JP2000078188A - Priority route control method and router device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a priority route control method and a router device capable of improving the band utilization efficiency of network by enabling the routing processing of optimum priority from the data of high real-time property to the data of low real-time property without securing any band. SOLUTION: A switch part 22 for outputting communication packets inputted from plural input ports 21 to plural output ports 23 is composed of a time switch 28 capable of arbitrarily switching the order of plural communication packets together with a space switch 27 for switching the route corresponding to destination addresses applied to the communication packets, and the time switch 28 is controlled so as to arrange the plural communication packets in the order corresponding to the desired delivery time added to each communication packet.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a priority route control method and a router which can route communication packets at high speed and efficiently.

[0002]

2. Description of the Related Art In the field of telecommunications, in order to efficiently exchange and transmit information via a communication network, a path control device (hereinafter, referred to as a "router device") such as an exchange or a router. .) Is used. This router device is widely used particularly in information data communication via the Internet, and plays an important role in configuring the Internet. Hereinafter, a conventional router device will be described with reference to FIG.

FIG. 1 shows an example of a conventional router device, in which 11 is a plurality of input ports, 12 is a switch unit, 13 is a plurality of output ports, 14 is a path control unit, 15
Is a communication packet.

Next, the operation of the conventional router will be described with reference to FIG.

[0005] The communication packet 15 input from the input port 11 is input to the switch unit 12. A destination address (for example, an IP address used in the Internet Protocol (hereinafter, referred to as “IP”)) 15 a indicating the destination of the packet is added to each communication packet 15, and a path for controlling the switch unit 12 is added. Control unit 14
Is the communication packet 1 based on the destination address 15a.
It is determined from which output port 13 the output port 5 should be output.

The switch unit 12 is composed of a space switch 16 and transmits a destination address 15 from a communication packet 15.
When the path control unit 14 retrieves a and inquires the path control unit 14, the path control unit 14 searches the routing table, obtains which output port 13 the communication packet 15 should be switched to, and returns the result to the space switch 16, The space switch 16 receives the result and switches the communication packet 15.

[0007] The communication packet 15 switched to the target output port 13 by the switch unit 12 is output from the output port 13.

When a large number of communication packets to the same destination are input from the input port 11, congestion occurs in the space switch 16, and the communication packets are discarded.

[0009]

The role of the router device is as follows.
This is to route the input communication packet and output it. With the conventional router device, routing is possible with a small delay if the traffic passing through is small, but the delay is large if the traffic is large, and When a communication packet exceeding the allowable capacity of the space switch is input, the communication packet is discarded. That is, the conventional router device provides a best-effort type network in which the best effort is made to transfer the communication packet, but communication may not be possible due to the limitation of the processing capacity or the speed limit of the communication line.

[0010] Voice data of a telephone, moving image data of a broadcast, and the like are data having a high real-time property. If a large fluctuation occurs in a communication delay time or a communication packet is lost, accurate data is received on a receiving side. Playback is not possible. For this reason, the communication has been hitherto performed by securing a necessary band of a communication channel in advance under the concept of "bandwidth guarantee" for data having a high real-time property.

However, when the bandwidth is secured, (a) the bandwidth must be secured on all the lines between the transmitter and the receiver.
(B) When the traffic fluctuates, the bandwidth must be secured at the maximum value, and the bandwidth utilization efficiency deteriorates.
(C) There is a problem that the processing load for securing the bandwidth is large in the router device.

From the above, it is possible to reliably deliver data with high real-time properties by means of securing a band, but there are many problems.

In addition to the bandwidth guarantee, a value indicating whether the priority is high or low is added to each communication packet, and a communication packet with a high real-time property is routed as a communication packet with a high priority. Thus, there is a method of improving reliability.

However, in this case, a communication packet having a high priority is treated preferentially in all the routers passing therethrough, and a communication packet having a low priority is treated low in all the routers. Will be. That is, since the degree of importance of a communication packet is represented by a fixed parameter indicating whether the priority is high or low, a situation occurs in which a high-priority communication packet is prioritized more than necessary. Also had the problem of being uniformly expensive.

SUMMARY OF THE INVENTION An object of the present invention is to solve such a problem and to perform routing processing of a wide range of data from high real-time data to bulk data regardless of real-time performance at an optimum priority without securing a band. ,
It is an object of the present invention to provide a priority route control method and a router device capable of improving the bandwidth utilization efficiency of a network.

[0016]

In order to achieve the above object, according to the first aspect of the present invention, a path control for switching and transmitting a path of a communication packet based on a destination address labeled for each communication packet. The method is characterized in that the desired delivery time is labeled together with the destination address, the priority of packet processing is determined based on the desired delivery time, and the route is controlled.

[0017] According to the above configuration, individual communication packets can be route-controlled in an order according to a desired delivery time, routing processing can be performed with high priority from high real-time data to low data, and network resources can be effectively used. it can.

Further, according to the present invention, a metric value between the sender and the receiver determined from the source address and the destination address of the communication packet, or a metric value between the transmission time and the desired delivery time determined from the transmission time and the desired delivery time of the communication packet. The communication fee is determined using the remaining time, the value obtained from the metric value and the remaining time until the desired delivery time, or a combination thereof as a parameter.

According to the above configuration, fine accounting can be performed for each packet based on the movement distance of information on the space axis or the time axis, instead of the accounting processing based only on the transmission band and distance of the data link.

According to the third aspect of the present invention, a metric value from the router device to the destination obtained from the address of the router device and the destination address of the communication packet, or a metric value from the current time and the desired delivery time obtained from the desired delivery time. The packet processing priority is determined by using the remaining time, the value obtained from the metric value and the remaining time until the desired delivery time, or a combination thereof as a parameter, and the route control is performed.

According to the above configuration, it is possible to differentiate processing according to the attribute of information without terminating an upper layer, to efficiently operate network resources, and to make each router device individually It is possible to know whether the delivery is delayed or advanced for each packet.

According to a fourth aspect of the present invention, a communication packet prepared by storing a transmission time and a desired delivery time of a communication packet in a relay point option header of the communication packet in the Internet Protocol version 6 specification is used. And

According to the above configuration, time axis information can be added to a communication packet without changing the existing IPv6 packet format.

According to the fifth aspect of the present invention, when congestion occurs due to a communication packet having a processing capacity or more, a communication packet having a low priority is temporarily evacuated to a storage device in the router device, and the desired delivery time is set. When the priority of a communication packet that is approaching and being evacuated rises, or when congestion has subsided, the communication packet is read from the storage device and the route control is restarted.

According to the above configuration, it is possible to reduce the discarding of packets at the time of traffic congestion.

In the invention according to claim 6, the communication packets are queued for each unit of the remaining time until the desired delivery time, and when the transfer processing is performed at a different probability for each unit, the communication packet is delayed until the desired delivery time during the waiting. When the unit of the remaining time changes, the unit of the waiting time is changed.

According to the above configuration, an appropriate transfer process corresponding to the remaining time until the desired delivery time can be performed.

Further, in the invention according to claim 7, when the communication packet is made to wait for each unit of the remaining time until the desired delivery time, and the transfer processing is performed with a different probability for each unit, the transfer probability for each unit is calculated as follows: It is characterized by changing according to the number of waiting packets for each unit.

According to the above configuration, it is possible to eliminate an extreme difference in the number of waiting packets according to the remaining time until the desired delivery time.

Further, in the invention according to claim 8, in the relay point option header of the communication packet in the Internet Protocol version 6 specification, the priority of the router device at the time of detour and the priority independent of the desired delivery time are set. In addition, a communication packet created by storing the handling when the desired delivery time is exceeded is used.

According to the above configuration, it is possible to change the route, to process the communication packet with the highest priority without waiting, to wait for the communication packet beyond the desired delivery time as it is, or to discard it. Can be indicated.

According to the ninth aspect of the present invention, the usage and free space of the storage device in the router device, the use status of the link, the number of adjacent routers on the link, the time required for a round trip to the adjacent router, and the like. Is exchanged between adjacent routers, a metric value to the adjacent router is obtained based on the information, and a metric value from the router to the destination is moved using the obtained metric value. Characteristically changed.

According to the above configuration, it is possible to perform adaptive routing even in a situation where the network is crowded, and it is possible to prevent the number of packets satisfying the desired delivery time from being significantly reduced.

In these priority route control methods, a plurality of input ports and a plurality of output ports, a space switch capable of arbitrarily switching connection between a plurality of input ports and a plurality of output ports, and an input from an input port are provided. A time switch capable of arbitrarily changing the order of a plurality of communication packets, a plurality of communication switches, and a path control unit for controlling a spatial switch so as to transmit the communication packet to an output port corresponding to the destination address. A router device provided with a priority control unit for controlling a time switch so that the order of the packets corresponds to the desired delivery time assigned to each communication packet, a sender and a receiver determined from a source address and a destination address of the communication packet Metric value, or the transmission request obtained from the transmission time and the desired transmission time of the communication packet. 11. The router device according to claim 10, further comprising means for determining a communication fee using a remaining time up to a time, a value obtained from a metric value and a remaining time until a desired delivery time, or a combination thereof as a parameter. Metric value from its own device to the destination obtained from the address of the communication packet and the destination address of the communication packet, or the remaining time until the desired delivery time obtained from the current time and the desired delivery time, or the metric value and the remaining time until the desired delivery time. 11. The router device according to claim 10, further comprising a priority control unit that determines a priority of packet processing using a value obtained from the above or a combination thereof as a parameter and controls a time switch. Stored in the relay point option header 13. The router device according to claim 10, further comprising a priority control unit that uses a transmission time and a desired delivery time of the communication packet, a storage device for temporarily storing the communication packet, and a communication device having a processing capacity or more. When congestion occurs due to input, communication packets with low priority are temporarily evacuated to the storage device, and when the desired delivery time approaches, the priority of the evacuated communication packet increases, or when congestion stops. To
14. The router device according to claim 10, further comprising control means for reading a communication packet from a storage device and restarting priority route control, a time switch divided for each unit of remaining time until a desired delivery time, and a queuing. Means for transferring a communication packet in one time switch in which the unit of the remaining time until the desired delivery time has changed to a time switch corresponding to the unit after the change.
Any one of the router devices, the time switch divided by the unit of the remaining time until the desired delivery time, and the extraction probability of the communication packet in the time switch of each unit,
And means for changing the number of packets waiting in the time switch for each unit.
5. The router device described in any one of 5 above, the ID of the detour source router device stored in the relay point option header of the communication packet in the Internet Protocol version 6 specification, the priority unrelated to the desired delivery time, and the desired delivery 17. The router device according to claim 10, further comprising a priority control unit that uses handling when the time is exceeded, the usage amount and free space of the storage device in the router device, the usage status of the link, and the adjacent status on the link. A means for creating a packet storing information such as the number of routers and the time required for a round trip to an adjacent router, and exchanging between adjacent routers, and a metric to the adjacent router based on the information in the packet. Means for calculating the value, and means for dynamically changing the metric value from the router device to the destination using the obtained metric value. DOO router device according to any one claims 10 to 17 including a is realized by.

[0035]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of a preferred route control method and a router according to the present invention will be described below in detail with reference to the drawings.

FIG. 2 shows a basic configuration of a router device using the priority route control method of the present invention. In FIG. 2, reference numeral 21 denotes a plurality of input ports, 22 denotes a switch unit, 23 denotes a plurality of output ports, and 24 denotes a plurality of output ports. A route control unit, 25 is a priority control unit, and 26 is a communication packet.

The switch unit 22 includes a plurality of input ports 21
And a plurality of output ports 23, and a time switch 28 capable of rearranging the output order of communication packets input to the router device. The space switch 27 is a path control unit. 24, and the time switch 28 is controlled by the priority control unit 25.

Communication packet 2 input to the router device
6 includes the “desired delivery time 26b” of the destination of the communication packet in addition to the destination address 26a of the network layer such as the IP address. Unlike a conventional router device including only a space switch as shown in FIG. 1, the switch unit 22 is configured by a space switch 27 and a time switch 28, and a route based on a network layer address (IP address) 26a indicating a destination. In addition to the control (switching of the space switch), the priority of the transmission order of the communication packets is newly controlled based on the "desired delivery time 26b", so that the network resources can be effectively used.

Note that the "desired delivery time" in the present invention refers to a time indicating only one point in time and a case having a certain time width (as a method of specifying the time width, the start time and the last It is also assumed that a time is specified, and a start time or an end time and a time (which can be omitted by arranging in advance, etc.) are also included.

The preceding is an explanation of the first and tenth aspects of the present invention.

Next, the principle of priority control of communication packets based on the "desired delivery time" will be described with reference to FIG.

In FIG. 3, the transmission of information is expressed as a movement in a two-dimensional space composed of a time axis and a space axis.
It is shown as an information transfer vector in a dimensional space. That is, communication and information transmission can be expressed as a space vector from a certain position vector to a position vector in a two-dimensional space as shown in FIG.

At this time, the time axis element is time, and the time itself can be used as an address on the time axis. On the other hand, the space axis element is a network layer address such as an IP address, and is an IP version 4 (hereinafter, “IPv4”).
It is written. ) Is a 32-bit one-dimensional space.
1 for version 6 (hereinafter referred to as “IPv6”)
It is a 28-bit one-dimensional space.

Since the time is a linear space and the IP address space is not a linear space, a metric value between routers is used as a distance between IP addresses.

In the present invention, the transmission of information is represented as a space vector in a two-dimensional space by the time-network layer address, and the time axis component, the space axis component,
Information attributes carried by a communication packet are classified according to factors such as size and declination to differentiate the processing priority and the like in the router device.

FIG. 4 shows elements of a space vector for information transmission. The time axis component of the start point of the information transmission space vector z = the information start point is the transmission time t _s , and the space axis component is the source network layer address s _s . Also,
The end point in the space vector of information transmission = the time axis component of the arrival (desired) point of information is the desired information delivery time (delivery time limit)
t _d , the spatial axis component is the destination network address s _d .

The moving distance | s _d -s _s | on the spatial axis is the moving distance of the end-to-end packet based on the metric value between the routers, and the moving distance | t _d -t _s on the time axis.
| Is the remaining (margin) time until the desired delivery time requested at the time of packet transmission.

In addition to these, the magnitude of information transmission | z | = {(t _d −t _s ) ² + (s _d −s _s ) ² } ^1/2 is the time-space in the end-to-end information transmission. represents the moving distance of the deflection angle _{Θ z = arctan {| s d} -s s | / | t d -t s |} corresponds to the moving distance on the spatial axis per unit time.

[0049] Even when the moving distance of the space-time are equal, argument theta _z is greater signaling needs to be transmitted farther than per unit time, argument theta _z as calculated parameters of the communication cost or communication fee Available. Even the value of argument theta _z are the same, the moving distance of the space-time | z |
Is large, it means that the number of passing router devices and the transmission distance are large, and | z | can also be used as a calculation parameter of the communication cost or communication fee.

Instead of calculating the fee based only on the transmission band and distance of the data link, the moving distance of the spatial axis | s _{d
-S s | t d -t s |} | and the magnitude of the vector on the space-time |, the moving distance of the time axis z |, declination theta _z such individual parameters or finer for each packet by a combination thereof, It will be possible to make a great charge.

The preceding is an explanation of the second and eleventh aspects of the present invention.

Also, y in FIG. 4 represents a space vector of information transmission in the router device during communication. At this time, the time axis component of the starting point of the vector represents the current time t _c at which there is information in the router device, the spatial-axis component represents the network layer address s _c of the router device. Therefore,
y is a space vector for information transmission from the router device in communication to the final destination.

| S _d -s _c | represents the remaining distance from the router device to the destination address based on the inter-router metric value, and | t _d -t _c | represents the remaining time until the current delivery deadline. (Margin) Indicates time. Further, the magnitude of the vector | y | = {(t _d −t _c ) ² + (s _d −s _c ) ² } ^1/2 is space-time from the current router device for processing the communication packet to the final destination. represents whether the remaining how moving distance above deflection angle _{Θ y = arctan {| s d} -s c | / | t d -t c |} is transmitted information before delivery desired time t _d To end
It represents the distance that must be moved on the spatial axis per unit time in the future.

Each component | s _d − of these information transfer vectors
The router device determines the priority of the packet routing process using any one of s _c |, | t _d −t _c |, | y |, Θ _y , or a combination thereof as a parameter.

Next, | y | About and theta _y priority processing in the router apparatus according will be described with reference to FIGS.
In FIG. 3, two types of different information transmissions, that is, real-time data transfer and file data transfer, are expressed as space vectors. Since real-time transfer requires short-term end-to-end information transmission, the argument ベ ク ト ル of the vector is larger than that of file data transfer.

[0056] In the middle of the router device, when none as large real-time data of the deflection angle theta _y small file data argument theta _y must be processed to determine the priority of processing the value of the argument theta _y In this case, the file data is stored in the router device, and by performing the queuing, it is possible to process and send high-priority real-time data first.

As described above, unlike the conventional router device,
The difference between the current time and the desired delivery time and the remaining distance on the spatial axis without having to terminate the upper layer enables processing to be differentiated according to information attributes, and efficient use of network resources it can.

The preceding is an explanation of the third and twelfth aspects of the present invention.

The desired delivery time can also be regarded as a measure of a discarding mechanism in the router device. For example, in the router device, it is possible to determine that a packet after the desired delivery time is meaningless information even if it is delivered to the destination (stream data reproduced in real time, etc.). If not, it can be actively discarded in the router device (when the current time t _c in the router device> the desired delivery time t _d ).

Since the argument Θ is a function of the current time t _c , by comparing the initial value Θ _z with the current value Θ _y in the router device, it is determined whether the delivery of the communication packet is behind schedule. It is possible to determine whether the process is proceeding, and this can be reflected in the priority of the processing in the router device.

The above is a supplementary explanation of the third and eighth aspects of the present invention.

FIG. 5 shows an example of storing time axis information in IPv6. The spatial axis components, that is, the source IP address s _s and the destination IP address s _d are each represented by 128 bits as an IPv6 header. The time axis components, that is, the transmission time t _s and the desired delivery time t _d, are represented and stored in 32 bits, respectively, using a relay point option header following the IPv6 header.

The priority of header processing in IPv6 is as follows:
IPv6 header> relay point option header (hop-b
y-hop options header> Destination option header
ns header)> Routing header (routin)
g header)> fragment header (fragmenth)
header)> authentication header (authenticati)
on header)> Encrypted payload (encaps)
updating security payload)
> End point option header> Upper layer header (TCP or UDP, etc.), and the relay point option header is processed with the highest priority as an option header.

At this time, “0” representing the relay point option header (HBH) is entered in the next header number in the head IPv6 header. In the next header number of the relay point option header storing the transmission time and the desired delivery time, for example, if the next header is a TCP header, a value of "6" is stored.

As the extension header length, a value obtained by expressing the relay point option header length (128 bits) in 8-byte units (64 bits) and minus 1, that is, "1" is entered.

The option number is an 8-bit identifier. The upper two bits indicate an operation when the router device cannot recognize the option, and the next one bit indicates whether the option can be changed on the delivery route. The lower 5 bits are an option number portion indicating the content of the option.
In the case of the IPv6 header used in this router device, the operation is “00” (meaning that if it cannot be recognized, it will be skipped),
The path changeable bit is “OFF”, and the option number part is IANA (Internet Assist).
Enter the new number that is described in the RFC by registering with the Gned Numbers Authority. The option data length is a byte representation of the option data length, and contains "12".

Thereafter, the transmission time t _s and the desired delivery time t _d are stored as 32-bit addresses as the start and end addresses of the information transfer vector. The last four bytes of the relay point option header are reserved for storing additional information such as the unit of time information, but the unit of time information is milliseconds (m
s), seconds (s), time (hour), when expressed roughly in three stages, priority of IPv6 header (4 bits)
May be used to represent (store) the unit of the time information.

The preceding is an explanation of the fourth and thirteenth aspects of the present invention.

By using the IPv6 relay point option header, it is possible to carry the time axis component without changing the existing IPv6 packet format, and process the packet in the same manner as a conventional packet having no time axis component as information. It is also possible to do.

Further, it can be said that the transmission time and the desired delivery time stored in the packet express the reproduction timing when carrying isochronous (isochronous) information such as an MPEG file (MPEG TS frame). like). In other words, since the desired delivery time attached to the communication packet can be used as a time stamp for reproducing the isochronous data at the destination, the network does not need to implement a transfer mechanism for isochronous assurance such as suppressing delay fluctuation. .

Furthermore, the desired delivery time can be regarded as the absolute sequence number of the communication packet, and the original information can be reconstructed. Therefore, the network must guarantee the order of arrival of the communication packet. Disappears,
The router device can implement a simpler transfer mechanism.

The above is a supplementary explanation of the fourth and thirteenth aspects of the present invention.

Next, a specific example of the router device of the present invention will be described.

FIG. 6 shows the overall configuration of the router device. FIG. 6A is an overall perspective view, and FIG. 6B is a wiring diagram. In the figure, 31A, 31B, 31C and 31D have a full-duplex communication path and the input / output section constituting the path control section, priority control section and time switch described in FIG. 2, and 32 is a disk array computer (PC ) And 33 are input / output units 31A to 31D.
And a switch (SW) unit for spatially switching between the disk array PCs 32; 34, a control computer (PC); 35, a dedicated backplane for connecting them; and 36, a disk array.

The control PC 34 is equipped with a routing protocol for exchanging link state information with another router device, and measures the distance on the spatial axis between the router devices.

The input / output sections 31A to 31D have the same configuration, and FIG. 7 shows an example of the detailed configuration, here, an example in which data exchange with a network is performed by an optical signal. In the figure, 311a is an optical input unit (O
pt in), 311b is an optical output unit (Opt out)
t), 312 is a serial / parallel converter (S / P),
313 is a data link layer termination unit, and 314 is an FPGA (F
field Programmable Gate ar
ray), 315 is a signal processor (DSP), 3
16a, 316b, 316c are synchronous SRAMs
(SSRAM) or the like, and 317 is a content addressable memory (Content Addressable Me).
memory: CAM), 318 is synchronous (syn)
c. ) FIFO memory, 319 is an interface unit (I
/ F).

Hereinafter, a description will be given of how the input / output section constitutes the above-described path control section, priority control section, and time switch, together with an outline of the operation thereof.

The packet data received from an unillustrated network via an optical fiber is photoelectrically converted by an optical input unit 311a, serial-parallel-converted by an S / P 312, and then converted to a Gigabit Ethernet (Gigabit E).
The data is extracted as an IP datagram through a data link layer terminating unit 313 such as the Internet
Are sorted according to the accuracy (unit) of the time axis component.

Here, the display unit of the time axis component is stored at the end of the relay point option header for storing the priority of the IPv6 header or the time axis component as shown in FIG. In the example of FIG. 7, the unit of the time axis component is classified into three stages of millisecond (ms), second (s), and time (hour).

The data classified by the FPGA 314 for each display unit of the time axis component is a high-speed memory 316a, 316b, 316c for millisecond (ms), second (s), and time (hour) as an IP datagram. Is stored in
At this time, the IP datagram is sequentially stored in the memory,
The memory is used as a ring buffer.

[0081] Meanwhile, the header portion of the IP datagram is stored in the fast memory 316A-316C (including relay point option header.) It is passed to the DSP315 for argument theta _y calculation of the route calculation and signaling vector.

[0082] DSP315 calculates a route from the destination address s _d of the IP header, it should sends the packet from Donoho path, i.e. whether to be sent next to which input-output unit via the SW unit 33 decide.

This route information is stored in the CAM 317 together with the write address (wp) when the IP packet is stored in the high-speed memories 316a to 316c, in units of time axis components. At this time, CA storing {direction, unit, wp}
As the write address of M317, the reciprocal of Θ _y calculated by the DSP 315 is used.

The CAM 317 is different from a normal memory.
Instead of giving an address to access stored data, the memory is provided with a search word and knows the address where the data matching the search word is stored.

For example, by giving the CAM 317 a route and a unit as a search word, only information on a specific route and unit can be extracted. When a plurality of pieces of information having the same route and unit are stored, the CAM 317 returns the minimum address from the stored data. Therefore, the write address of the CAM 317 storing {route, unit, wp} is { _y }. by using the inverse of, it is possible to know the storage position wp maximum theta _y information for a particular unit of a particular route maps, namely the high-speed memory 316a~316c the IP datagram having the maximum theta _y.

The thus obtained wp is stored in the high-speed memory 316.
a to 316c are used as the read addresses (rp) so that the IP having the maximum Θ _y of the specific unit for the specific route is always used.
Datagrams can be retrieved from high speed memories 316a-316c.

The IP datagram having the maximum Θ _y extracted from the IP datagram in this manner is represented by I
The data is transferred to the SW unit 33 via the backplane 35 via the / F 319.

On the other hand, from the SW unit 33 to the back plane 35
Via the I / F 319, the IP datagram arriving via Sync. Received by FIFO memory 318. S
ync. The IP datagrams stored in the FIFO memory 318 are sequentially sent to the data link layer terminating unit 313 and the S / P3
12. The signal is transmitted to the optical output unit 311b, converted into an optical signal, and transmitted to the network.

FIGS. 8A and 8B show the detailed configuration of the above-described switch unit (SW unit). FIG. 8A is a circuit configuration diagram, and FIG. 8B is an explanatory diagram of switching connection. In the figure, reference numeral 331 denotes a crossbar switch having a 5 × 5 switch configuration composed of a high-speed crossbar switch chip. The five input ports and the output ports are connected to the disk array PC 32 together with the four input / output units 31A to 31D. Have been. 332
Denotes a switching control CPU, and a crossbar switch 331
Is switched to five phases covering all the one-to-one connections of each input / output port as shown in FIG. Note that a guard time is set between the phases.

The disk array PC 32 is a large-capacity storage device for storing low-priority IP datagrams for which the desired transmission time t _d is set to ∞ (infinity), in this case, the disk array 36. Is to control the

When there is a large margin before the desired delivery time,
Communication packets for which the desired delivery time is set to infinity are sent from the input / output units 31A to 31D to the SW unit 33 and P
Via the C32, the data is temporarily stored in the disk array 36, and after an appropriate time has passed, the PC 32 and the SW unit 33
Is read out to the input / output unit again.

At this time, by setting the desired delivery time to infinity, it is possible to specify that the data can be delivered regardless of the time, so that the data is reliably transmitted while being stored in the storage device in the router device. A new information transmission service class can be defined. Further, even when traffic is congested, low priority packets can be saved in the storage device instead of being discarded, so that the packet discard rate can be improved. As a result, a transformer having a simple end-to-end delivery confirmation mechanism can be achieved. Reliable data transfer can be realized by the port layer.

In the above description, saving data to the disk array 36 corresponds to the fifth and fourteenth aspects of the present invention.

By the way, in the above-described embodiment,
The packet data is stored in the high-speed memories 316a to 316c in the order in which they are simply received in units of time axis components (milliseconds, seconds, and time).
The extraction of the packet data from 16c is simply performed in the descending order of the priority (Θ _y ) for each route and for each unit of the time axis component.

Therefore, in a situation where the network is crowded, the remaining (margin) time until the desired delivery time of the packet data stored in the memory 316c corresponding to a unit of a certain time axis component, for example, the time (h), is Even if it decreases, the transfer of the packet data in the other memories 316a and 316b is prioritized, and as a result, the memory 3
There has been a problem that the desired delivery time of the packet data stored in 16c may not be kept.

FIG. 9 shows a main part of another embodiment of the router apparatus according to the present invention, in which the above-mentioned points are improved, in which packet data is stored (queued) and taken out in high-speed memories 316a to 316c. It shows the parts involved.

In this embodiment, each high-speed memory 316a
Queuing the packet data priority within ~316c (Θ _y) performs in order to change the memory itself to be queued if a unit of time remaining until the delivery desired time (order) is changed by such waiting. The packet data is taken out of the high-speed memory 31.
For each of the 6a to 316c, the priority (higher priority) is changed, but the fetch (transfer) probability (ratio) from each of the high-speed memories 316a to 316c is changed according to the amount of queued data.

Each of the aforementioned high-speed memories 316a to 316c
The queuing of the packet data in the priority order is realized by rearranging the packet data by the address taken out in the priority order from the associative memory 317, and the higher the priority (Θ _y ), the higher the speed of each of the high-speed memories 316a to 316.
The array is arranged so that c is closer to the head (exit) and lower is closer to the tail (entrance).

The packet data stored in each of the high-speed memories 316a to 316c is taken out from the high-priority data, that is, A, B and C in each memory.

At this time, if the order of the remaining time until the desired delivery time changes due to a long waiting time or the like, for example, the order is the remaining time until the desired delivery time of the packet C in the memory 316c at the time (h). When the time becomes 59 minutes 59 seconds or less, the order transition control unit 41 promptly moves to the end of the memory 316b of the second (s) through the route of c1. Similarly, the memory 3 whose order is second (s)
When the remaining time until the desired delivery time of the packet B in the packet 16b becomes 999 ms or less, the order transition control unit 41 passes the route b1 through the route b1 to store the millisecond (ms)
Move to the end of 6a. This is realized by rewriting the bit of the unit of the time axis component in the associative memory 317.

The extraction of the packet data from the high-speed memories 316a to 316c is controlled by the extraction control unit 42 in more detail.

That is, the take-out control unit 42
6a to 316c, the change in the number of stored packet data (queue length) is observed.
The extraction probability is changed for every 〜316c. However, the retrieval probability is assumed to be a fluctuation in the range of 0 <probability <l, and unless there is a memory having no waiting and arriving packets, there is no possibility that the retrieval is performed only from a specific memory or that no specific memory is retrieved. .

FIGS. 10 and 11 are flowcharts showing a queuing and order transition control algorithm at the time of receiving a packet. FIGS. 12 to 14 are flowcharts showing an extraction control algorithm at the time of packet transmission.

The preceding is an explanation of the sixth, seventh and fifteenth and sixteenth aspects of the present invention.

FIG. 15 shows another example of storing time axis information in IPv6, as in FIG. Here, there are two types of desired delivery times, td1 and td2, which enable designation of desired delivery times after td1 and before td2. That is, td1 may be set to 0 in order to handle td as in FIG. 5, and if td1 = td2, the packet is transferred so as to be delivered at that time.

When a detour is performed when the transmission destination is crowded, the router ID of the detour source is described in order to prevent a packet loop. It is "0" when not detouring. In order to perform this change, the changeable bit on the path of the option type is “1”, which is different from FIG.

Further, a field of the priority and the order of the time td is provided. When the priority is “1”, the packet received by the router is sent to the next router quickly without performing queuing, assuming that the priority is the highest. In this case, the field of the desired delivery time is not referred to, and therefore, the calculation of the priority Θ is not performed. In the case of "0", queuing is performed for each order of the desired delivery time as usual.

Using the last bit of this field, the handling of a packet exceeding the desired delivery time is specified.
In the case of "1", delivery is performed without discarding even after the desired delivery time. In the case of "0", packets exceeding the desired delivery time are discarded.

The preceding is an explanation of the invention according to claims 8 and 17.

In the above description, the routing protocol uses a fixed metric value, and changes in the processing time to the destination due to fluctuations in available bandwidth due to fluctuations in traffic flowing through the network. Can not take effective measures against In other words, as long as the packet cannot be reached, such as when the router in the middle becomes inoperable, the transfer efficiency is reduced due to congestion, and even if packet loss starts to occur, the route that was regarded as “shortest” when creating the routing table The problem of continuing to take it arises. For this reason, the number of packets satisfying the desired delivery time may be significantly reduced depending on the traffic situation of the network.

The above-mentioned problem can be solved by exchanging information on the network status between adjacent router devices and enabling the metric value to be dynamically changed.

FIG. 16 shows an AS (Autonomous S).
Open Shortes (OSPF) is an example of propagation of routing information in a system (autonomous system).
tPath First) LSA (Link Sta)
9 is an example of storing exchange information with an adjacent router in a te Advertisement packet.

Since detailed information is exchanged only with the adjacent router, the flooding scope is set to Link Lo.
cal. The CLASS here matches the number of queues (high-speed memory), and CLASS = 3 if there are three queues of ms, s, and h.

When a large-capacity semiconductor memory and a hard disk are used as storage for saving low-priority packets, a subsequent STR bit is set. Also, if_num is the total number of interfaces inserted into the router (excluding communication ports and other interfaces used only during management). qlen_used is the number of packets queued in each queue, and qlen_unuse
d indicates the vacancy of each queue, that is, the number of packets that can be stored.

Mbuf_used, mbuf_unus
ed is the memory buffer usage and free space, respectively.
It is a value expressed in units of ytes. Similarly, memory_u
sed, memory_unused and disk_u
“sed” and “disk-unused” are also values indicating the used amount and the free space of the storage in Kbytes.
These packet numbers and capacities are not instantaneous values at the time of measurement, but are average values for a certain time section.

Further, tr_if is a structure having information for each interface as a member. tr_if. inte
rface_id is an ID for uniquely identifying an interface in the router, and tr_if. if_baudra
te indicates the physical speed of the interface.

[0117] tr_if. The statistics indicates the number of bytes of packets that have entered the router via the interface per second. point-t
In the case of an o-point link, this numerical value indicates the use state of the link, and the point-to-multi
If the link is a point link, the usage of the link is the sum of the statistics of all routers on the link.

[0118] tr_if. route_num is the number of neighboring routers on the link. tr_if. rout
er_id is the id of the adjacent router, and tr_if. r
tt is the time required for the round trip to the neighboring router RTT
(Round Trip Time).

The information is exchanged between the adjacent routers, and the metric values up to the adjacent router are calculated. At this time, three types of metric values for each router are calculated for ms, s, and h. This formula is calculated as a function of the information obtained. Therefore, the metric value is dynamic, and the metric value of ms is used as a representative value for AS.
Distributed to all routers by flooding. Assuming that this metric value corresponds to the distance on the spatial address in FIG. 3, Θ _y is calculated.

Here, it is assumed that the exchange information sent from the adjacent router A to the router C is given as a member of the structure A, and the average value of the number of packets existing in the ms queue in the router A and the storable packets The average of the numbers of A. qlen_used [0] and qlen_u
used [0], and similarly, the used amount and the remaining capacity of the memory buffer in A are A. mbuf_used
And A. mbuf_unused, and the RTT with router A is A.mbuf_unused. RTT, the physical speed of the link where router A is located if_baudrate, the usage status of the link calculated from the usage status obtained from all the neighboring routers for router C on the link where router A is present. s
metric values in ms to neighboring router A using these as |
If S _A −S _C | _ms is expressed, | S _A −S _C | _ms = f (A. Qlen_used [0],
A. qlen_unused [0], A. mbuf_u
sed, A .; mbuf_unused, A. rtt,
A. if_baudrate, A.I. static
s) Hereinafter, similarly, the metric values at s and h are expressed as | S _A −S _C | _s = g (A. qlen_used [1],
A. qlen_unused [1], A. mbuf_u
sed, A .; mbuf_unused, A. memor
y_used, A. memory_unused, A.
rtt, A.R. if_baudrate, A.I. stati
sts) | S _A −S _C | _h = h (A. qlen_used [2],
A. qlen_unused [2], A. mbuf_u
sed, A .; mbuf_unused, A. disk_
used, A. disk_unused, A. rtt,
A. if_baudrate, A.I. static
s)

As specific examples of f (), g () and h (),
The following formula is shown.

| S _A -S _C | _ms = ｛α _ms (A. qlen_
used [0] + β _ms qlen_unused
[0]) + γ _ms (A.mbuf_used + δ _ms A.m
buf_unused)｝ A. rtt / $ 1.1 ・
(A. if_baudrate-A.statisti
cs)｝ | S _A −S _C | _s = ｛α _s (A. qlen_used [1]
+ Β _s A. qlen_unused [1]) + γ _s (A.
mbuf_used + δ _s A. mbuf_unuse
d) + ε _s (A. memory_used + ζ _s A. me)
more_unused) ｅｄA. rtt / $ 1.1 ・
(A. if_baudrate-A.statisti
cs)｝ | S _A −S _C | _h = ｛α _h (A. qlen_used [2]
+ Β _h A. qlen_unused [2]) + γ _h (A.
mbuf_used + δ _h A. mbuf_unuse
d) + ε _h (A.disk_used + ζ _h A.disk
_Unused)｝ ・ A. rtt / ｛1.1 · (A.i
f_baudrate-A. (statistics)} In the above equation, all 1s represent one second, and are inserted in order to make each equation dimensionless. Α is an average packet size, β to ζ are real constants, γ and ε are positive, and β, δ, and 負 are negative.

Using the three types of metric values in ms, s, and h, a routing table is created for each. However, the metric value distributed to the other router is only the metric value of ms as described above, and the metric value distributed from the other router is only the metric value of ms. That is, the three metric values are only the metric values from the own router to the adjacent router, and the routing table based on the s and h metric values is used for detour.

That is, m distributed to other routers as a representative
Although the metric value of s is usually used as the optimal path, when the path is crowded, the packets in the queue of the order of s and h can be detoured by referring to the routing table based on the metric value of s and h. If there is, make a detour. At this time, the detour source router ID is described in the IPv6 relay option header shown in FIG.

The preceding is an explanation of the ninth and eighteenth aspects of the present invention.

[0126]

As described above, according to the present invention,
The following effects can be obtained.

That is, according to the first aspect of the present invention, individual communication packets can be route-controlled in an order according to a desired delivery time, and routing processing can be performed with high priority from high real-time data to low data. , Network resources can be used effectively.

According to the second aspect of the present invention, instead of the charging process based only on the transmission band and distance of the data link, fine charging for each packet is performed based on the moving distance of information on the space axis or the time axis. It can be performed.

According to the third aspect of the present invention, processing can be differentiated according to information attributes without terminating an upper layer, and network resources can be operated efficiently. In each router device, it is possible to know whether the delivery is delayed or advanced for each individual packet.

According to the fourth aspect of the present invention, time axis information can be added to a communication packet without changing the existing IPv6 packet format.

According to the fifth aspect of the present invention, it is possible to reduce packet discarding at the time of traffic congestion.

Further, according to the invention of claim 6, appropriate transfer processing corresponding to the remaining time until the desired delivery time can be performed, and delivery can be performed so as to satisfy the desired delivery time as much as possible.

Further, according to the seventh aspect of the invention, it is possible to eliminate an extreme difference in the number of waiting packets according to the remaining time until the desired delivery time, and to satisfy the desired delivery time as much as possible. Delivery becomes possible.

According to the invention of claim 8, the route can be changed, the communication packet is processed with the highest priority without waiting, and the communication packet exceeding the desired delivery time is directly waited. You can instruct them to dispose or discard.

According to the ninth aspect of the present invention, it is possible to perform adaptive routing even in a situation where the network is crowded, and it is possible to prevent the number of packets satisfying the desired delivery time from being significantly reduced.

Further, according to the tenth to eighteenth aspects of the present invention, it is possible to provide a router device capable of executing the above-described priority route control.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of a conventional router device.

FIG. 2 is a block diagram showing a basic configuration of a router device of the present invention.

FIG. 3 is an explanatory diagram of the principle of priority control of a communication packet based on a desired delivery time;

FIG. 4 is an explanatory diagram of elements of a space vector of information transmission.

FIG. 5 is an explanatory diagram of a storage example of time axis information in IPv6.

FIG. 6 is a configuration diagram showing an example of an embodiment of a router device of the present invention.

FIG. 7 is a detailed configuration diagram of an input / output unit in FIG. 6;

FIG. 8 is a detailed configuration diagram of a switch unit in FIG. 6;

FIG. 9 is a configuration diagram showing a main part of another example of the embodiment of the router device of the present invention;

FIG. 10 is a flowchart showing a queuing and order transition control algorithm when receiving a packet;

FIG. 11 is a flowchart showing a queuing and order transition control algorithm when receiving a packet;

FIG. 12 is a flowchart showing an extraction control algorithm at the time of packet transmission.

FIG. 13 is a flowchart illustrating an extraction control algorithm when transmitting a packet.

FIG. 14 is a flowchart showing an extraction control algorithm when transmitting a packet.

FIG. 15 is an explanatory diagram of another storage example of time axis information in IPv6.

FIG. 16 is an explanatory diagram of an example of storing exchange information with an adjacent router device in an LSA packet;

[Explanation of symbols]

21: input port, 22: switch unit, 23: output port, 24: route control unit, 25: priority control unit, 26: communication packet, 26a: destination address, 26b: desired delivery time, 27: space switch, 28: Time switch, 31
A, 31B, 31C, 31D: input / output unit, 32: disk array PC, 33: switch (SW) unit, 34: control PC, 35: backplane, 36: disk array, 311a: optical input unit (Opt in), 311b:
Optical output unit (Opt out), 312: Serial / parallel converter (S / P), 313: Data link layer termination unit, 314: FPGA, 315: Signal processor (DSP), 316a to 316c: High-speed memory, 31
7: associative memory (CAM), 318: synchronous (s)
ync. ) FIFO memory, 319: interface unit (I / F), 331: crossbar switch, 332: control CPU, 41: order transition control unit, 42: takeout control unit.

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Kazuaki Obana 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo F-term in Nippon Telegraph and Telephone Corporation (reference) 5K030 GA03 HC01 HD03 LB05 9A001 JJ18 KK56

Claims (18)

[Claims] 1. A route control method for switching and transmitting a route of a communication packet based on a destination address labeled for each communication packet, wherein a desired delivery time is labeled together with the destination address, and the desired delivery time is assigned to the desired delivery time. A priority routing control method comprising: determining a priority of packet processing based on the priority; and performing routing control. 2. A metric value between a sender and a receiver determined from a source address and a destination address of a communication packet, or a remaining time until a desired delivery time obtained from a transmission time and a desired delivery time of a communication packet, or a metric value and a delivery. 2. The priority route control method according to claim 1, wherein a communication fee is determined using a value obtained from the remaining time until the desired time or a combination thereof as a parameter. 3. A metric value from the router device to the destination obtained from the address of the router device and the destination address of the communication packet, or a remaining time until the desired delivery time obtained from the current time and the desired delivery time, or a metric value and delivery. 2. The priority route control method according to claim 1, wherein a priority determined in packet processing is determined by using a value obtained from the remaining time until a desired time or a combination thereof as a parameter, and route control is performed. 4. Internet Protocol version 6
4. The priority route control method according to claim 1, wherein a communication packet created by storing a transmission time and a desired delivery time of the communication packet in a relay point option header of the communication packet in the specification is used. 5. When congestion occurs due to a communication packet having a processing capacity or more, a communication packet having a low priority is temporarily saved in a storage device in the router device, and the communication packet being saved when the desired delivery time is approaching. 5. The priority route control method according to claim 1, wherein when the priority rises or when the congestion stops, a communication packet is read from the storage device and the route control is restarted. 6. Waiting for a communication packet for each unit of the remaining time until the desired delivery time, and performing transfer processing with a different probability for each unit, when the unit of the remaining time until the desired delivery time changes during the waiting. 6. The priority path control method according to claim 1, wherein a unit of the waiting is changed. 7. When a communication packet is queued for each unit of the remaining time until the desired delivery time and transfer processing is performed with a different probability for each unit, the transfer probability for each unit is determined by the packet waiting for each unit. 7. The priority route control method according to claim 1, wherein the priority route control method is changed according to the number. 8. Internet Protocol version 6
Communication created by storing in the relay option header of the communication packet in the specification the ID of the detour source router device, the priority unrelated to the desired delivery time, and the handling when the desired delivery time is exceeded, in the relay option header of the communication packet in the specification. 8. The priority route control method according to claim 1, wherein a packet is used. 9. A packet storing information such as the amount of use and free space of a storage device in a router device, the state of use of a link, the number of adjacent routers on a link, and the time required for a round trip to an adjacent router. The metric values from the router device to the destination are dynamically changed using the obtained metric values. The priority route control method according to claim 1. 10. A plurality of input ports and a plurality of output ports, a spatial switch capable of arbitrarily switching connection between a plurality of input ports and a plurality of output ports, and a communication packet input from the input port as a destination address. A router device having a route control unit for controlling a spatial switch to send to a corresponding output port, wherein a time switch capable of arbitrarily changing the order of a plurality of communication packets, and the order of the plurality of communication packets A router device, comprising: a priority control unit that controls a time switch so that the order corresponds to the assigned desired delivery time. 11. A metric value between a sender and a receiver determined from a source address and a destination address of a communication packet,
Alternatively, the communication fee is determined using the remaining time until the desired delivery time obtained from the transmission time and the desired delivery time of the communication packet, or the value obtained from the metric value and the remaining time until the desired delivery time, or a combination thereof, as a parameter. The router device according to claim 10, further comprising: 12. A metric value from the own device to the destination obtained from the address of the own device and the destination address of the communication packet, or the remaining time until the desired delivery time obtained from the current time and the desired delivery time, or a metric value and the delivery The value obtained from the remaining time until the desired time,
11. The router device according to claim 10, further comprising a priority control unit that determines a priority of packet processing using the combination thereof as a parameter and controls a time switch. 13. The communication apparatus according to claim 10, further comprising a priority control unit that uses a transmission time and a desired delivery time of the communication packet stored in a relay point option header of the communication packet in the Internet Protocol version 6 specification. The router device according to any of the above. 14. A storage device for temporarily storing communication packets, and, when a communication packet having a processing capacity or more is input and congestion occurs, a communication packet having a low priority is temporarily evacuated to the storage device and transmitted. Control means for reading out the communication packet from the storage device and resuming the priority route control when the priority of the evacuation communication packet is increased due to the desired time approaching or when the congestion is reduced. The router device according to any one of claims 10 to 13. 15. A time switch divided for each unit of remaining time until a desired delivery time, and a communication packet in one time switch in which the unit of remaining time until the desired delivery time has changed during a wait is changed. 15. The router device according to claim 10, further comprising: means for shifting to a time switch corresponding to the unit of (1). 16. A time switch divided for each unit of the remaining time until the desired delivery time, and a probability of taking out a communication packet in the time switch for each unit is determined according to the number of packets waiting in the time switch for each unit. 16. The router device according to claim 10, further comprising: 17. The I of the detour source router device stored in the relay point option header of the communication packet in the Internet Protocol version 6 specification.
17. The router device according to claim 10, further comprising: D, a priority control unit that uses a priority irrelevant to the desired delivery time and handles the case of exceeding the desired delivery time. 18. A packet storing information such as the amount of use and free space of a storage device in a router device, the use state of a link, the number of adjacent routers on a link, and the time required for a round trip to an adjacent router. Means for exchanging between adjacent router devices, means for calculating a metric value to the adjacent router device based on the information in the packet, and a metric from the router device to the destination using the obtained metric value. 18. The router device according to claim 10, further comprising means for dynamically changing a value.