on the way. Reduce the average number of times a packet switch router processes a packet by arriving at its destination by passing as much traffic as possible through the formed optical cut-through link and reducing traffic through the default link. It becomes possible. As a result, it is possible to reduce the processing load on the router, and it is possible to improve packet quality by reducing packet delay and jitter. Therefore, it is desirable that the counterpart node apparatus forming the optical cut-through link has a traffic as much as possible flowing through the optical cut-through link and is far away to some extent.

  In the present invention, almost all or almost all of the optical transmission / reception of each node device is always used, and an optical cut-through link that can be stretched is stretched. In order to select a partner node device that can establish an optical cut-through link from its own node device, a partner node device that is desirable at that time, i.e., a partner node device in which as much traffic as possible flows through the optical cut-through link. In the present invention, one or more of the optical cut-through links starting from the own node device periodically is released from the side with the least traffic, and the partner node device different from the partner node device of the released optical cut-through link Signaling is performed so as to create a new cut-through link (reconfiguration of the optical cut-through link).

  In the present invention, after the optical cut-through link is formed, it is advertised as one of the links. The neighboring node device that has received this examines the metric value of the optical cut-through link, and if the total of the metric value up to the arrival destination becomes smaller through the metric value, it passes through the optical cut-through link. Rewrite the routing table entry to be a route. The rewritten entry is advertised to nearby node devices as its own route information. In this way, setting the optical cut-through link may rewrite the routing table of the entire optical network. In the optical network of the present invention, since each node device operates by an autonomous distributed routing protocol, it cannot be predicted how other node devices rewrite the routing table. After all, after establishing a new optical cut-through link, it is necessary to predict in advance how much traffic will flow through the optical cut-through link until the routing table of all node devices is converged by the advertised route information. Is almost impossible.

  In the present invention, when determining the partner node device to which the new optical cut-through link is connected, the partner node device is selected based on the measured value at the packet switch router of the traffic to the partner node device at that time. However, there is a high possibility that the traffic after the optical cut-through path is extended is significantly different from the traffic measured at this time. Therefore, the traffic measured at this time is only a reference for selecting a counterpart node device, and is merely an expectation that a certain amount of traffic may flow. Therefore, it is not necessary that the traffic measured by the packet switch router to the newly selected partner node device is larger than the traffic of the optical cut-through link to be released. That is, the threshold for releasing and the threshold for selecting the partner node device of the new optical cut-through link are different, and the latter may be lower.

  In this way, the traffic of the optical cut-through link cannot be understood unless it is stretched. Each node device has a function of measuring the traffic of an optical cut-through link whose end point is its own node device stretched at that time. The traffic measured on the optical cut-through link is reliable as the link performance. Therefore, by periodically releasing an optical cut-through link with low traffic and trying to establish a new link that is unknown but can be expected to have some traffic, how much is actually in the optical cut-through link? To see if traffic flows. If a large amount of traffic flows on the new link, the optical cut-through link will not be released for the next optical cut-through link reconfiguration, and if there is not much traffic, it will be listed for release. Eventually, it may be freed. By periodically reconfiguring such an optical cut-through link, the result is that an optical cut-through link with high traffic is selected and survives forever. In other words, it is possible to effectively use a limited number of optical transceivers and always keep an optical cut-through link preferentially at a place where traffic is high.

After the optical cut-through link is released and newly spread and the change is advertised to the adjacent node device, it takes time in minutes until the result converges and is reflected in the entire network. If route information changes are advertised too frequently, the route recalculation at the node equipment will be frequent, and in some cases the routing may be disrupted due to the next change before convergence. There is. Therefore, it is desirable that the reconfiguration period of the optical cut-through link is sufficiently longer than the time taken to converge after advertising. The overall traffic in the network changes relatively slowly because traffic from many terminals in the lower network is statistically multiplexed, particularly in a network close to a trunk line using an optical cross-connect. That is, it changes greatly in time, day, and even in loose units such as the month and holiday season, such that the trend is different between day and night, and the trend is different between weekdays and holidays. By using the system of the present application, it becomes possible to cope with such a long-term change in traffic trend, and a finite number of optical transceivers are allocated to the most necessary parts at each time, and the network shape Can be adapted to change in traffic trends.

  The optical network according to the second aspect of the present invention performs the signaling for releasing the optical cut-through link and performing a new setting at substantially the same time in the optical network.

  That is, in the invention according to claim 1, each node device periodically releases the optical cut-through link and performs signaling for new setting. In the invention according to claim 2, the timing for performing the signaling is as follows: It is performed at substantially the same time in a plurality of node devices or all node devices in the optical network.

  In the invention according to claim 1, route information that has been changed due to release or new setting is advertised to the neighboring node device. The route information is changed one after another within the affected range. Packets for the advertisement fly and converge after some time. Each node device repeats the release / new setting in a cycle longer than the time required for convergence. When the timing of each node device release and new setting is different, the advertisement packet fluctuates randomly in the network, and before the effect of one change converges, another change is notified, Route information may not converge. Accordingly, in the invention according to claim 2, such release / new setting is performed at substantially the same time in a node apparatus within a certain range. “Substantially the same time” is defined as the time when each node device reconfigures the optical cut-through link so that the time when the routing table of each node device in the optical network does not converge is acceptable for operation. Because.

  As a result, route information changes occur almost simultaneously and are advertised. It may happen that a plurality of route information change advertisements are received almost simultaneously. Also, several advertisements for changing route information are received with a very short time difference. Each node device can recalculate the route by combining several advertisement contents, or even recalculate the route for each one and advertise the result The time until convergence in the network does not change in order from the case where the route is changed at only one place. Depending on the size of the network, it will probably converge in a matter of minutes.

  The range of node devices to be changed synchronously is preferably all node devices in the optical network if possible. If an optical network operating with the same protocol is very wide, it is usually partitioned into several autonomous systems to facilitate operation. In such a case, it is usually sufficient to synchronize in units of autonomous systems because routers that limit the boundaries of the autonomous system are limited and route information is filtered inside and outside the autonomous system. In this way, it is desirable to release and set the optical cut-through link in synchronization with one unit of network.

  Furthermore, in the present invention, when a new optical cut-through link is set, a partner node device different from the optical cut-through link that is simultaneously released is selected. Therefore, the different counterpart node device must have an optical transceiver for creating a new end point of the optical cut-through link. In the optical network of the present invention, each node device always uses the finite number of optical transceivers of the device almost fully. Therefore, unless the node device as a new partner also releases the optical cut-through link at the same time, the probability of successful signaling of the new optical cut-through link is significantly reduced. From this point of view, it is desirable to release and set the optical cut-through link in a certain manner in the network.

  By doing so, the route information in the network can be changed smoothly, and the time during which the network is confused can be shortened. Furthermore, it becomes easy to determine a new connection partner when reconfiguring the optical cut-through link.

  Next, in the optical network according to the third aspect of the present invention, the number of optical transceivers for optical cut-through links is determined corresponding to the number of optical cut-through link endpoint formation requests for each node device.

  In a large-scale network, the traffic from any node device in the network to any other node device is greatly different depending on the combination, that is, the traffic is often biased. The number and scale of lower-level networks and external networks connected to a node device are different for each node device, and traffic is concentrated on a specific node device or traffic between specific node devices is often larger than others.

  In the present invention, each node device has a limited number of optical transceivers, but the number is a value suitable for each node device and does not have to be the same. In other words, a node device that tends to concentrate traffic has more optical transceivers, a node device that does not have much traffic, or a node that does not have a large number of nodes that always have the same connection destination where traffic increases. A device, a node device that does not have many connections to which traffic increases simultaneously, has a small number of optical transceivers. That is, a node device that is desired to be an end point of many optical cut-through links includes many optical transceivers. As signaling for establishing an optical cut-through link, one of the node devices including its own node device is requested to be the start point of the optical cut-through link, and the node device requested to be the start point is the end point. An optical cut-through link formation request is issued to the node device. What is necessary is just to determine the number of optical transmitters / receivers corresponding to the frequency with which it is desired to be the start point or the end point.

  By doing so, each node device can be provided with the minimum number of optical transceivers, and the cost of the network can be reduced. In addition, traffic can be accommodated in the optical cut-through link according to the trend in the network, and the transmission quality of many traffics can be kept high.

  An optical network according to a fourth aspect of the present invention is the optical network according to the second aspect, wherein the packet switch router is a label switch router that performs multi-protocol label switching, and the traffic measurement means is a label switch for measuring packet traffic. A new optical cut-through link is set corresponding to one of the label switched paths, and the ingress node device of each label switched path has a metric value after the setting of the new optical cut-through link. The route of the label switched path is changed when the metric value of the routing table that has been advertised and converged after the routing table has converged is improved from the value before setting a new optical cut-through link.

  Each node device has a packet switch router that switches traffic on a packet basis and an optical cross-connect for circuit-switching the optical cut-through link. In the fourth invention of this application, the packet switch router is a label switch router (LSR) corresponding to MPLS (multiprotocol label switching). Therefore, the data packet flows through the label switched path (LSP) set by MPLS. The traffic measurement at the router of each node device is performed in LSP units. At least a mesh-like LSP is extended so that a packet can be transmitted from any node device to any other node device. The LSP is a tunnel, and its start point (ingress) node device and end point (egress) node device are clear, and its route is fixed. When measuring traffic that is used as a reference when setting up a new optical cut-through link, it is possible to measure the traffic with the start and end points clearly defined. It is very easy to determine a route for a new optical cut-through link before starting signaling. That is, an optical cut-through link is extended along the LSP route in units of any LSP ingress node device to egress node device.

  It should be noted that the LSP must be properly established before the packet is sent, and when changing the route via the LSP, a new route must be established before the old route is demolished or suspended. In order to establish a route and establish an LSP, it is necessary to perform separate signaling. In the present invention, the light route is merely a link, not a path. It is like a link connecting the start and end points of an optical cut-through link that can be made in time or disappear. Even if a link is made and the route information is advertised, the LSP does not automatically change the route of the LSP unless the transit link disappears. Therefore, in the invention according to claim 4, in response to the change of route information, when changing to a route passing through an optical cut-through link, or changing to a route passing through a different optical cut-through link, Based on the changed route information, signaling for establishing an LSP is performed again.

  By doing so, it is possible to realize an optical network that facilitates route determination using LSP.

  An optical network according to a fifth aspect of the present invention is the optical network according to the fourth aspect of the present invention, wherein the node device stores a route in the latest past in which a label switched path is set only via a default link. The ingress node device that sets a new optical cut-through link has a new optical corresponding to the route in the latest past set only via the default link stored in the storage means. Signaling is performed to set up a cut-through link.

  When the optical network according to the present invention is performed on an MPLS basis, a part of, and in some cases, all of the LSP transit links may pass through the optical cut-through link. The newly formed optical cut-through link is extended along the route of any LSP. At this time, a part of the LSP may already pass through the existing optical cut-through link. In such a case, it is naturally impossible to pass the new optical cut-through link stretched corresponding to the LSP into the existing optical cut-through link through which it passes. Also, in the case of a route that partially passes through an existing optical cut-through link, if the same route is traced only with the default link, that route is the most desirable route (with a small metric value) for that LSP. Not necessarily. It goes without saying that even with an optical cut-through link, it is desirable to have as few node devices as possible.

  Any LSP must have followed a route that passes only the default link before a part of the LSP is set to pass through the existing optical cut-through link. Remember the last (last past) route. Since the optical cut-through link cannot pass through another optical cut-through link, the route formed only by the default link is the shortest (smallest metric value) route for the new optical cut-through link. As for the storage of the route, the ingress node device stores all routes together, or the transit node device of the LSP stores the previous and subsequent node devices with respect to the LSP. And, if the LSP that is a new optical cut-through link candidate is partly passing through an existing optical cut-through link, the control packet for signaling the optical cut-through link is only the default link. Interact through. In this way, the new optical cut-through link is set along the route when the LSP is only via the default link, that is, the shortest route.

  An optical network according to a sixth aspect of the present invention is the optical network according to the second aspect of the present invention, wherein the node device aggregates packet traffic measurements in the routing table of each node device by means of packet traffic measurement means possessed by the node device. This is performed for each entry, and before releasing the optical cut-through link, route determination signaling is performed to determine the start point, end point, and route candidate of the newly set optical cut-through link.

  That is, in the invention according to claim 6, instead of the LSP of claim 4, a unit of entries aggregated in the routing table of each node device is used. Routing is performed on the basis of an ordinary IP (Internet Protocol). If the unit of entries aggregated based on the normal routing protocol is determined to have a lot of traffic as a result of traffic measurement, it becomes a new optical cut-through link candidate. At this time, for the following reason, a step of performing route determination signaling for determining the start point, the end point, and the transit node device of the new optical cut-through link candidate is necessary.

  That is, in a general IP router, the aggregated entries described in the routing table include an aggregated entry of destination addresses, a subnet mask as a unit of aggregation, and an output interface corresponding to the entry (or its interface) Only information about the level of the adjacent node device to which is connected. The router looks at the destination information described in the arrived packet, searches for an entry that matches the destination information, and outputs the packet to the corresponding output interface. If the routing table of each node device is properly prepared, packets are sequentially sent to the destination by that method. However, when a new optical cut-through link is set in this application, the start node device, the end node device, and the transit node device, that is, the route must be clearly determined. I have no idea where it starts and where it ends. Aggregated entries vary, and the address of the entry may be a network outside the optical network. For this reason, when measuring traffic in units of aggregated entries, a route determination signaling step for determining the route of the optical cut-through link in advance is performed.

  In this way, unlike the case of creating a network only by LSP, the optical network of the present application can be constructed based on an IP network by a more general IP router, so that the compatibility with the existing network is good. Easy to upgrade.

  The router may be an IP router that simultaneously supports MPLS. In the case of MPLS, the aggregated entry is a unit of MPLS label. If LSP is selected from the traffic measurement result, route determination signaling for setting an optical cut-through link is not required.

  An optical network according to a seventh aspect of the present invention is the optical network according to the sixth aspect of the invention, wherein the node device belongs to the aggregated entry in the routing table to the own node device in the route to the destination related to the entry. Information on the presence / absence of passage of any optical cut-through link in the optical network is stored, and information on the presence / absence of passage of the optical cut-through link stored for each entry in the route information advertisement based on the routing table If the information on the presence or absence of the passage of the optical cut-through link of any entry changes from none to existence due to the route recalculation by the advertisement of the route information received from other node devices, the latest past related to the entry Default link output interface Storing the scan or connected adjacent node device, and sends a control packet for use in route determination signaled to the output interface or connected adjacent node device of the stored most recent previous default link.

  As in the case of the above-mentioned LSP, the route of the new optical cut-through link cannot pass through the existing optical cut-through link. Therefore, the route with the smallest metric value (the shortest) is established when only the default link is constructed. It is desirable to be determined along When route determination signaling is performed, various control packets are transmitted and received on a route that can be a route, but in order to create a route for a new optical cut-through link candidate via only the default link, Send and receive via link only. Therefore, each node device stores a plurality of output interfaces as necessary for each aggregated entry. That is, when the output interface for the data packet described in the entry is an optical cut-through link, the output destination of the most recent default link is also stored for the entry. That is, when the route information is changed with respect to the entry and the output interface is changed to the optical cut-through link, the output interface information of the default link used so far is also left without being erased. Of course, the routing of the data packet related to the entry is output to the interface of the optical cut-through link after the change, but when the packet used for route determination signaling is for the destination described in the entry, the default link described at the same time is used. Output to the output interface. In addition, when advertising the route information, the local node device belongs to the destination before the packet corresponding to each entry arrives at the destination as well as the local node device is the start point of the optical cut-through link. When passing through any one of the optical cut-through links in the optical network, information on the presence or absence of the passage is included and sent for each entry. The routing table of each node device has an area for storing the presence / absence of optical cut-through link passage for each entry. Of course, when the own node device is the starting point, the route information from other If it is notified that the optical cut-through link is passed within the optical network, the fact that it passes is stored. Then, when the entry passage presence / absence column changes from none to yes, the information of the output interface (or connection destination node device) of the entry immediately before the change is stored without being erased. In the switching of the data packet related to the entry, it is naturally output to the output interface passing through the optical cut-through link. The control packet used for route determination signaling is stored at the same time as the entry, if the latest entry corresponding to the destination does not contain an optical cut-through link. To the output interface (connection destination adjacent node device) that passes only the nearest default link.

  In this way, a new optical cut-through link can be set with the shortest path.

  An optical network according to an eighth aspect of the present invention is the optical network according to the sixth aspect of the present invention, wherein in the route determination signaling, each node device is grouped in units of entries in the routing table of the node device by the packet traffic measurement means. Based on the measurement result of the packet traffic performed, the entry having the traffic larger than the threshold is selected, and the address of the route formation request source node device to the connection destination adjacent node device of the default link of this entry, the network address of the entry, and the subnet mask The node device that has sent the route formation request including the measurement result and received the route formation request is the node device that the connection destination adjacent node device corresponding to the entry included in the route formation request has received the route formation request. If it is a node device in the optical network to which the device belongs, the default link corresponding to the network address and subnet mask of the entry included in the received route formation request and the routing table of the node device that has received the route formation request The network address and subnet mask of the entry of the node device that has received the route formation request are compared, and the network address and subnet mask of the corresponding entry included in the route formation request are wider than the network address and subnet mask of the entry included in the route formation request. Or the same, or an adjacent node with the same connection destination of the output interface of the default link described in which there are a plurality of corresponding entries of the node device that has received the route formation request. If it is a node device, the address of the node device that has received this request is added to the route formation request as a transit node device address, and relayed and transmitted to the adjacent node device to which the default link of the entry is connected. Whether there are multiple entries of the node device that has received the route formation request corresponding to the network address and subnet mask of the entry included in the request, and whether the connection destination of the output interface of the described default link is a different adjacent node device A node outside the optical network to which the node device to which the connection destination adjacent node device corresponding to the entry included in the route formation request has received the route formation request has a narrow range of the network address and subnet mask of the corresponding entry When it is a device The node device that has received the route formation request is recognized as the end point of the route of the new optical cut-through link candidate.

  Route determination signaling has significant challenges. Normally, each node device aggregates route information and collects route information so that the number of entries is as small as possible. When aggregating IP address entries, it is generally aggregated with a range of upper bits that can be aggregated (subnet mask) and addresses within that range. Even if the algorithm for summarizing route information is the same for all node devices, the way in which entries are aggregated due to slight differences in route information notified to each node device, the position of each node device in the network, etc. May be different.

  In the present application, the traffic is measured for each entry in each node device. Route determination signaling is performed prior to optical cut-through reconfiguration for entries that are heavily trafficked and that are deemed desirable to transition to the optical cut-through link. Because there are various possibilities for each node device for the same address, the entries that are aggregated for the same address can be aggregated in the same range if you try to determine the route with the entries in the same address and range. There is a possibility that the optical cut-through link can be extended only within the range of the neighboring node devices, and the number of via node devices may be remarkably reduced. As described above, the optical cut-through link of the present application is intended to be a highway-like link that allows packet devices to pass between node devices that are somewhat apart from each other without being subjected to packet switching. If the number cannot be secured, the effect is diminished.

  In the route determination signaling according to the present invention, a route formation request packet is first generated for an entry for which each node device determines that there is a lot of traffic. This includes the address of the own node device, the aggregated network address of the entry, the subnet mask of the aggregated network address, and the traffic measurement result, and the connection destination adjacent node described in the corresponding entry in the routing table. Send a packet to the device (to the interface to which it is connected). The adjacent node device that has received the route formation request finds an entry in the routing table of its own node device corresponding to the network address and subnet mask included therein, and performs comparison. Wide range in which the corresponding entry of the local node device is the same as the entry described in the route formation request and the network address and subnet mask, or the entry of the local node device completely includes the network address described in the route formation request If there is more than one corresponding entry in the routing table of the own node device, but the connected destination node device described there is the same, the route formation request is relayed. To do. The destination is the connection destination adjacent node device described in the corresponding entry of the own node device. When relaying, in order to record the route, the address of the own node device is added as a transit node device.

  As a result of comparison, if the entry range of the local node device is narrower or the range to be aggregated is slightly different, the destination range specified by the network address and subnet mask described in the route formation request is In the routing table of the device, when the connection destination adjacent node device is different corresponding to a plurality of entries, or the connection destination of the corresponding local node device entry is outside the optical network to which the node device belongs, If the optical cut-through link that passes through the node device in that direction cannot be established, the node device recognizes the node device as the node device at the end point with respect to the route formation request.

  In this way, it is determined that a certain node device has a lot of traffic, and for the created route formation request, the farthest node device including the network address and subnet mask range of the entry of the node device may be the end point. it can. When the range is small or the range is divided, the end point is that the traffic measurement is in units of entries, and what is the address that mainly contributes to the traffic increase in the smaller range included in the entry? This is because when the range is divided, it is not known where the optical cut-through link can be connected to ensure that traffic can be cut-through. When measuring traffic, this situation can be prevented by measuring traffic for each destination IP address, not for each aggregated entry. However, measuring traffic for each IP address places a heavy burden on the node equipment. Pass.

  If the local node device is at the end of the network, an optical cut-through link cannot be established before that. Here, the entry to be compared is that of the default link and does not include the optical cut-through link. As described in the seventh aspect, route determination signaling is performed along a route that does not pass through the optical cut-through link.

  By doing so, the end point of the new optical cut-through link can be identified.

  The optical network according to the ninth aspect of the present invention is the optical network according to the eighth aspect of the present invention, wherein a node device that receives a route formation request in route determination signaling and recognizes its own node device as an end point is the first as an end point. Among a plurality of route formation requests that arrive at the node device within a predetermined time from the time of recognition, and that the route that has passed through the route formation request is in an inclusive relationship, and that leads to the result that the node device is recognized as the end point. Select one or more of Then, the request source node device that has generated the route formation request selected by the node device that has recognized its own node device as the end point is selected as the start node device of the route of the new optical cut-through link candidate and is selected. The transit node device added to the route formation request sent from the starting node device is selected as the node device through which the new optical cut-through link candidate route passes.

  When the end point is determined by signaling using the route formation request packet, the node device that has identified its own node device as the end point waits for a similar route formation request for a certain period of time. When traffic from one subordinate network to another subordinate network is increasing, a plurality of transit node devices of that traffic become the request source, and many similar route formation requests reach the end node device. Because there is a possibility. While the end node device is waiting, if only one route formation request has arrived at its own node device, the request source node device described therein is determined as the start node device. Even if multiple route formation requests arrive, if the route consisting of the source node device and the transit node device described in the route formation request is not inclusive of each other, that is, different routes, The request source node device described in the route formation request is the start node device. However, if the routes described in multiple route formation requests are inclusive of each other, a comparison is made according to the contents (the amount of traffic, the number of via node devices), etc. One of them, or two or more as necessary, are selected as start node devices, and a corresponding via node device is determined.

  In this way, it is possible to select the start node device and its transit node device without causing unnecessary duplication for the same traffic.

  An optical network according to a tenth aspect of the present invention is the optical network according to the second aspect of the present invention, wherein route information advertised to an external network connected to the optical network is periodically synchronized within the optical network. Even if the optical cut-through link is released and reconfigured by a new setting is performed a predetermined number of times or more, the information is limited to information on the optical cut-through link that is not released and is not changed.

  That is, the invention according to claim 10 relates to how to advertise internal route information outside the optical network of the present application. In the present application, each node device, particularly a boundary node device that is a connection point with an external network, has route information for the inside of the network and route information for the outside. The inside and outside of the optical network referred to here is inside and outside of the optical network of the present application if it is a single network, and inside and outside of the autonomous system if it is divided into autonomous systems. The optical network of the present application is characterized in that an optical cut-through link with low traffic is periodically released and a new optical cut-through link is provided, so that the optical cut-through link is always supplied to a place where the traffic is large. Depending on the period, the contents of the routing table frequently change greatly in units of several minutes to several tens of minutes. Usually, only rough route information aggregated in units of autonomous systems is advertised outside the autonomous system, and detailed internal route information is often not advertised. However, for example, a metric from a border node device of one optical network to another border node device in the same optical network is advertised to the outside as information for connecting a plurality of networks. When the metric value of such a route changes due to the optical cut-through link, it is usually necessary to advertise the change to the outside. However, if there is a lot of traffic on the route, the route is frequently subject to the release / new setting of the optical cut-through link, and the value of the metric changes frequently. It will be. The node device of the external network needs to recalculate the routing table entry with the changed metric value, and the load of the node device increases over a wide range.

  Therefore, in the present application, the route information advertised to the outside includes only the optical cut-through link that is stably stretched. Specifically, even if one of the optical cut-through links is installed, the metric that is advertised to the external network is not included in the calculation immediately after the extension, but is released several times after being installed. It is not included in the calculation of the value of the metric to be advertised to the outside until it is continuously stretched even after the set period has passed. The fact that it is continuously stretched even after several cycles passes is because there is a high possibility that the traffic of the optical cut-through link is stable and high.

  By doing so, it is possible to prevent an extra load from being applied to the external node device of the optical network due to the frequent change of the metric value due to the periodic release / new setting of the optical cut-through link.

  The node device according to an eleventh aspect of the present invention is a node device suitable for configuring the optical network according to the first aspect of the present invention. That is, the node device according to the invention according to claim 11 is a node device that is connected by a default link and an optical cut-through link overlaid on the default link, and forwards a packet through the default link and the optical cut-through link. A packet switch router for switching an input packet, a traffic measuring means for measuring the packet traffic switched by the packet switch router and the traffic of the optical cut-through link whose own node device is an end point, and a metric An autonomous distributed routing protocol stack having a routing table in which entries including values are determined and determining an output destination of the packet; and the packet used when becoming an end point of the cut-through link. An optical transceiver for a number of optical cut-through links smaller than the number of all other node devices in the communication network, and an optical cross-connect that defines a route of the optical cut-through link, and a starting point of the optical cut-through link; In this case, after the setting of the optical cut-through link, the metric value changed by the setting of the optical cut-through link is advertised to the adjacent node device and the end-point node device connected by the optical cut-through link. Among the optical cut-through links, the at least one cut-through link having a small traffic value of the optical cut-through link measured by the traffic measurement means is periodically released, and the own node device and other nodes New destination nodes in descending order of packet traffic measured by the device Select location, to perform signaling for setting new optical cut-through link against the selected end point node device.

  The node device according to the inventions according to claims 12 and 13 is a node device suitable for configuring the optical network according to the invention according to claim 2. Details are omitted.

  The node devices according to the inventions according to claims 14 to 20 are node devices suitable for constituting an optical network according to the inventions according to claims 4 to 10, respectively. Details are omitted.

  As described above, according to the present invention, after an optical cut-through link is installed, it is advertised as a new link instead of a path, and is periodically advertised. By opening up a new and different combination of optical cut-through links, it becomes possible to cope with long-term changes in traffic trends, and a finite number of optical transceivers are most needed at each time. It is possible to change the shape of the network to adapt to changes in traffic trends. That is, according to the present invention, by using a limited number of optical transceivers more efficiently, an optical network having a high utilization rate in terms of time and bandwidth at a low cost and a high throughput as a whole and its optical It is possible to provide a node device for realizing a network.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing a first embodiment of an optical network according to the present invention. The optical network 1 includes, for example, eight node devices 2-1 to 2-8. The node devices are connected by optical fibers 3-1 to 3-10. In the description of FIG. 1, basically all types of links are formed in both directions, and description of each unidirectional is omitted. In the optical fiber 3, the wavelength-multiplexed optical signal is transmitted. One of the wavelengths, for example, λ0 is used as the wavelength for the default link. The default links 4-1 to 4-10 connect the adjacent node devices, and the data packet input from the default link 4 is converted into an electrical signal by each node device, and switching is performed on a packet basis. FIG. 1 shows an example in which each node device has two optical transceivers for an optical cut-through link. Light of n types of wavelengths other than λ0 is used for the optical cut-through link.

  FIG. 2 is a configuration example of the node device 2 in FIG. The node device 2 is mainly composed of a packet switch router 15 that performs packet-based switching and an optical cross-connect 11. In the configuration of the node device shown in FIG. 2, there are three examples of node devices adjacent to the node device, and the adjacent node devices are connected by optical fibers. For example, the optical fiber 7-1 is connected to the same adjacent node device as a pair with 7-4. The same applies to the other two sets of optical fibers.

  Next, an operation example of the node device 2 will be described. The optical signal input from the optical fiber 7-1 is separated for each wavelength by the wavelength multiplexer / demultiplexer 6-1. Of the separated wavelengths, λ0 corresponding to the default link is converted into an electrical signal by the default link optical receiver 8-1. In this way, the data packet input via the default link is switched by the packet switch router 15 and output to an appropriate output port. Data packets input from the port 16 connected to the lower level or external network are similarly processed. Light having a wavelength other than the default wavelength λ 0 demultiplexed by the wavelength multiplexer / demultiplexer 6-1 is input to the optical cross connect 11. Here, n optical switches are provided for each wavelength in the optical cross-connect 11, and light of each wavelength output from the wavelength multiplexer / demultiplexer 6-1 is input to the corresponding optical switch. For example, light having a wavelength of λ1 is input to an optical switch denoted as λ1 in FIG. An optical switch can connect an arbitrary input port to an arbitrary output port, and is logically a matrix switch.

  If this node device is not an end node, the light input to the optical switch is output to an output port that leads to an optical fiber 7 connected to an appropriate adjacent node device. The optical signal of the optical cut-through link whose end is the node device is output to the port connected to the wavelength multiplexer / demultiplexer 6-8 or 6-10 by the optical switch, and the wavelength of the optical transceiver 12 for the optical cut-through link After being converted into an electrical signal by the variable optical receiver 14, the packet switch router 15 receives packet-based switching.

  Packets input to the optical cut-through link starting from the local node device are output by the packet switch router 15 to a port connected to the wavelength tunable optical transmitter 13 of the optical transceiver 12 for the optical cut-through link. The wavelength of the tunable optical transmitter 13 is tuned to a wavelength suitable for the optical cut-through link, and the output light is input to the optical switch via the wavelength multiplexer / demultiplexer 6-7 or 6-9. The route is determined by an optical switch as a cut-through link and output.

  The control packet addressed to the own node device is output to the port connected to the node control unit 10 by the packet switch router 15, and the control packet from the node control unit 10 is switched by the packet switch router 15, and the appropriate adjacent node device or Output to optical cut-through link.

  The optical cross-connect in the node device of FIG. 2 is a photonic cross-connect that does not perform wavelength conversion, optical / electrical conversion, or electrical / optical conversion internally, and the optical cut-through path is a path of light that is not subjected to wavelength conversion in the middle of the node device. Continue from the start point to the end node device. For this purpose, a tunable optical transmitter and a tunable optical receiver that are tuned to an appropriate wavelength so as not to be subjected to wavelength blocking in the middle are used.

  Although the photonic cross-connect is used in the configuration example of the node device shown in FIG. 2, the present invention shown in FIG. 1 can be used even if an optical cross-connect capable of wavelength conversion is used instead of the photonic cross-connect. It does not affect the operation of the optical network of the embodiment according to the invention. In short, the wavelength multiplexed signal input from the adjacent node device is separated, the separated optical signal is converted into an electrical signal and packet switched, and the optical signal is switched and output to other adjacent node devices. It is only necessary to have a function to perform, a function to multiplex optical signals of a plurality of wavelengths, and a function to form an end point of an optical cut-through link.

  The routing table used in the packet switch router is created by the node control unit 10 based on the information of the optical cut-through link whose end point is the local node device and the advertisement of the route information received from the other node device. The node control unit 10 instructs the packet switch router 15 to perform switching based on the routing table. Further, the node control unit 10 controls the route of the switch in the optical cross-connect and the transmission / reception wavelength of the optical transceiver for the optical cut-through link. In FIG. 2, the control line is indicated by a one-dot chain line. The node control unit 10 further collects the measurement result of the traffic of the optical cut-through link whose end point is the own node device and the measurement result of the packet traffic switched by the packet switch router. In addition, the node control unit 10 performs determination / instruction of an optical cut-through path release / new setting.

  Returning to FIG. 1, the description will be continued. In the configuration of the node device 2 as shown in FIG. 2, the optical cut-through link is stretched at a certain point of time, for example, 5-1 to 5-8 (gray line). Since each node device has two sets of optical transceivers for optical cut-through links and the optical cut-through links are extended in both directions, eight optical cut-through links are set for the eight node devices in the network. Has been. For example, in the node device 2-1, an optical cut-through link 5-2 to the node device 2-7 and an optical cut-through link 5-6 to the node device 2-8 are set.

  FIG. 1 is an example of an optical network including eight node devices, and the number of optical transceivers for optical cut-through links in each node device is uniformly 2. However, in the case of a large-scale optical network with 10 node devices or 20 node devices, the number of optical transceivers for optical cut-through links in each node device is larger, for example, 4 or 8 respectively. That's fine. In that case, the number of optical transceivers for optical cut-through link may differ depending on each node device, and node devices that tend to concentrate traffic outside the optical network, such as border node devices and node devices with large subordinate networks May be provided with many optical transceivers for optical cut-through links.

  When an optical cut-through link is set, the starting node device, which is bi-directional in the example of FIG. Include in route information advertisement based on routing table. For example, in FIG. 1, when there is an optical cut-through link 5-2 whose end point is the node device 2-1, when the setting 5-6 is newly performed, the node device 2-1 The route information including 5-6 in the route calculation is advertised to the node devices 2-2 and 2-5 and the node devices 2-7 and 2-8 which are adjacent node devices by the optical cut-through link. FIG. 3 shows an image in which an optical cut-through link is handled as a link instead of a path. It seems as if a new link has been generated for a node device that is not physically adjacent.

  An example of the routing table of each node device is shown in FIG. Each entry describes the destination network address and subnet mask, and the corresponding output destination interface (port), and switching is performed accordingly. Each entry corresponding to the entry number indicates the destination network address and the prefix length corresponding to the subnet mask, the output interface number, and the metric value. Metrics vary by routing protocol. For example, in the case of OSPF (Open Shortest Path First), the metric is a cost. The metric may be the number of hops. The entry interface 1 and the metric value item in FIG. 4 are described in two ways separated by slashes. The slash means that the packet corresponding to this entry passes through the optical cut-through link in this optical network before reaching the destination. The number before the slash is the output interface and metric value when the destination is via an optical cut-through link. On the other hand, the number after the slash is the output interface and metric value when only the default link is used for the same destination. The value described in each entry is basically the result of route calculation toward the middle of the network. In addition to this, the boundary node device has a metric value for advertising outside the network. It is updated based on information on when the optical cut-through link was stretched.

  When receiving the route information based on the changed routing table, each node device recalculates the routing table of its own node device based on the route information. When the result of the recalculation is different from the routing table before the recalculation, the result is advertised as new route information to the adjacent node devices including the node devices that are adjacent nodes by the optical cut-through link. In response to these advertisements, each node device further recalculates routes. This is repeated until the calculation result in each node device does not change from that before the calculation. Depending on the size of the network, the calculation results will not change if route information is exchanged several times. This state is called a converged state.

  As a result, many of the packets are sent via a route via the optical cut-through link. In the example shown in FIG. 1, a packet sent from the node device 2-7 to the node device 2-3 is sent via the node devices 2-8 and 2-4 if only the default link is passed. However, depending on the metric value of the optical cut-through link, the node device 2-6 is sent to the node device 2-3 via the optical cut-through link 5-4. There are various methods for determining the metric value depending on the routing protocol. A metric value is generally a closer (better) route with a smaller value, and is often given by the sum of all links (hops) from the destination to the source.

  In the present invention, the metric value of the optical cut-through link is set smaller than the metric value passing only through the default link along the optical cut-through link. For example, it is assumed that the metric value of the optical cut-through link is set to half of the metric value passing only through the default link along the optical cut-through link. By making the metric value in this way, the optical cut-through link exists like a highway that cuts through several node devices, and on the metric value, the optical cut-through link is used rather than passing the route of only the default link. The one that passes is closer (smaller value). Therefore, according to the routing table calculated for each node device, the number of cases through the optical cut-through link increases.

  Each node device measures the traffic of an optical cut-through link that is an end point. Each node device periodically releases the optical cut-through link with the smallest measured traffic volume among the optical cut-through links each serving as an end point. If the number of optical transceivers for optical cut-through link is clearly larger than that of other node devices, it can be judged that the demand for optical cut-through links with the node device as an end point is high. A plurality of optical cut-through links with a small amount may be simultaneously released. The optical cut-through link is reconfigured periodically by releasing or newly setting the optical cut-through link.

  The evaluation of traffic volume, which is a criterion for releasing an optical cut-through link, is made by comparing the traffic volume of a past fixed period with an appropriate weighted averaging process using a function of an appropriate time. However, since the present invention has an object of reducing the switching process of the node device by the optical cut-through link, it is important how many node devices can be cut-through. If the optical route is treated as a link instead of a path, the short line may carry more traffic, and only a short line may survive the optical cut-through link reconfiguration when compared with a simple time average. There is. In such a situation, one of the objects of the present invention cannot be achieved. Therefore, preferably, the traffic is weighted and compared by the number of node devices that can be cut through by the optical cut-through link. For example, when the cut-through of the one-node device is performed, the traffic amount is evaluated as 1.5 times when the cut-through is not performed, for example, twice when the cut-through is performed for the two-node device.

  An end point of one optical cut-through link is a two-node device, and one of them may release it, but the other may release other optical cut-through links. In such a case, one or more optical transceivers are released per node device.

  If you do not want to release and set a lot at once, for example, if you do not just select one with the least traffic and release it, but set a release threshold in conjunction with this and do not meet this However, release signaling is not performed from the own node device, but may be released if there is a release request from the other end. At this time, it is possible to control how much the network is released by adjusting the threshold value.

  In the present invention, basically, all of the node devices in a group of optical networks periodically reconfigure the optical cut-through link by releasing and newly setting the optical cut-through link at substantially the same time. I do. There are various ways to synchronize. For example, the node device with the youngest address in the optical network transmits a signal of a signal for starting the operation of releasing / newly setting the optical cut-through link to the adjacent node device. The adjacent node device that has received the signal of the signal can synchronize the release / new setting of the optical cut-through link by a method in which the node device itself also outputs the signal of the signal. In this case, the node device that receives the signal of the signal transmits the signal of the signal to the node device adjacent to the node device only once so that the packet storm of the signal does not occur. To do. This method is a subordinate system. As another method of synchronizing the reconfiguration of the optical cut-through link by releasing and newly setting the optical cut-through link, each node device is equipped with a timer. There is also a method of starting the release / new setting of the optical cut-through link when time has elapsed. This method is a so-called independent method.

  The period for periodically releasing and newly setting the optical cut-through link is preferably set to a time sufficiently longer than the convergence time of the route information. Typical values of the period are about 10 minutes or 20 minutes.

  In this embodiment, each node device has two sets of optical transceivers for optical cut-through links. For example, one of the node devices is periodically released, and a new node device with another partner device is released. Set the optical cut-through link. It is an image of changing opponents who hold hands together. At this time, the criterion for selecting the partner node device is to select a route with a lot of traffic based on the traffic amount measured by the packet switch router of the own node device, or in some cases, another node device. It is even better if the evaluation is weighted by the number of hops as in the case of release. At this time, a plurality of new opponent candidates are ranked and selected. Even if it is said that the optical cut-through link should be newly established, each node device only releases one or two optical cut-through links, so that one of the node devices is a partner to be newly connected. This is because a device does not always select as a partner with which it wishes to newly connect its own node device. A specific signaling procedure for establishing an optical cut-through link may be the same as that in the conventional example.

  One of the features of the present invention is to perform signaling several times as necessary. Signaling is performed on the selected new opponent candidate from the top. If this signaling fails, the order is lowered and signaling is performed. You can prepare many opponents and repeat until you are successful, or you can reduce the number of candidates and stop when the candidates are exhausted. Furthermore, the signaling deadline may be determined as an optical network, and signaling may be stopped if it is not finished by the deadline.

  If a new partner node cannot be determined, some of the optical transceivers may be left open. In such a case, the next optical cut-through link will be reconfigured to search for another partner. Become. If most of the newly installed node equipment or optical transceivers for optical cut-through links of the node equipment are unused at the initial stage of operation of the optical network, almost all of the optical cut-through link light is used. What is necessary is just to increase candidates so that a transmitter / receiver is used, or to extend the termination period of signaling.

  Here, the outline of the release / new setting procedure is as follows. First, a list of optical cut-through links released by each node device and a list of newly set partner candidates are listed. Next, synchronization for reconfiguration of the optical cut-through link is performed by the node devices in the network. When synchronization is established, route information that changes due to release is advertised to neighboring node devices. Next, signaling for releasing the optical cut-through link is performed to release the optical cut-through link. Next, signaling for new setting is performed, and when the signaling is completed, an empty optical cut-through link optical transceiver is set for an optical cut-through link for which signaling was successful. When the setting is completed, the route information including the link is advertised to the neighboring node device.

  All signaling packets for releasing / setting the optical cut-through link are performed via the default link. In order to release the optical cut-through link, signaling is performed on a hop-by-hop basis with a default link along the optical cut-through link to be released, so that a release notification is also transmitted to the transit node device. The same applies to the setting. For this purpose, the route of the default link is left in the routing table as described above. Since this route is the shortest route hop-by-hop, a new optical cut-through link can be established with the shortest possible route.

  For example, when a new optical cut-through link is to be extended from the node device 2-3 to the node device 2-7, in the state shown in FIG. The route from -6 to the node device 2-7 is taken, but the shortest route in the default link is the route 2-3, 2-4, 2-8, 2-7, so the setting signaling is performed by the latter route .

  In the embodiment shown in FIG. 1 and FIG. 3, the optical cut-through link is stretched in two-way symmetry. Accordingly, the optical cut-through link optical transmitter and the optical cut-through link optical receiver of each node device are used as a set and connected to the same counterpart node device. As shown in FIG. 5, the optical cut-through link may be unidirectional, and the optical cut-through link optical transmitter and the optical cut-through link optical receiver may be used independently. FIG. 5 is an example in which each has one optical cut-through link optical transmitter and one optical cut-through link optical receiver, and unidirectional optical cut-through links indicated by arrows are stretched. In the calculation of the route, it is possible to make the metrics different in both directions, and thus the above procedure hardly changes. Since Internet traffic includes a lot of traffic that is asymmetrical in uplink and downlink, by allowing unidirectional optical cut-through links in this way, it is possible to effectively use optical transceivers for optical cut-through links.

  By doing so, we can effectively use a limited number of optical transceivers for optical cut-through links where there is a lot of traffic, and flexibly adapt the network configuration in response to long-term changes in traffic trends. Network throughput can be improved.

  Next, an embodiment in which such a configuration is performed based on the LSP will be described.

  FIG. 6 is a diagram showing an embodiment in which LSP is used as a base. The configuration of the network such as the use of optical fiber is almost the same as that shown in FIG. The node device configuration is the same as that in FIG. The packet switch router is a label switch router. In the embodiment shown in FIG. 6, data packets are switched at the MPLS layer, not the IP layer. That is, all data packets are labeled MPLS. In addition to the routing table shown in FIG. 4, each node device has a forwarding table that associates the label of the input packet with the output interface, and determines the output interface according to the label of the input packet. In this case, the routing table is used only for control packet transmission and LSP setting, and not for data packet switching. In the network shown in FIG. 6, first, a mesh-like LSP is set as in the conventional example. This LSP is indicated by a two-dot chain line in the figure. The LSP is originally unidirectional, but in order to prevent the figure from becoming too complicated, the bidirectional is shown by a single line for convenience. All data packets input from outside the network or lower networks are labeled to pass through the LSP, and data packets other than MPLS do not pass through the network. Therefore, it is possible to perform traffic measurement at the label switch router in each node device in units of LSPs. In order to prevent the figure from becoming too complicated, the optical cut-through link is extracted and shown in FIG.

  The basic operation of release / setting is almost the same as in FIG. However, in the case of the LSP, simply rewriting the routing table does not change the LSP route itself, so it is necessary to re-establish the LSP route along with the change of the routing table. When a new optical cut-through link is set, route information is advertised, routes are recalculated in each node device, and routes converge in the network, each ingress of each LSP is sent to the egress of each LSP. The latest route is confirmed and compared with the route so far. If it is desired to change to the LSP route including via the optical cut-through link, the route is re-established. Here, whether or not the route change is desirable is determined when the metric value is improved compared to the value before the route change.

  At that time, the “make before break”, which is the principle of re-rooting the LSP, shall be observed. However, when the route before the change is a route that passes only the default link and the route after the change includes an optical cut-through link, the route before the change is not broken down. That is, while maintaining the LSP, if the packet of the label is put there, communication through the LSP can be performed at any time, and the route of only the default link is stored in correspondence with the changed route. deep. When the optical cut-through link through which the LSP passes is released, the traffic is immediately returned to the route of only the stored default link. At this time, the starting node device of the optical cut-through link to be released knows what LSP is passing through the optical cut-through link through the setup signal of the LSP passing therethrough. It is better to notify the ingress of the LSP that passes. It is even better if the optical cut-through link is released after those ingresses have moved the LSP traffic to the default link-only route.

  In the case of being based on LSP, the step of re-sizing LSP and the step of migrating traffic are added in this way. In the embodiment shown in FIG. 1, the release / new setting cycle is sufficiently longer than the time for the route information to converge. However, in the embodiment shown in FIG. It is necessary to set the cycle to a sufficiently long time.

  In the embodiment shown in FIG. 6, since the optical cut-through link is extended along the LSP and the route of the LSP is determined when the LSP is extended, the start point and the end point are clear. The LSP ingress node device knows its route. In most cases, traffic can be completely measured at the ingress node device.

  By doing so, it becomes easy to determine the route of a newly set optical cut-through path and to compare traffic.

  The next embodiment is not based on LSP but based on normal IP routing. The basic operation related to the release / new setting of the optical cut-through link is the same as that of the embodiment shown in FIG. The feature of this form is that, even if a new setting is made, only the network address and its output interface are recorded in the routing table, and the effective optical cut-through link start point and end point cannot be determined easily. is there. Therefore, in the present application, route determination signaling for determining a start point, an end point, and a transit node device is performed before new setting is performed. The timing for performing the route determination signaling is one after the synchronization between the node devices until the new setting is performed.

  At this time, as in the case of the embodiment shown in FIG. 1, since the new optical cut-through link should be connected by the shortest path, the packet of the route determination signaling is transmitted only to the default link. In the example of the routing table of FIG. 4, whether or not to pass through the optical cut-through link, and when passing through the optical cut-through link, the route through only the default link is also shown at the same time. Select only the output destination and send it.

  In this embodiment, in order to determine the end point, the transit node device, and the start point, the following is performed.

  First, how to determine the end point is as follows. FIG. 7 shows an example in which only a specific route in the network is extracted. Each node device measures traffic for each routing table entry, ie, each aggregated network address in FIG. If it is determined that the traffic corresponding to one of the entries is large enough to make a new optical cut-through link, a route formation request packet is sent to the adjacent node device of the default link described in the entry. Since the node device that issues the request does not know in advance the number of hops to the end node device, and it is uncertain whether the end node device will select its own node device as the start point, traffic for determining that a route formation request is issued It is preferable to set a lower threshold.

  At that time, the request source node device address, the network address of the entry, the subnet mask, and the measured traffic amount are described in the route formation request packet. For example, in FIG. 7, the node device 2-1 sends a request to the adjacent node device 2-2 described therein for an entry addressed to a network address of 123.134.156.0/24, for example.

  The adjacent node device 2-2 that has received this looks for an entry in the routing table of its own node device including 123.134.156.0/24. For example, it is assumed that the node device 2-2 has the same entry 123.134.156.0/24. The node device 2-2 adds the address of its own node device as a transit node device to the route formation request packet, and outputs this to the connection destination adjacent node device 2-3.

  The node device 2-3 that has received this retrieves the entry in the same manner. In 2-3, for example, if the entry is only one-digit large entry of 123.134.156.0/23, the node device 2-3 is connected to this request by adding its own node device as a transit node device. The data is output to the adjacent node device 2-4. Similarly, when searching for entries in the node device 2-4, there are two entries, 123.134.156.128/26 and 123.134.156.192/26, and connection destinations on these default links are adjacent. Assume that the node devices are different. Since the node device 2-4 is unsure which of these should be selected, the node device 2-4 is selected as the end point. In this way, the end node device is determined.

  As a condition of the end node device, if the node device 2-4 has only the entry of 123.134.156.128/26, the range is narrowed, so that it is recognized as the end point. Also, if the connection destination of 123.134.156.0/24 is a node device of an external network that cannot continue the optical cut-through link, the local node device is recognized as the end point. However, if it is summarized as 123.134.156.0/25 and 123.134.156.128/25, it becomes 123.134.156.0/24, and the connection destination of these default links When the adjacent node device is the same, the local node device is not set as the end point, but is relayed to the connected adjacent node device.

  In this form, it is not known from where the entry that is judged to have a lot of traffic originates. Therefore, each node device creates and sends out a new route formation request based on the measurement result of the node device separately from the relay of the route formation request. That is, in the example of FIG. 7, when it is determined that not only the node 2-1 but also the node device 2-2 has a large amount of traffic related to 123.134.156.0/24, the node device 2-2 also forms a route. Send a request. This similarly reaches the node device 2-4. Further, the node device 2-3 similarly issues a request for 123.134.156.0/23. If the request of 123.134.156.0/24 ends at the node device 2-4, the larger range 123.134.156.0/23 will naturally end at 2-4. In some cases, 123.134.156.0/24 may reach the destination more than 123.134.156.0/23, and the destination node device may be different. In this case, since the end points are different, it is possible that an optical cut-through link is extended to each end point. There is no particular problem because the links will be deceived.

  Next, a starting point node device and a route are determined. The node device 2-4 that recognizes its own node device as the end node device waits for a certain time, for example, about 3 seconds, from the first arrival of a route formation request from a certain input interface by route determination signaling. Similarly, when other route formation requests having the node device 2-4 as the end point arrive during that period, they are classified and inspected together. This waits for each input interface of the default link of the own node device. In the example of FIG. 7, the node device 2-4 receives the route formation request from the node devices 2-1, 2-2, 2-3 while waiting for the route formation request. Further, even with the same input interface, a different route such as a route that branches off from the node device 2-3 may arrive. These requests are classified by route. That is, the route groups are classified into inclusive relations. When issuing a request, if there are several entries that exceed the threshold for the same adjacent node device, a request is issued for all of them, but some of them may end up at the same end point at the same via node device. is there. Since the optical cut-through link of this application is a link to the last, even if the contents of the entry are different, everything that goes through the same route can be put in the link. Therefore, the end point node device performs grouping according to the inclusion relation of the route, and does not compare the entry values.

  As a result of grouping, one group is composed of, for example, three requests shown in FIG. For each group, the node device 2-4 determines its starting node device. The number of start node devices may be one as long as the route has a relatively small number of hops as shown in FIG. The selection is based on the amount of traffic described in the route formation request. It may be simply compared by traffic, or it is better if weighted by the number of hops and compared as in the case of FIG. When weighting is performed by the number of hops, a node device that is somewhat distant is easily selected as a starting point. When selecting a plurality of nodes, it is preferable to select a node device that is higher in the traffic comparison result and is somewhat distant from each other in order to cope with a situation where traffic suddenly increases in the middle. In this case, a route determination notification packet to be described later may be created and sent separately for each starting point.

  In the example of FIG. 7, it is assumed that, for example, 2-1 is selected as the starting point. The end node device creates a route determination notification in which a route including the start node device, the end node device, and the transit node device is described, and sends this to the adjacent node device 2-3 in the reverse order. 2-3 relays this and sends it to 2-2. Similarly, it reaches 2-1 and 2-1 recognizes its own node device as a starting point, and the route determination signaling ends.

  Route determination signaling is performed almost simultaneously in synchronization. As a result, each node device has several new routes starting from the own node device. The starting and ending points of the route know the traffic of these routes, possibly weighted traffic with the number of hops, at the stage of route determination signaling. In signaling for setting up an optical cut-through link, a partner to which signaling is preferentially performed is selected based on these values.

  By doing so, it is possible to establish an optical cut-through link by determining a route even in an IP routing network in which the route is not clear.

  In such an IP routing network, load balancing is sometimes performed in which packets of the same destination are divided into several routes and output in order to distribute the load of the node device. The optical cut-through link of the present application is a highway and a broadband high-speed link. Therefore, it is not necessary to use the optical cut-through link as one path for load balancing, and it is sufficient to put all of them in the optical cut-through link. Therefore, when issuing a route determination request, a node device that creates a route division by load balancing counts the total traffic volume of multiple entries (output interfaces) corresponding to the destination network address, and calculates the network address. As traffic. When an optical cut-through link is established for a route that has been load balanced, the load from the end node device is too heavy, and there is no room for another optical cut-through link. From there, load balancing may be performed again.

  When the IP routing is used, a packet loop or a packet storm may occur in a transition period of link release / setting. Therefore, preferably, all the traffic is returned to the default link before the link is released, and the route change to the optical cut-through link is not performed until the routing table converges. This is because each node device has an entry that passes only the default link corresponding to all destinations, and the route of only the default link is rarely changed unless a failure or the like occurs. As a result, the number of data packets that only want to pass through the default link temporarily increases. However, since the optical cut-through link to be released is a link with little traffic in the first place, even a slight increase does not have a significant effect. However, if it is not guaranteed that the route that goes only through the default link will not be guaranteed, a loop may occur. Therefore, for example, a flag or the like that permits passage of only the default route may be included in the header or the like of the data packet. These flags may be set using, for example, any unused area of the IP header or an option header.

  Each node device also includes in the information whether or not each entry passes through the optical cut-through link when advertising the route information. When releasing the optical cut-through link, the source node device first detours the packet to the destination network address output to the optical cut-through link by setting the above-mentioned flag at the connection destination of the corresponding default link. Let Thereafter, the optical cut-through link is released, and the route information changed by the release is advertised to the neighboring node device. This flag continues to be set in the same manner until the route information advertisement converges. When setting up an optical cut-through link, advertisement of route information based on the new setting is started on almost all node devices at the same time, so new route information including the new setting on the own node device is received and the route is set. When recalculating, remember the routing table just before receiving the first new route information, continue routing with the old routing table until the route converges, and when it converges, synchronize with the network all at once It is sufficient to move to the routing table after convergence. At this time, the synchronization of the network is performed by transmitting a packet of a signal that each node device has converged to the adjacent node device, and the adjacent node device performs this only once for each signal generation node device. This is possible by a method of relaying to a node device other than the node device that sent the message by the adjacent node device. If this signal arrives for all the node devices in the network, it may be determined that convergence has been completed.

  Regardless of whether the routing base is LSP or IP, the network of the present application creates a highway cut-through link where there is a lot of traffic to reduce the switching load of the node equipment, and a finite number of optical transceivers The purpose is to change the network configuration according to traffic trends. Therefore, in some cases, it may be desired to attach a plurality of optical cut-through links to the same partner. The above procedure doesn't create a new link to the same opponent you just released, but it doesn't mean you shouldn't add more to the one you are currently doing. If the configured optical cut-through link has been busy for a long time, the traffic may be distributed by extending the second line. Since the network of the present application periodically reconfigures the optical cut-through link, depending on how the change is made, the newly set-up optical cut-through link may reduce traffic that has been overcrowded. Therefore, such a second line should be stretched only when the condition of being too crowded for a long time is satisfied.

  In the present invention, the number of optical transceivers used at the end points of the optical cut-through link of the node device need not be the same depending on the node device. If you decide the partner of the new optical cut-through link as described above, you can get long-term statistics on how much endpoint formation is requested to the node device, how much is accepted and how much is rejected Become. If the number of endpoint formation requests is large and the rejection ratio is large, the number of optical transceivers for the optical cut-through link of the node device is not appropriate, and it is considered desirable to be larger. Therefore, it is preferable to have a function of notifying such information to the node device or the network administrator. If the administrator increases or decreases the number of optical transceivers for the optical cut-through link according to the statistical value, the number of optical transceivers for the optical cut-through link can be made more appropriate and efficient use is possible. In general, a large number of optical cut-through link optical transceivers are provided for a node device with a large traffic volume, and a small optical cut-through link optical transceiver is provided for a node device with a small traffic volume.

  Finally, a mode for notifying the metric outside the network of the present application will be described. FIG. 8 shows a configuration in which there are other networks 6-1 and 6-2 in addition to the optical network 1 of the present application, and these are connected via the node device 2-1 and the node device 2-7. For example, if the network addresses managed by eight node devices in the optical network are eight from 123.134.152.0/24 to 123.134.159.0/24, these are 123.134.152. It can be aggregated at 0/21. Therefore, the node device 2-1 and the node device 2-7, which are boundary node devices, notify only the aggregated network address to the node devices 7-1 and 7-2 in the other networks. The metric value at that time does not generally depend on the metric value in the network. Therefore, the release / setting of most optical cut-through links in the network does not affect the metric value advertised to the outside. However, the route information for the node device 7-1 to send the packet to the node device 7-2 via the node device 2-1, the node device in the optical network 1, and the node device 2-7 is obtained as the node device 2-. 1 needs to be advertised to the node device 7-1. At this time, the value of the metric changes depending on whether there is an optical cut-through link between the node devices 2-1 and 2-7. Since the network of this application has a side to test whether large traffic flows through the optical cut-through link, the half-cut optical cut-through link is frequently set and released between the same start and end points. There are also cases where the robot vibrates repeatedly. If the metric advertised to the outside frequently changes for this reason, the external router needs to recalculate route information each time, and the load increases. In the network of the present application, the node device is designed to withstand the frequent setting and releasing of the optical cut-through link, but if it affects the external node device, in the worst case, it may lead to packet loop or packet loss. . Therefore, in this application, when it is necessary to advertise the change in the metric due to the optical cut-through link to the outside, the optical cut-through link is set and then continued even after several optical cut-through link reconfigurations are completed. Only those that have been stretched will be included in the advertised metric. In the example of FIG. 8, even if the optical cut-through link is reconfigured about 5 times after the optical cut-through link 5-2 from the node device 2-1 to 2-7 is set, for example, If 2 is continued, it is included in the route information advertised to the outside. By doing so, it is possible to minimize external influences.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the components without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  For example, the connection between the node devices using the default link in the description of the embodiment has been shown in a case where the connection is realized by the optical transmission technology using the default wavelength. However, in the present invention, the transmission technique in the default link is not necessarily limited to that based on the optical transmission technique. For example, the default link may be realized by a telecommunication technology including a wireless communication technology.

It is a figure which shows 1st Embodiment of the optical network by this invention. It is a figure which shows the structural example of embodiment of the node apparatus by this invention. It is the figure which showed the form of FIG. 1 typically. It is a figure which shows the example of the routing table of the node apparatus by this invention. It is a figure which shows embodiment which comprises an optical cut-through link by the one-way thing in the optical network by this invention. It is a figure which shows embodiment which is an embodiment of the optical network by this invention, and comprises based on a label switch path. FIG. 3 is a diagram for explaining route determination signaling in an embodiment of an optical network according to the present invention; It is a figure for demonstrating the connection of embodiment of the optical network by this invention, and an external network. It is a figure which shows the structural example of the conventional optical network.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Optical network 2 (2-1 to 2-8), 18 (18-1 to 18-2) Node apparatus 3 (3-1 to 3-10), 7 (7-1 to 7-6) Optical fiber 4 (4-1 to 4-10) Default link 5 (5-1 to 5-8) Optical cut-through link 6 (6-1 to 6-10) Wavelength multiplexer / demultiplexer 8 (8-1 to 8-3) Default link optical receiver 9 (9-1 to 9-3) Default link optical transmitter 10 Node controller 11 Optical cross-connect 12 (12-1 to 12-2) Optical cut-through link optical transceiver 13 ( 13-1 to 13-2) Tunable optical transmitter 14 (14-1 to 14-2) Tunable optical receiver 15 Packet switch router 16 Port 17 (17-1 to 17-2) connected to lower level or external network Other networks

Claims (22)

An optical network in which an optical cut-through link is introduced into a packet communication network provided in each node device with a packet switch router for switching a packet transferred on a default link connecting adjacent node devices,
The optical cut-through link includes a number of optical transceivers for optical cut-through links smaller than the number of all other node devices in the packet communication network included in each node device serving as an end point, and the optical cut-through link. Each node device to be cut-through has an optical cross-connect that defines its route,
Each of the node devices determines an output destination of the packet by an autonomous distributed routing protocol based on a routing table that describes an entry including a metric value,
The start node device of the optical cut-through link advertises the value of the metric changed by the setting of the optical cut-through link after the setting of the optical cut-through link to the adjacent node device and the end-point node device connected by the optical cut-through link. And
Each node device has traffic measuring means for measuring the packet traffic switched by the packet switch router and the traffic of the optical cut-through link whose end point is the own node device, and the optical device starting from the own node device. Among the cut-through links, at least one of the cut-through links having a small traffic value of the optical cut-through link measured by the traffic measuring means is periodically released, and the own node device and other nodes Selecting a new end node device determined based on the value of the packet traffic measured by the device, and performing signaling for newly setting an optical cut-through link for the selected end node device An optical network characterized by   2. The optical network according to claim 1, wherein the release of the optical cut-through link and the execution of signaling for new setting are performed at substantially the same time in the optical network.   2. The optical network according to claim 1, wherein the number of optical transceivers for the optical cut-through link is determined in correspondence with the number of endpoint formation requests for the optical cut-through link for each of the node devices.   The packet switch router is a label switch router that performs multi-protocol label switching, the traffic measurement means measures packet traffic in units of label switched paths, and a new optical cut-through link corresponds to one of the label switched paths. The ingress node device of each label switched path sets the new optical cut-through link, the metric value is advertised after the setting of the new optical cut-through link, and the routing table converges after the routing table converges. 3. The optical network according to claim 2, wherein the route of the label switched path is changed when the value is improved from a value before setting the link.   The node device further includes means for storing a route in the most recent past in which the label switched path is set only via a default link, and the ingress node device for setting the new optical cut-through link includes 5. The signaling for setting the new optical cut-through link corresponding to the most recent route set via only the default link stored in the storage means is performed. The optical network described.   The node device performs packet traffic measurement by packet traffic measurement means included in the node device in units of aggregated entries in the routing table of each node device, and is newly set before releasing the optical cut-through link. The optical network according to claim 2, wherein route determination signaling for determining a start point, an end point, and a route candidate of the optical cut-through link is performed. The node device stores, in the aggregated entry in the routing table, information on whether or not any optical cut-through link in the optical network to which the node device on the route to the destination related to the entry belongs belongs. ,
Include information on the presence or absence of passage of the stored optical cut-through link for each entry in the advertisement of route information based on the routing table,
When the information about the presence / absence of passage of the optical cut-through link of any of the entries changes from none to existence by recalculation of the route by advertisement of the route information received from the other node device, the latest past information about the entry Store the output interface of the default link or the connected adjacent node device,
7. The optical network according to claim 6, wherein a control packet used in the route determination signaling is transmitted to the output interface of the last past default link stored or the connected adjacent node device. In the route determination signaling,
Each node device selects from the entries whose traffic is greater than the threshold based on the packet traffic measurement results performed by the packet traffic measurement means in units of aggregated entries in the routing table of the node device. Send the route formation request including the address of the route formation request source node device, the network address of the entry, the subnet mask, and the measurement result to the connection destination adjacent node device,
When the node device that has received this route formation request is a node device in the optical network to which the connected adjacent node device corresponding to the entry included in the route formation request belongs to the node device that has received the route formation request The network address and subnet mask of the entry included in the received route formation request are compared with the network address and subnet mask of the corresponding default link entry in the routing table of the node device that has received the route formation request, and The network address and subnet mask of the corresponding entry of the node device that has received the route formation request are wider than or identical to the network address and subnet mask of the entry included in the route formation request, Or, when there is a plurality of corresponding entries of the node device that has received the route formation request and the connection destination of the output interface of the default link is the same adjacent node device, this request is received in the route formation request. Add the address of the node device as a transit node device address and relay / send it to the adjacent node device to which the default link of the entry is connected,
Neighboring node device having a plurality of entries of the node device that has received the route formation request corresponding to the network address and subnet mask of the entry included in the route formation request, and the connection destination of the output interface of the default link described is different The network address of the corresponding entry, the range of the subnet mask is narrow, or the optical network to which the node device to which the connected adjacent node device corresponding to the entry included in the route formation request has received the route formation request belongs 7. The optical network according to claim 6, wherein when the node device is an external node device, the node device that has received the route formation request is recognized as an end point of a route of a new optical cut-through link candidate. In the route determination signaling,
A node device that receives the route formation request and recognizes its own node device as an end point arrives at the node device within a predetermined time from the time when the node device was first recognized as an end point, and the route through the route formation request includes an inclusive relationship And selecting one or more of a plurality of route formation requests that lead to a result recognized as having the node device as an end point,
The request source node device that has generated the route formation request selected by the node device that has recognized its own node device as the end point is selected as the start point node device of a new optical cut-through link candidate route, and the selected start point 9. The optical network according to claim 8, wherein the transit node apparatus added to the route formation request sent from the node apparatus is selected as a node apparatus through which a new optical cut-through link candidate route passes.   The route information advertised to the external network connected to the optical network has been reconfigured by a predetermined number of times or more after the optical cut-through link is released and newly set in synchronization with the optical network periodically. 3. The optical network according to claim 2, wherein the information is limited to information relating to an optical cut-through link that is not released and is not changed. A node device for an optical network, which is connected by a default link and an optical cut-through link and transfers packets via the default link and the optical cut-through link,
A packet switch router that switches incoming packets;
Traffic measuring means for measuring packet traffic switched by the packet switch router and traffic of the optical cut-through link whose own node device is an end point;
An autonomous distributed routing protocol stack that has a routing table that describes entries including metric values and determines an output destination of the packet;
A number of optical transceivers for the optical cut-through link, which is used when becoming the end point of the cut-through link, less than the number of all other node devices in the packet communication network;
An optical cross-connect that defines the path of the optical cut-through link;
When it is the start point of the optical cut-through link, the metric value changed by the setting of the optical cut-through link after the setting of the optical cut-through link is sent to the adjacent node device and the end-point node device connected by the optical cut-through link. Advertise,
Periodically releasing at least one of the cut-through links having a small traffic value of the optical cut-through link measured by the traffic measurement means from among the optical cut-through links starting from the own node device,
Selecting a new end node device determined based on the value of packet traffic measured by the own node device and other node devices;
A node device that performs signaling for newly setting an optical cut-through link for the selected end node device.   A timer for starting signaling for periodically setting a new optical cut-through link, or means for detecting that an adjacent node device has started signaling for newly setting an optical cut-through link 12. The node device according to claim 11, further comprising:   12. The node device according to claim 11, further comprising a timer synchronized with another node device, and starting signaling for newly setting an optical cut-through periodically according to the timer.   The packet switch router is a label switch router that performs multi-protocol label switching, and the traffic measurement unit performs packet traffic measurement in units of label switched paths, and each of the label switched paths is an ingress node device. When the metric value is advertised after the cut-through link is set and the routing table converges, the converged routing table metric value is improved from the previous value of the new optical cut-through link. The node device according to claim 12 or 13, wherein the node device is changed.   When the label switched path further includes means for storing the latest past route set via only the default link, and is the ingress node device for setting the optical cut-through link, the label switched path is stored in the storage means. 15. The node device according to claim 14, wherein signaling for setting the new optical cut-through link is performed in correspondence with the route in the latest past set via only a default link.   The traffic measurement means performs packet traffic measurement in units of aggregated entries in the routing table, and before releasing the optical cut-through link, the start point, end point, and route of the newly set optical cut-through link are determined. The node apparatus according to claim 12 or 13, wherein route determination signaling for determining a candidate is performed. In the aggregated entry in the routing table, information on whether or not any optical cut-through link in the optical network to which the own node device in the route to the destination related to the entry belongs is stored,
Include information on the presence or absence of passage of the stored optical cut-through link for each entry in the advertisement of route information based on the routing table,
When the information on the presence / absence of passage of the optical cut-through link of any of the entries changes from none to existence due to route recalculation by advertisement of route information received from another node device, the latest past default for the entry Memorize the output interface of the link or the connected adjacent node device,
17. The node device according to claim 16, wherein a control packet used in the route determination signaling is transmitted to the output interface of the latest past default link stored or the connected adjacent node device. In the route determination signaling,
Based on the measurement result of the packet traffic performed by the aggregated entry unit in the routing table by the traffic measuring means, the traffic is selected from the entries whose traffic is larger than the threshold, and the route formation request is made to the adjacent node device to which the default link of this entry is connected Send a route formation request including the address of the original node device, the network address of the entry, the subnet mask, and the traffic measurement result,
When a route formation request is received from another node device, the connection destination adjacent node device corresponding to the entry included in the route formation request is a node device in the packet communication network to which the node device that has received the route formation request belongs. If there is, the network address and subnet mask of the entry included in the received route formation request are compared with the network address and subnet mask of the corresponding default link entry in the routing table of the local node device, and the local node device The network address and subnet mask of the corresponding entry are wider than or identical to the network address and subnet mask of the entry included in the route formation request, or there are a plurality of corresponding entries of the own node device. When the connection destination of the output interface of the default link described in the above is the same adjacent node device, the address of the own node device is added as a transit node device address to the route formation request, and the connection of the default link of the entry is made Relay / send to destination adjacent node device,
Corresponding to whether there are multiple entries of the local node device corresponding to the network address and subnet mask of the entry included in the route formation request, and the connection destination of the output interface of the default link described is different. When the range of the network address and subnet mask of the entry is narrow or the connection destination adjacent node device corresponding to the entry included in the route formation request is a node device outside the packet communication network to which the own node device belongs, The node apparatus according to claim 16, wherein the apparatus is recognized as an end point of a route of a new optical cut-through link candidate. In the route determination signaling,
When receiving the route formation request and recognizing the own node device as an end point, the route arrives at the own node device within a predetermined time from the time when it was first recognized as the end point, and the route via the route formation request has an inclusion relationship. Select one or more of a plurality of route formation requests that lead to a result that is recognized as having the node device as an end point, and
The request source node device that has generated the selected route formation request is selected as the start point node device of the route of the new optical cut-through link candidate, and is added to the route formation request sent from the selected start point node device. 19. The node device according to claim 18, wherein the transit node device is selected as a node device through which a route of a new optical cut-through link candidate passes.   When connected to an external network, the route information advertised to the external network is reconfigured by releasing the optical cut-through link and reconfiguring the packet communication network periodically in synchronization. The node device according to claim 12 or 13, wherein the node device is limited to information relating to an optical cut-through link that has been performed a predetermined number of times or more and is not released and not changed. A packet switch router for switching a forwarded packet and a traffic measurement means for measuring the traffic of the packet and the traffic that is the end point of the packet, and based on a routing table that describes an entry including a metric value, A node device that determines an output destination of the packet;
A default link that connects adjacent node devices and forwards the packet by the packet switch router;
When the traffic is increased by the traffic measuring means of the node device, an optical cut that sets a link that forwards the packet by connecting and cutting through fewer node devices than the node device connected in the default link Through link,
When a start point node device that is a start point of the node device to which the packet is transferred and an end point node device that is an end point are determined, the metric value changed by this setting after the setting of the optical cut-through link is set to the adjacent node. Advertising means for advertising to the end node device connected by the device and the optical cut-through link;
A cut-through that periodically releases at least one of the cut-through links of which the traffic value of the optical cut-through link measured by the traffic measurement means is small among the optical cut-through links starting from the start node device. Link release means;
Endpoint node device selection means for selecting a new endpoint node device determined based on the traffic value measured by the start node device and the other node devices;
Signaling executing means for executing signaling for newly setting an optical cut-through link for the end point node apparatus selected by the end point node apparatus selecting means;
An optical network comprising: A packet switch router for switching a forwarded packet and a traffic measurement means for measuring the traffic of the packet and the traffic that is the end point of the packet, and based on a routing table that describes an entry including a metric value, A node device that determines an output destination of the packet;
A default link that connects adjacent node devices and forwards the packet by the packet switch router;
When the traffic is increased by the traffic measuring means of the node device, an optical cut that sets a link that forwards the packet by connecting and cutting through fewer node devices than the node device connected in the default link An optical cut-through link setting method for an optical network comprising a through link,
When a start point node device that is a start point of the node device to which the packet is transferred and an end point node device that is an end point are determined, the metric value changed by this setting after the setting of the optical cut-through link is set to the adjacent node. Advertising to the device and the end node device connected by the optical cut-through link;
A cut-through that periodically releases at least one of the cut-through links of which the traffic value of the optical cut-through link measured by the traffic measurement means is small among the optical cut-through links starting from the start node device. A link release step;
An endpoint node device selection step of selecting a new endpoint node device determined based on the value of traffic measured at the start node device and the other node devices;
A signaling execution step for executing signaling for newly setting an optical cut-through link for the end point node device selected by the end point node device selection step;
A method for setting an optical cut-through link in an optical network.