JP4500589B2 - Network equipment - Google Patents

Description

  The present invention relates to a network device, and is suitably applied to 1 + 1 redundancy switching control using an Ethernet (registered trademark) line in a network that transmits IP packets, layer 2 frames, and the like.

  A packet-based network may employ a 1 + 1 redundant configuration of lines in order to increase network robustness in the event of a failure. The 1 + 1 redundancy configuration is not limited to the SONET / SDH network OC-n (n = 3, 12, 48, 192, etc.), but each Ethernet (first (10/100 Mbps), gigabit (1 Gbps), 10 gigabit (10 Gbps) Ethernet ( Often used on registered lines.

The conventional 1 + 1 redundancy switching method has been limited to a remedy process for line failures (line errors) such as LOS (Loss of Signal), LD (Link Down), and the number of received frame errors discarded (Patent Document 1).
JP2003-324438 (summary, figure)

  By the way, in an interface (IF) card in an actual packet transmission apparatus, a plurality of lines (ports) are accommodated in one IF card, and only some of the ports are used as a 1 + 1 redundant configuration. The remaining ports are used as separate and independent ports. For this reason, in the working side card and the backup side card that accommodates 1 + 1 lines on the receiving station side, the remaining ports in each card receive different independent traffic, and the amount of traffic is also different from that of the working side card. This is different from the backup card. Therefore, the congestion state is different between the working card and the backup card, and even if there is no congestion on the backup card, the working card generates a frame discard due to traffic congestion, causing unnecessary frame discard. There are cases. Hereinafter, this will be specifically described with reference to the drawings.

  FIG. 4 shows a spanning (STP) tree change of a network accommodating 1 + 1 lines, and FIG. 4A shows a case where there is no failure on the network. In the figure, the B station includes four IF cards C1 to C4, and each card can accommodate eight ports. Further, the E station and the B pole are connected by a 1 + 1 redundant line, the link L1a is accommodated in the port 1 of the IF card C3, and the link L1b is accommodated in the port 1 of the IF card C4.

  Further, in FIG. 4A, the STP tree of the network accommodated in each of the remaining ports 5 to 8 of the IF cards C3 and C4 is indicated by a bold line. Ports 5-8 of IF card C3 accommodate link L5 connected to station A, and ports 5-8 of IF card C4 accommodate link L4 connected to station D. When there is no failure on the network, communication between the A and B poles is set to be performed via the links L2, L3, and L4, and the traffic is processed by the IF card C4.

In this state, assuming that the link L1a of the 1 + 1 line is the working side (W) and the link L1b is the protection side (P), in the IF card C3 that accommodates the working link L1a, the link L5 is reserved (indicated by a dotted line). ) And the amount of communication traffic in the ports P5 to P8 and the like that accommodate them is relatively small, so the card as a whole is not congested.

  FIG. 4B shows a case where a line failure has occurred in the link L4, and this figure shows the result of rebuilding the STP tree after avoiding the line failure. In this state, communication with the A station or the like is performed via the ports 5 to 8 of the IF card C3 that accommodates the link L5 (shown by a solid line) instead of the links L2, L3, and L4. For this reason, in the IF card C3 of the B station, the traffic volume increases rapidly, and frame discard due to traffic congestion may occur.

  Normally, a separate tree is configured and managed for each domain (customer) accommodated in the network, and the STP tree is changed independently for each domain, so the tree has 1 + 1 lines. The customer may be affected by a change in the STP tree of another customer, and the service may be deteriorated.

  The present invention has been made in view of the above-mentioned problems of the prior art, and the object of the present invention is that when a 1 + 1 redundant line is accommodated in each of two IF cards accommodating a plurality of ports, only when a working line failure occurs. In addition, the present invention is to provide a network device that can quickly relieve against a sudden increase in the amount of traffic in a card.

The above problem is solved by the configuration of FIG. In other words, the network device of the present invention (1) is an interface unit 20 that accommodates a plurality of ports, and queues received frames transferred from a plurality of input ports IP _{1 to} IP _m to a predetermined output port OP. Switching control of the 1 + 1 redundant line accommodated in each of the _two interfaces 20 ₁ and 20 ₂ and the one having the common congestion absorbing buffer 24 in the plurality of ports capable of absorbing traffic congestion a control unit 10, which congestion absorption buffer interface unit 20 ₁ for accommodating the working line W performs switching control to the standby line P by detecting a predetermined congestion states, and a city.

According to the present invention (1), each multiple-port two to accommodate the interface unit 20 _1, 20 ₂ to 1 + 1 redundancy line in case that houses respectively, not only when the line failure as in the prior art working lines, working the interface unit 20 ₁ for accommodating the circuit traffic (i.e., common resources) the configuration switching to the protection line even when congested, it becomes possible to high-speed relief during traffic congestion, improve reliability and transmission efficiency of the network To do.

  In the present invention (2), in the present invention (1), the control unit 10 performs the switching control based on the congestion detection only when there is no line failure (line error) of the working line. Therefore, it is possible to improve the transmission efficiency of the 1 + 1 line while maintaining the conventional remedy capability in case of a line error.

  In the present invention (3), in the present invention (1) or (2), the congestion absorption buffer performs frame discard control of each corresponding traffic according to a plurality of discard profiles according to the priority of the traffic, and a control unit Performs switching control to a protection line when a predetermined congestion state is detected for traffic of a specific priority. Therefore, it is possible to preferentially relieve the congestion state of traffic having a specific priority (preferably high priority).

  In the present invention (4), in the present invention (1) or (3), the congestion absorption buffer detects the buffer occupancy rate due to traffic on the working line, and the control unit determines that the detected buffer occupancy rate is predetermined. The switching control to the protection line is performed by the following.

  When the main cause of congestion in the working side interface unit is on the 1 + 1 working line side, the main cause of congestion is shifted to the protection side even if this is simply switched to the 1 + 1 working line side. Therefore, in the present invention (4), the switching control to the protection line is performed when the buffer occupation ratio of the 1 + 1 work line traffic in the working side congestion absorption buffer is not more than a predetermined value. Accordingly, not only the 1 + 1 line but also the transmission efficiency of both interface units (networks) including these lines is improved.

In the present invention (5), in the above-mentioned present invention (1) or (4), the congestion absorption buffer detects a first threshold value for detecting a congestion state based on a traffic congestion frame discard rate and a non-congestion state. The controller has a second threshold value smaller than the first threshold value, and the control unit has a congestion frame discard rate of traffic on the working line exceeding the first threshold value and a congestion frame discard rate of traffic on the protection line Is below the second threshold value, switching control to the protection line is performed. Therefore, it is possible to effectively prevent the working / standby switching control from fluttering in a network in which the traffic volume changes dynamically.

  As described above, according to the present invention, it is possible to provide not only remedy means at the time of 1 + 1 line failure as in the prior art, but also high-speed remedy means for an increase in traffic volume in the interface unit. Can be improved.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the accompanying drawings. Note that the same reference numerals denote the same or corresponding parts throughout the drawings. FIG. 1 is a diagram showing the configuration of the 1 + 1 redundancy switching system according to the embodiment, and shows an application example to the network configuration described in FIG. However, the figure mainly shows a configuration relating to 1 + 1 redundancy switching provided between a certain transmitting station 50 and receiving station 100 in the network. In addition, in the receiving station 100, the relationship between the m input ports IP _{1 to} IPm and the output port OP to which received frames are transferred / multiplexed is shown in detail.

In the transmitting station 50, 51 ₁ and 51 ₂ are line cards (LC) that transmit Ethernet frames (IP packet interpolation), and in this embodiment, frames of the same content are transmitted to 1 + 1 redundant lines. .

In the receiving station 100, 20 _{1 to} 20 n are IF cards that terminate Ethernet (registered trademark) lines, and 21 _{1 to} 21 m convert received optical signals at the input ports IP _{1 to} IPm into electrical signals and block optical signals ( The photoelectric conversion circuit (O / E) 22 that detects and outputs the LOS (Loss of Signal) state terminates the Ethernet frame of each port, and also discards errors in the link down (LD) and received frames. Physical interface (PHY) capable of detecting and outputting the number, 23 is a packet forwarding unit (FWD) that forwards / stops the received packet of 1 + 1 line to the output port according to the current / spare setting, and 24 is the output port at the time of traffic congestion Packets to OP are queued separately in buffers according to their priority and quality of service (QoS) In addition, the congestion absorption buffer (BUF) that can be output by measuring the buffer usage rate due to 1 + 1 traffic and 25 if there is a line failure (LOS, LD, frame error discard number, etc.) or congestion frame discard rate 1 + 1 line switching trigger information is collected and notified to a CPU card 10 to be described later, and a 1 + 1 line switching trigger control unit 10 that performs 1 + 1 line switching control in accordance with a switching command from the CPU card 10 is provided from each IF card. A CPU card 11 for monitoring the switching trigger information and determining whether or not line switching is necessary and switching control, and 11 is a bus line for connecting the CPU card 10 and the IF cards 20 _{1 to} 20 n.

  With this configuration, the transmitting station 50 and the receiving station 100 are connected by 1 + 1 redundant lines L1a and L1b, and the other remaining lines are used as independent lines as described in FIG. Hereinafter, the operation of 1 + 1 redundant line switching control will be mainly described. In this case, the congestion absorption buffer 24 preferably performs active queue management of traffic having different priorities and service qualities separately, for example, by a known WRED (Weighted Random Early Drop) method.

  FIG. 3A shows a congestion discard profile according to the WRED (Weighted Random Early Drop) method. The horizontal axis represents the queue occupation rate of the congestion absorption buffer 24, and the vertical axis represents the frame discard rate due to congestion. The figure shows two congestion discard profiles corresponding to the priorities of traffic mixed in the same IF card, P1 is a discard profile for high priority traffic, and P2 is a discard profile for low priority traffic. .

  In the present embodiment, the congestion frame discard rates according to the profiles P1 and P2 are separately monitored to distinguish between a congestion state mainly composed of low priority traffic and a congestion state mainly composed of high priority traffic. Preferably, by linking the congestion discard rate of the high priority traffic with the 1 + 1 switching trigger, it is possible to preferentially relieve the congestion state mainly composed of the high priority traffic.

  FIG. 3B shows the hysteresis characteristics of congestion / non-congestion detection. The horizontal axis represents time t, and the vertical axis represents, for example, the frame discard rate of packets managed by the congestion discard profile P1. The frame discard rate in each IF card 20 (that is, the congestion absorbing buffer 24) changes dynamically as time passes as shown in the figure. Two threshold values TH1 and TH2 (<TH1) are set in advance at predetermined intervals, and a congestion state is detected when the frame discard rate exceeds the threshold value TH1, and the frame discard rate is set to the predetermined threshold value TH2. A non-congested state is detected by falling below. Thereby, even if the congestion state of the network changes dynamically, it is possible to effectively avoid the flutter caused by unnecessary line switching.

  FIG. 2 is a flowchart of the 1 + 1 redundancy switching process according to the embodiment, which is executed in the CPU card 10. This process is periodically input. In step S11, it is determined whether or not the working side line failure (LOS, LD or the number of received frame errors discarded) is based on the switching trigger information. After performing the 1 + 1 switching process based on this, this process is exited. Therefore, it is possible to maintain the conventional remedy capability in case of a line error.

  If the line failure is not determined in step S11, it is further determined in step S13 whether or not the congestion frame discard rate of the working line exceeds the threshold value TH1. In this way, it is possible to improve the transmission efficiency of 1 + 1 lines while maintaining the conventional ability to recover from line errors by determining the presence or absence of congestion only when there is no switching trigger based on a line failure. It becomes. However, if the congestion frame discard rate does not exceed the threshold value TH1, this process is exited.

  If exceeded, it is further determined in step S14 whether or not the usage rate of the congestion absorbing buffer by 1 + 1 traffic is below the threshold th1, and if it is below, the process proceeds to step S15. In a congestion state due to an increase in 1 + 1 traffic itself, even if switching from working to backup, intra-card congestion just transitions to the backup card as it is, and there is no point in switching. In the present embodiment, the configuration that enables the switching process only when the 1 + 1 traffic is not the main congested state can effectively avoid the frame discard due to the rapid increase of traffic on other lines. However, if the usage rate due to 1 + 1 traffic is not less than the threshold th1, this process is exited.

  If it is lower, it is further determined in step S15 whether or not the congestion frame discard rate on the spare side is lower than the predetermined threshold value TH2. If it is lower, the 1 + 1 line is switched to the backup side in step S16, and this process is exited. Therefore, even when the amount of network traffic changes dynamically, switching flapping can be effectively avoided.

To explain specifically with the state of FIG. 1 or more, the conditions for switching the 1 + 1 line from the use side L1a to the backup L1b are use side card 20 ₁ Congestion frame loss rate exceeds the predetermined threshold value TH1, and when the That is, the usage rate of 1 + 1 traffic in the current-side congestion absorption buffer 24 is lower than the predetermined threshold th1, and the congestion frame discard rate on the spare side at that time is lower than the predetermined threshold TH2. In this way, no congestion has occurred on the spare card side, so switching to the spare card side can improve the transmission efficiency of traffic on the 1 + 1 line.

In the above embodiment, the application example to the switching control of the 1 + 1 redundant line has been described, but the present invention is not limited to this. The present invention can also be applied to switching control of n + 1 redundant lines.
Further, although the preferred embodiment of the present invention has been described, it goes without saying that various changes in the configuration, processing, and combination of each part can be made without departing from the spirit of the present invention.

It is a figure which shows the structure of the 1 + 1 redundant switching system by embodiment. It is a flowchart of the 1 + 1 redundancy switching process by embodiment. It is a figure explaining the congestion determination processing by embodiment. It is a figure explaining a prior art.

Explanation of symbols

10 CPU card 11 bus lines 20 ₁ ~20n IF card 21 ₁ ~21m photoelectric converter (O / E)
22 Physical interface (PHY)
23 Packet forwarding part (FWD)
24 Congestion Absorption Buffer (BUF)
25 1 + 1 switching trigger control unit 50 transmitting station 51 ₁ , 51 ₂ line card (LC)
100 receiving stations

Claims (5)

An interface unit that accommodates a plurality of ports, and a congestion common to the plurality of ports that can absorb traffic congestion by queuing received frames transferred from a plurality of input ports to a predetermined output port in a buffer With an absorption buffer;
A control unit that performs switching control of 1 + 1 redundant lines accommodated in two interface units, respectively, and switches to a protection line when the congestion absorption buffer of the interface unit accommodating the working line detects a predetermined congestion state. A network device comprising: a device that performs control. 2. The network device according to claim 1, wherein the control unit performs switching control based on the congestion detection only when there is no line failure of the working line. The congestion absorption buffer performs frame discard control for each corresponding traffic according to a plurality of discard profiles corresponding to the priority of the traffic, and the control unit reserves a predetermined congestion state for traffic of a specific priority. 3. The network apparatus according to claim 1, wherein switching control to a line is performed. The congestion absorbing buffer detects a buffer occupancy ratio due to traffic on the working line, and the control unit performs switching control to the protection line when the detected buffer occupancy ratio is not more than a predetermined value. Item 4. The network device according to Item 1 or 3. The congestion absorption buffer includes a first threshold value for detecting a congestion state based on a congestion frame discard rate of traffic and a second threshold value smaller than the first threshold value for detecting a non-congestion state. The control unit performs switching control to the protection line when the congestion frame discard rate of the traffic on the working line exceeds the first threshold and the congestion frame discard rate of the traffic on the protection line is below the second threshold. 5. The network device according to claim 1, wherein the network device is a network device.

Images