JP3746282B2 - Network system, node device, and network management device - Google Patents

Description

  The present invention is suitable for a virtual LAN (VLAN) system, and more particularly, relates to a network system, a node device, and a network management device that can change a network topology in accordance with traffic conditions in the VLAN.

  LAN (Local Area Network) technology advances rapidly. As for Ethernet (trademark), standardization of 10BASE5 was made by the IEEE802.3 committee in the 1980s, and now Gigabit Ethernet has been used in companies and general households around the world. In recent years, the application of LAN technology to the general public network, that is, the world of WAN, has been actively performed.

  One reason why LAN has come to be used drastically is that the server client system can be configured in a tree structure based on the LAN switch. In the case of a conventional LAN with shared media using a yellow cable, when one client is transmitting data to the server, the other clients cannot transmit any data. However, when a LAN switch is used, data can be sent simultaneously as long as there is no data collision at the LAN switch output port. For example, while one client is transferring a data frame to a printer, another client can simultaneously transfer data to the server.

For these reasons, tree-structured LANs using LAN switches have become popular. Since the LAN switch is not a shared media type, a client connected to the LAN switch can freely transmit data.
Ethernet has an address resolution mechanism. If the source client does not know the destination MAC (Media Access Control) address to be transmitted, the source client transmits the destination address of the address resolution frame as a broadcast address. For example, when the host protocol is IP (Internet Protocol), the destination client whose IP address matches tells the source client the destination MAC address. Since this address resolution frame is a broadcast address, the LAN switch connected to the client transmits address resolution frames from all ports. Therefore, the broadcast frame reaches the entire physically connected LAN segment, and eventually the efficiency of the entire LAN decreases. Such inefficient transmission of broadcast frames is particularly problematic when a large-scale WAN is configured with Ethernet. In addition, when a normal LAN switch is used as a WAN, frames are spread over the entire LAN segment, which is a serious security problem.

  For such a problem, in a virtual LAN (hereinafter referred to as “VLAN”), the scalability and security of the network can be improved by using the function of the VLAN switch. For example, a range in which a broadcast or multicast frame can be reached can be logically limited. For this, various methods such as a method of distinguishing the frame input to the port of the VLAN switch on a port basis and a method of distinguishing on a higher protocol basis can be considered. For example, in the VLAN method in which frames are distinguished on a port basis, when a frame is input to a preset port, the frame is output only from ports included in the group. In VLAN, there are a method of explicitly describing a VLAN tag following a frame header, and a method of not implicitly describing a tag.

  According to VLANs, even if they are physically in the same segment, frames are not transferred to different VLAN groups, so that security can be improved.

  By using a VLAN, it is possible to logically limit the broadcast domain of the LAN, and the broadcast packet can be limited even if the network scale increases, so that scalability is improved. In addition, security is improved because unnecessary packets do not reach the LAN.

  However, in the conventional VLAN, when there is a traffic bias in a group having the same VLAN identifier, there is a problem that not only a congestion state occurs in the same VLAN group, but also the traffic of other VLAN groups is adversely affected. is there.

  The present invention has been made in view of such circumstances, and an object thereof is to provide a network system, a node device, and a network management device capable of measuring and controlling traffic on a virtual LAN and avoiding congestion. To do.

  A network system according to one aspect of the present invention is connected to a user end system, identifies a virtual LAN corresponding to a frame to be transferred, and transfers a plurality of virtual LAN edge nodes according to the identification result. A plurality of virtual LAN core nodes that are connected to each other via a transmission path, identify a virtual LAN corresponding to the frame, and transfer the frame according to the identification result; and a plurality of virtual LAN edge nodes Based on the measurement result of the inflow traffic amount and the outflow traffic amount of the frame in each, and the measurement result by the measurement means, the first virtual LAN edge node and the outflow traffic amount that maximize the inflow traffic amount The largest second virtual LAN edge node Determining means for determining, a specifying means for specifying a path including at least one of the first and second virtual LAN edge nodes and the plurality of virtual LAN core nodes, and a transmission corresponding to the path specified by the specifying means And a bandwidth control means for expanding the bandwidth of the road.

  According to another aspect of the present invention, a network system is connected to a user end system, identifies a virtual LAN corresponding to a frame to be transferred, and transfers a plurality of virtual LAN edge nodes according to the identification result; A plurality of virtual LAN core nodes which are connected between a plurality of virtual LAN edge nodes via a first transmission path and a second transmission path, identify a virtual LAN corresponding to the frame, and transfer the frame according to the identification result; , Measuring means for measuring the inflow traffic amount and outflow traffic amount of the frame at each of the plurality of virtual LAN edge nodes, and a first virtual in which the inflow traffic amount is maximized based on a measurement result by the measurement means LAN edge node and second temporary traffic flow amount maximizing Determining means for determining a LAN edge node; specifying means for specifying a path including at least one of the first and second virtual LAN edge nodes and the plurality of virtual LAN core nodes; and a path specified by the specifying means The virtual LAN edge node and the virtual LAN core node so that the second transmission path is used corresponding to the other path and the first transmission path is used corresponding to the other path to transfer the frame. And a control means for controlling the network system.

  According to the present invention, even when there is a traffic bias in a virtual LAN, a route with relatively large traffic can be estimated, and traffic control that increases the bandwidth of a transmission path connecting nodes can avoid network congestion. It becomes possible.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(First embodiment)
FIG. 1 is a diagram showing a configuration example of a network system according to the first embodiment of the present invention. For convenience of explanation, in this configuration example, it is assumed that three VLAN edge nodes and three VLAN core nodes are provided for constructing a virtual LAN (VLAN), and these VLAN edge nodes and VLAN core nodes are installed. The number is arbitrary. The physical layer of the network system according to the present embodiment is, for example, Ethernet (trademark).

  The three VLAN edge nodes are configured by VLAN edge switches 107, 108, and 109 and first control units 114, 115, and 116, respectively. The VLAN edge switch and the first control unit are connected via a control line and a data line. In FIG. 1, the control line is indicated by a dotted line, and the data line is indicated by a solid line. End systems 101 and 102 are connected to a VLAN edge switch 107 of a VLAN edge node via a transmission line, end systems 103 and 104 are connected to the VLAN edge switch 108 via a transmission line, and a VLAN edge switch 109 is connected to a VLAN edge switch 109. End systems 105 and 106 are connected via a transmission line. The number of end systems that can be connected to the VLAN edge node is arbitrary.

  Each of the three VLAN edge nodes is connected to one VLAN core node via a transmission path. As shown in FIG. 1, the three VLAN core nodes are configured by VLAN core switches 110, 111, and 112 and second control units 117, 118, and 119, respectively. The VLAN core switch and the second control unit are connected via a control line and a data line.

  Specifically, the connection between the VLAN edge node and the VLAN core node is a connection between the VLAN edge switch and the VLAN core switch constituting each node, and these two channels are paired. Connected through. The VLAN core switches constituting the three VLAN core nodes are connected to each other (in a ring) by two pairs of transmission paths.

In this embodiment, the network management device 113 is connected to the VLAN core switch 110 of the VLAN core node.
It is assumed that the end systems (101 to 106) do not need to have a network function corresponding to VLAN. For this reason, the network interfaces of the VLAN edge switches 107, 108, and 109 for the end system can process an implicit frame having no VLAN tag transmitted / received to / from such an end system. On the other hand, the network interfaces of the VLAN edge switches 107, 108, and 109 for the VLAN core switches 110, 111, and 112 can process explicit frames with VLAN tags that are sent to and received from these edge switches that constitute the VLAN. It is.

  Since the network interfaces of the VLAN core switches 110, 111, and 112 transmit and receive frames having VLAN tags, in these VLAN core switches, there is no need to perform switching by MAC (Media Access Control) address. You can decide the output port for forwarding. However, since these VLAN core switches use a link aggregation (aggregation) function, which will be described later, they can be switched by looking at the input VLAN tag and have a MAC forwarding function for forwarding by looking at the MAC address. Shall.

  As described above, the second control units 117, 118, and 119 are connected to the VLAN core switches 110, 111, and 112 through the control lines and the data lines, respectively. The second control units 117, 118, and 119 can rewrite the tag forwarding tables of the VLAN core switches 110, 111, and 112 using the control lines. The tag forwarding table is a table indicating the relationship between the VLAN identifier of the input frame and the output port. A VLAN tag of a frame input to the VLAN core switch is read to obtain a VLAN identifier, and a tag forwarding table is searched using this VLAN identifier as a key to derive an output port to be output.

  As described above, the first controllers 114, 115, and 116 are connected to the VLAN edge switches 107, 108, and 109 via control lines and data lines, respectively. In the network on the VLAN edge switch side, a VLAN tag is explicitly included in a frame to be transmitted / received. On the other hand, the network on the end system side is an implicit VLAN, and the VLAN tag is not included in the frame. For this reason, the VLAN edge switch serves as a boundary between a non-VLAN network and a VLAN.

  Specifically, for example, a port VLAN method in which a VLAN identifier is assigned to each port of a VLAN edge switch to which a frame from an end system is input can be considered. Further, a protocol VLAN that assigns a VLAN identifier according to a higher-level protocol is also conceivable. In the present embodiment, description will be made with a port VLAN, but the present invention does not depend on the VLAN method. It is assumed that the first control unit rewrites a VLAN (tag forwarding) table using a control line. In the case of the port VLAN method, the VLAN table indicates the relationship between the end system side port number, the network side port number, and the VLAN identifier. The VLAN edge switch selects one output port from a plurality of output ports with the same VLAN identifier and outputs a frame based on the MAC forwarding table based on the MAC address. Data lines to which the first control units 114, 115, 116 are connected to the VLAN edge switches 107, 108, 109 are used for data communication with the network management apparatus 113.

  The network management apparatus 113, the first control units 114, 115, and 116, and the second control units 117, 118, and 119 are connected to each other via data lines, and can communicate with each other with a frame with a VLAN tag. is there. For this communication, a VLAN identifier VLAN-D (Default) is set.

  Since the VLAN core switches 110, 111, and 112 are ring-connected, there are two paths when performing data transfer between the VLAN edge switches. In order to avoid this, STP (Spanning Tree Protocol) that implements a single path for data transfer is implemented in the VLAN core switch. Alternatively, it is also possible to control the VLAN core switch from the network management device 113 through the second control units 117, 118, and 119, and make a single path using the tag forwarding table. If the VLAN core switches are not ring-connected, it is not necessary to unify the route like STP.

  As described above, the VLAN edge switches 107, 108, 109 and the VLAN core switches 110, 111, 112 are each connected by a pair of transmission paths. The VLAN core switches 110, 111, and 112 are also connected to each other by a pair of transmission paths (two pairs between two adjacent VLAN core switches). If an independent MAC address is used and a frame is transferred to two paired transmission paths, the expected operation is not considered because there are two paths. Therefore, in the present embodiment, a link aggregation control protocol (LACP) that aggregates a plurality of links and virtually displays them on one link (transmission path) is used. LACP is standardized by IEEE802.3ad. LACP is a protocol that allows a plurality of MAC addresses per port to be treated as a single MAC address virtually by a link aggregation layer called an aggregator.

An operation for avoiding congestion by measuring and controlling traffic in the network system of the present embodiment configured as described above will be described.
In the initial state, it is assumed that the VLAN edge switches 107, 108, 109 and the VLAN core switches 110, 111, 112 are each connected by one transmission line. That is, in the initial state, only one link is allocated to the VLAN identifier among the two links forming a pair of the VLAN edge switches 107, 108, 109 and the VLAN core switches 110, 111, 112.

  At the input / output of the ports of the VLAN edge switches 107, 108, and 109, frame traffic is measured in units of VLAN identifiers. The first controllers 114, 115, and 116 access the VLAN edge switches 107, 108, and 109, respectively, and acquire the measured traffic information. Furthermore, the network management device 113 sucks up traffic information from each of the first control units 114, 115, and 116.

  The network management device 113 collects input / output traffic information with the external network at the VLAN edge node. In the case of three VLAN edge nodes as in the present embodiment, there are six routes connecting the VLAN edge nodes, and there are six pieces of traffic information input and output from the external network to the three VLAN edge nodes. . In the case of the three VLAN edge nodes shown in FIG. 1, the traffic of the route between the VLAN edge nodes can have an explicit solution.

  In general, it is necessary to consider a case where there are n VLAN edge nodes. When there are n VLAN edge nodes, there are n * (n-1) paths connecting the respective VLAN edge nodes. Since the traffic volume that can be measured is 2n, when n is 4 or more, the traffic passing through each path cannot be obtained.

Assume that there are n edge nodes in the VLAN network as shown in FIG. Let Ii be the amount of traffic flowing into the edge node number i from the outside, and Ei be the amount of traffic flowing out of the edge node number i. Also, let t _km be the amount of traffic flowing through the path from the edge node number k to the edge node number m. The traffic volume that can be measured is Ii (i = 1 to n) and Ei (i = 1 to n).

Here, first, it is considered that the traffic flowing through the network is equal. At that time, Ii (i = 1 to n) is equal, and Ei (i = 1 to n) is also equal. As a condition for such a situation, first, it is considered that the amount of traffic flowing in all routes is equal. At this time, the traffic t _kj of the path from the edge node k to the edge node other than k is all equal (j = 1 to n, excluding k), and the traffic t of the path from the edge node other than m to the edge node m _jm (j = 1 to n, excluding m) are all equal. Next, when traffic is unbalanced, Ik is the largest among Ii (i = 1 to n) and Em is the largest among Ei (i = 1 to n) think of. In order for the traffic to change and the traffic Ik and the traffic Em to become the largest, the amount of traffic t _km flowing from the edge node k to the edge node m must be large. When the traffic volume other than this route increases, Ik and Em do not become maximum simultaneously.

As described above, from the traffic information of the edge node, it is possible to estimate a route having the maximum traffic among routes between the edge nodes.
Here, consider a case where the end systems 101, 103, and 105 belong to the same VLAN group. This group is called VLAN1. The VLAN group can be set manually for each node. Further, the network management device 113 can set the parameters of the VLAN edge switch and the VLAN core switch via the first control unit and the second control unit.

  The network management apparatus 113 uses the traffic information sucked from each edge node to derive the route with the largest traffic by the above estimation. Of the traffic from the derived ingress VLAN edge node to the egress VLAN edge node, the bandwidth of the route with the largest traffic is doubled. For this reason, the tag forwarding table of the VLAN edge switch and the VLAN core switch is rewritten so that all of the transmission paths that are paired between nodes on the corresponding path (that is, two transmission paths) are used.

  For example, when the traffic on the path from the VLAN edge switch 107 to the VLAN edge switch 108 is the largest, the VLAN forwarding table of the VLAN edge switch 107 is rewritten, and the two tagged transmission lines corresponding to the VLAN 1 tagged frame as the VLAN identifier To be output. Also, the tag forwarding table of the VLAN core switch 110 is rewritten so that two transmission paths (that is, paired transmission paths) connecting the VLAN core switch 111 are used. Specifically, the output port of the tag having the identifier VLAN1 is changed to two. Further, the tag forwarding table of the VLAN core switch 111 is similarly rewritten.

  By rewriting the tag forwarding table as described above, the transmission path between the ingress VLAN edge switch 107 and the egress VLAN edge switch 108 becomes two. LACP doubles the transfer bandwidth of these two transmission paths. As a result, it is possible to double the bandwidth of the route with the largest traffic in the VLAN, thereby avoiding congestion.

  FIG. 3 is a flowchart more specifically showing an operation of avoiding congestion by measuring and controlling traffic. Note that this operation is for one VLAN (for example, VLAN 1 above), and the same operation is repeated for each VLAN. First, in steps S0 to S3, the network management apparatus 113 obtains the ingress traffic volume and the egress traffic volume from all the edge nodes i (i = 1 to n). Next, in step S4, the network management apparatus 113 determines that, for example, the node k has the largest ingress traffic volume, and determines in step S5, for example, that the node m has the largest egress traffic volume. To do. Next, in step S6, the network management apparatus 113 determines that the traffic on the route from the node k to the node m is large from the results of steps S4 and S5. Subsequently, the network management apparatus 113 determines the shortest path from the node k to the node m based on the network topology information set in advance and the information indicating the band in use. Next, the network management apparatus 113 gives a message (command) to double the bandwidth to the first control unit of the edge node k on the route determined in step S7 (step S8). In addition, a message (command) is given to the second control unit of the core node on the path determined in step S7 so as to double the bandwidth (step S9). Further, the network management device 113 gives a message to double the bandwidth to the first control unit of the edge node m on the route determined in step S7 (step S10). Then, the first control unit and the second control unit of each node rewrite the tag forwarding table corresponding to the VLAN.

  The traffic information is always monitored by the network management device 113, and when there is another route where traffic has increased, the route added above is released and the bandwidth of the new route is doubled. . It is also conceivable to release the route added above at specific time intervals.

  In the present embodiment, for convenience of explanation, the VLAN edge switches 107, 108, 109 and the VLAN core switches 110, 111, 112 are connected by two transmission lines that make a pair, and the VLAN core switch 110, 111 and 112 have been described as being connected to each other via two transmission lines. However, the number of transmission lines between the switches is not limited to one pair (= 2). Such a transmission line pair may actually have two transmission lines, or actually has one transmission line, but by utilizing various multiplexing techniques such as time division multiplexing, Equivalently, a transmission path pair (or more transmission path sets) may be secured. In short, if the bandwidth can be controlled in response to traffic control for avoiding congestion, the number of physical transmission lines is arbitrary.

  In the present embodiment, the network management apparatus 113 is described as being connected to the VLAN core switch 110 of the VLAN core node. However, the network management apparatus 113 may be connected to the VLAN edge node. Alternatively, the network management device 113 may be connected to any other VLAN core node in the VLAN.

(Second Embodiment)
FIG. 4 is a diagram showing a configuration example of a network system according to the second embodiment of the present invention. Similar to the first embodiment, the second embodiment also assumes that there are three VLAN edge nodes and three VLAN core nodes.
The VLAN edge node includes VLAN edge switches 207, 208, and 209 and first control units 214, 215, and 216, respectively. The VLAN edge switch and the first control unit are connected via a control line and a data line. End systems 201 and 202 are connected to the VLAN edge switch 207 of the VLAN edge node via a transmission line, end systems 203 and 204 are connected to the VLAN edge switch 208 via a transmission line, and the VLAN edge switch 209 is connected to the VLAN edge switch 209. End systems 205 and 206 are connected via transmission lines. The number of end systems that can be connected to the VLAN edge node is arbitrary.

  Each of the three VLAN edge nodes is connected to one VLAN core node via a transmission path. As shown in FIG. 4, the three VLAN core nodes are configured by VLAN core switches 210, 211, 212, optical switches 220, 221, 222, and second control units 217, 218, 219, respectively. The VLAN core switch and the second control unit are connected via a control line and a data line, and the optical switch and the second control unit are connected via a control line.

  Specifically, the connection between the VLAN edge node and the VLAN core node is a connection between the VLAN edge switch and the VLAN core switch constituting each node, and a connection between the VLAN edge switch and the optical switch. In the present embodiment, the VLAN edge switch and the VLAN core switch are connected via a single transmission line, and the VLAN edge switch and the optical switch are also connected via a single transmission line. These transmission lines form a pair. The VLAN core switches constituting the three VLAN core nodes are connected to each other by a single transmission line (in a ring), and the optical switches are also connected to each other by a single transmission line (in a ring). These two transmission lines form a pair. That is, as in the first embodiment, one VLAN core node is connected to the other two VLAN core nodes by two pairs of transmission paths. The second embodiment is different from the first embodiment in that one of two transmission lines forming a pair (between optical switches) is an optical path.

The network management device 213 is connected to the VLAN core switch 210 of the VLAN core node.
It is assumed that the end systems (201 to 206) do not have to have a network function corresponding to the VLAN. Therefore, the network interface of the VLAN edge switches 207, 208, and 209 for the end system can process an implicit frame having no VLAN tag transmitted / received to / from such an end system. On the other hand, the network interfaces of the VLAN edge switches 207, 208, and 209 for the VLAN core switches 210, 211, and 212 can process explicit frames with VLAN tags that are sent to and received from these edge switches that constitute the VLAN. It is.

  Since the network interfaces of the VLAN core switches 210, 211, and 212 transmit and receive frames with VLAN tags, there is no need to perform switching by MAC address in the VLAN core switch, and output ports that perform forwarding only with VLAN tags. You can decide.

  The second control units 217, 218, and 219 are connected to the VLAN core switches 210, 211, and 212 through control lines and data lines, respectively. The second control units 217, 218, and 219 can rewrite the tag forwarding tables of the VLAN core switches 210, 211, and 212 using the control lines. The tag forwarding table is a table indicating the relationship between the VLAN identifier of the input frame and the output port. A VLAN tag of a frame input to the VLAN core switch is read to obtain a VLAN identifier, and a tag forwarding table is searched using this VLAN identifier as a key to derive an output port to be output.

  Data lines connecting the second control units 217, 218, and 219 and the VLAN core switches 210, 211, and 212 are used for data communication between the second control unit and the network management device 213. . The second control units 217, 218, and 219 are also connected to the optical switches 220, 221, and 222 that constitute the VLAN core node, respectively, so that the state of the optical switch can be controlled via the second control unit. It has become.

  The first control units 214, 215, and 216 and the VLAN edge switches 207, 208, and 209 are the same as the corresponding first control units 114, 115, and 116 and the VLAN edge switches 107, 108, and 109 in the first embodiment. Therefore, detailed description is omitted here.

An operation for avoiding congestion by measuring and controlling traffic in the network system of the present embodiment configured as described above will be described.
In the initial state, it is assumed that the VLAN edge switches 207, 208, and 209 and the VLAN core switches 210, 211, and 212 are each connected by one transmission line. That is, in the initial state, data does not flow through the link between the VLAN edge switches 207, 208, and 209 and the optical switches 220, 221, and 222 of the VLAN core node. Further, no data flows between the optical switches 220, 221 and 222. Further, in the input / output of the ports of the VLAN edge switches 207, 208, and 209, frame traffic is measured in units of VLAN identifiers. The first control units 214, 215, and 216 access the VLAN edge switches 207, 208, and 209 to acquire information, respectively, and the network management device 213 receives traffic from the first control units 214, 215, and 216, respectively. Suck up information.

  Here, consider a case where the end systems 201, 203, and 205 belong to the same VLAN group. This group is called VLAN1. The VLAN group can be set manually for each node. The network management device 213 can set the parameters of the VLAN edge switch and the VLAN core switch via the first control unit and the second control unit.

  As described in the first embodiment, the route with the largest traffic amount is estimated from the input / output traffic amount between the external network and the edge node. The external network is a network outside the VLAN (network between core nodes in this embodiment). By this estimation, the route with the largest traffic amount from the ingress VLAN edge switch to the egress VLAN edge switch can be found using the VLAN edge switch input / output traffic information in units of VLAN identifiers.

  Specifically, the network management apparatus 213 derives a route with the largest traffic using the traffic information sucked from each edge node. Then, the network management device 213 identifies the ingress VLAN edge node and the egress VLAN edge node of the route. For example, when the amount of traffic input from the outside to the VLAN edge switch 207 is large and the amount of traffic flowing from the VLAN edge switch 208 to the outside is large, the traffic flowing through the route from the VLAN edge switch 207 to the VLAN edge switch 208 is the largest. Estimated. The network management device 213 instructs the ingress VLAN edge switch 207 on the route with the largest traffic to rewrite the tag forwarding table through the first control unit 214. That is, in the tag forwarding table of the VLAN edge switch 207, a frame having a VLAN identifier of VLAN1 is initially set to be transferred to the VLAN core switch 210 of the VLAN core node. The data is transferred to the optical switch 220. Further, the network management device 213 controls the optical switches 220 and 221 through the second control units 217 and 218, and all the optical paths from the VLAN edge switch 207 to the VLAN edge switch 208 through the optical switches 220 and 221. Set.

  The path between the ingress VLAN edge switch 207 and the egress VLAN edge switch 208 is switched to the optical path by rewriting the tag forwarding table as described above and controlling the optical switches 217 and 218. By keeping the bandwidth of the optical path sufficiently large, as a result, the bandwidth of the route with the largest traffic in the network is increased. Therefore, congestion can be avoided by increasing the bandwidth. Here, the network management device 213 performs setting control for the first control unit 214 and the second control units 217 and 218 of the nodes on the path between the ingress VLAN edge switch 207 and the egress VLAN edge switch 208. Although described as what is performed, such a route control method is not limited to the above-described one. For example, the network management device 213 may issue an instruction only to the first control unit 214 of the ingress VLAN edge node. In this case, the first control unit 214 of the ingress VLAN edge node that has received the instruction has a GMPLS (Generalized Multi-Protocol Label Switch) mechanism that autonomously switches toward the node on the instructed path. GMPLS is standardized in ITU-T G.8080 or IETF sub-IP.

  Also, the traffic information is always monitored by the network management device 213, and when there is another route where traffic has increased, the route added by the above traffic control is released and the bandwidth of the new route is increased. You may do it. Further, control may be performed so that the optical path is released at specific time intervals.

  In this embodiment, the VLAN edge switches 207, 208, and 209 and the VLAN core switches 210, 211, and 212 are each connected by a single transmission line, and the VLAN edge switches 207, 208, and 209 are connected to the VLAN. It has been described that the optical switches 220, 221, and 222 of the core node are also connected by a single transmission line, and these two transmission lines form a pair. The VLAN core switches 210, 211, and 212 are connected to each other by a single transmission line, and the optical switches 220, 221, and 222 are also connected to each other by a single transmission line. It explained as what makes. However, instead of using two transmission lines as a pair, a larger number of transmission lines may be used for bundling. In this case, it is conceivable to connect many optical paths between nodes by controlling the optical switch.

  The transmission path between the edge node and the core node or between the core nodes may be a single optical fiber. Even with a single optical fiber, a transmission line equivalent to a plurality of optical fibers can be secured by using wavelength multiplexing technology. In order to dynamically set and release the optical path using wavelength multiplexing, it is preferable to make the light source wavelength of the optical transmitter tunable in order to avoid overlapping of wavelengths.

  Specifically, in a communication method such as SONET / SDH (Synchronous Optical NETwork / Synchronous Digital Hierarchy), a frame can be logically multiplexed into a plurality of channels even with a single transmission path. The frame may be transferred with different resource allocations of the multiplexed channels according to the measured traffic.

  In the present embodiment, the network management device 213 is connected to the VLAN core switch 210 of the VLAN core node. However, the network management device 213 may be connected to the VLAN edge node, or may be connected to the VLAN core node at any location. . Also, a plurality of network management devices 213 may be installed to increase availability.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements 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, although the first and second embodiments have been described on the assumption that the edge node and the core node are separated, the present invention can also be implemented in a network having a node having both an edge function and a core function. In this case, the node includes a VLAN switch, an optical switch, a control unit, and the like. The interface with the outside of the network handles a frame to which no VLAN tag is added. Also, the inside of the network handles frames with a VLAN tag added.

  The present invention can be used for all virtual LAN technologies for constructing a virtual network independently of a physical connection form, and is particularly suitable for realizing a backbone for performing high-speed digital communication using an optical fiber.

The figure which shows the structural example of the network system which concerns on 1st Embodiment of this invention. The figure for demonstrating the traffic measurement in the network system which concerns on 1st Embodiment of this invention. The flowchart which concerns on 1st Embodiment of this invention and shows more specifically the operation | movement which avoids congestion by measuring and controlling traffic. The figure which shows the structural example of the network system which concerns on 2nd Embodiment of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 101-106 ... End system, 107-109 ... VLAN edge switch, 110-112 ... VLAN core switch, 113 ... Network management apparatus, 114-116 ... 1st control part, 117-119 ... 2nd control part, 201 -206 ... End system, 207-209 ... VLAN edge switch, 210-212 ... VLAN core switch, 213 ... Network management device, 214-216 ... First controller, 217-219 ... Second controller, 220- 222: Optical switch

Claims (9)

A plurality of virtual LAN edge nodes connected to the user end system, identifying a virtual LAN corresponding to a frame to be transferred, and transferring the frame according to the identification result;
A plurality of virtual LAN core nodes connected via a transmission path between the plurality of virtual LAN edge nodes, identifying a virtual LAN corresponding to the frame, and transferring the frame according to the identification result;
Measuring means for measuring an inflow traffic amount and an outflow traffic amount of the frame in each of the plurality of virtual LAN edge nodes;
Determination means for determining a first virtual LAN edge node having the maximum inflow traffic amount and a second virtual LAN edge node having the maximum outflow traffic amount based on a measurement result by the measurement means;
Specifying means for specifying a path including at least one of the first and second virtual LAN edge nodes and the plurality of virtual LAN core nodes;
Bandwidth control means for expanding the bandwidth of the transmission path corresponding to the path specified by the specifying means;
A network system comprising: In the node device connected to the transmission line,
A forwarding table representing the relationship between the identifier of the virtual LAN and the port;
Means for identifying a virtual LAN corresponding to a frame to be transferred and transferring the frame from a port determined by referring to the forwarding table according to the identification result to another node device on the transmission path;
Control means for receiving a message representing an increase in traffic in the virtual LAN from a network management device and rewriting the forwarding table in accordance with the message, thereby expanding the bandwidth of the transmission path;
A node device comprising: A virtual LAN corresponding to a frame to be transferred is identified by being connected to the user end system, and a plurality of virtual LAN edge nodes that transfer the frame according to the identification result, and a transmission path between the plurality of virtual LAN edge nodes. A network management device connected to a network system including a plurality of virtual LAN core nodes that are connected to each other, identify a virtual LAN corresponding to the frame, and transfer the frame according to the identification result;
Measuring means for measuring an inflow traffic amount and an outflow traffic amount of the frame in each of the plurality of virtual LAN edge nodes;
Determination means for determining a first virtual LAN edge node having the maximum inflow traffic amount and a second virtual LAN edge node having the maximum outflow traffic amount based on a measurement result by the measurement means;
Specifying means for specifying a path including at least one of the first and second virtual LAN edge nodes and the plurality of virtual LAN core nodes;
Bandwidth control means for expanding the bandwidth of the transmission path corresponding to the path specified by the specifying means;
A network management apparatus comprising: A plurality of virtual LAN edge nodes connected to the user end system, identifying a virtual LAN corresponding to a frame to be transferred, and transferring the frame according to the identification result;
A plurality of virtual LAN core nodes that are connected between the plurality of virtual LAN edge nodes via a first transmission path and a second transmission path, identify a virtual LAN corresponding to the frame, and transfer the frame according to the identification result When,
Measuring means for measuring an inflow traffic amount and an outflow traffic amount of the frame in each of the plurality of virtual LAN edge nodes;
Determination means for determining a first virtual LAN edge node having the maximum inflow traffic amount and a second virtual LAN edge node having the maximum outflow traffic amount based on a measurement result by the measurement means;
Specifying means for specifying a path including at least one of the first and second virtual LAN edge nodes and the plurality of virtual LAN core nodes;
The virtual transmission is performed so that the second transmission path is used corresponding to the path specified by the specifying means, and the frame is transferred using the first transmission path corresponding to another path. Control means for controlling a LAN edge node and the virtual LAN core node;
A network system comprising: The network system according to claim 4, wherein the second transmission path includes an optical path. 6. The network system according to claim 4, wherein the second transmission path is released after a predetermined time has elapsed. The network system according to claim 4 or 5, wherein the second transmission path is released in accordance with a decrease in traffic on the path specified by the specifying means. In a node device connected to a first transmission path and a second transmission path composed of an optical path,
A forwarding table representing the relationship between the identifier of the virtual LAN and the port;
The virtual LAN corresponding to the frame to be transferred is identified, and the frame is transferred from the port determined by referring to the forwarding table according to the identification result to either the first or second transmission path. Means,
A message representing an increase in traffic detected in the virtual LAN is received from the network management apparatus, and the forwarding table is rewritten so that the frame is transferred onto the second transmission path composed of the optical path according to the message. Means,
A node device comprising: A plurality of virtual LAN edge nodes each having a virtual LAN edge switch connected to a user end system and identifying a virtual LAN corresponding to a frame to be transferred and forwarding the frame according to the identification result; and the plurality of virtual LAN edges A plurality of virtual LANs each having a virtual LAN core switch connected between nodes via a first transmission path and a second transmission path, identifying a virtual LAN corresponding to the frame, and transferring the frame according to the identification result In a network management device connected to a network system comprising a LAN core node,
Measuring means for measuring an inflow traffic amount and an outflow traffic amount of the frame in each of the plurality of virtual LAN edge nodes;
Determination means for determining a first virtual LAN edge node having the maximum inflow traffic amount and a second virtual LAN edge node having the maximum outflow traffic amount based on a measurement result by the measurement means;
Specifying means for specifying a path including at least one of the first and second virtual LAN edge nodes and the plurality of virtual LAN core nodes;
The virtual transmission is performed so that the second transmission path is used corresponding to the path specified by the specifying means, and the frame is transferred using the first transmission path corresponding to another path. Control means for controlling a LAN edge node and the virtual LAN core node;
A network management apparatus comprising:

Images