JPH0969852A - Band reserving system for network - Google Patents

Abstract

PROBLEM TO BE SOLVED: To transmit all protocol frames at a fixed rate without concentrating processing at a priority frame by allocating bands to queues provided according to protocols in a prescribed ratio and successively repeatedly performing frame processing in a network with which repetition is performed while using plural protocols. SOLUTION: When transmitting the frame received from an 'Ethernet<(> R<)> ' 101 to a leased line 102 at a multiprotocol router having the queue of frames, the router provides the queues as many as the number of networks to be used at a buffer memory 37 and allocates the rate of band usage in the prescribed ratio. Then, the frames received from the 'Ethernet<(> R<)> ' 101 are divided according to protocols and respectively distributed to the queues, the allocated frames are successively transmitted according to the rate of band usage and at the point in time when the transmission procedure completes one cycle, the transmission is returned to the first queue and repeatedly performed. Thus evey protocols can be transmitted at a fixed ratio.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-protocol router for relaying between networks by a plurality of protocols, and more particularly to a network bandwidth reservation system capable of transmitting any protocol frame at a constant rate.

[0002]

2. Description of the Related Art The prior art will be described. In the network system shown in FIG. 3, the Ethernet 101 and the Ethernet 120 of 10 Mbit / sec are 64 kbit / sec.
Connected via the sec high-speed digital dedicated line 102,
It is a network system that relays traffic such as TCP / IP, IPX, and SNA by routers 103 and 123.

As shown in FIG. 4, the router 103 includes an Ethernet interface circuit 30, a dedicated line interface circuit 31, internal buses 32, 33 and 34, and a router operation control section 36 including a microprocessor and a memory.
It is composed of a buffer memory 37 for storing transmission / reception frames, signal lines 38, 39, and the like.

Consider a case where the router 103 relays a packet flowing through the Ethernet 101 to the leased line 102. When a frame received from the Ethernet 101 side is relayed to the dedicated line 102 side, the transmission rate of the dedicated line 102 is 64 kbit / sec, which is slower than the transmission rate of 10 Mbit / sec of the Ethernet 101.
When a plurality of frames to be relayed arrive on the 01 side in succession, it is necessary to temporarily store these frames in the buffer memory 37 in the router 103. Router 10
Within 3, these frames are held in a queue called a queue. This queue is, for example, FIG.
It is realized by a list structure in which a series of memory blocks called a frame descriptor are connected by a pointer. The router 103 holds a pointer indicating the beginning of the frame stored inside, and this pointer points to the beginning frame descriptor.
The frame descriptor is composed of a pointer (next pointer) that points to the next frame descriptor, a frame length, a pointer that indicates the memory address where the actual transmission frame or the reception frame is stored, and the like. The frame descriptor that indicates the last frame is the next. As a pointer
Set a special value that can identify the end. Here, by setting 0, the end is shown. The frame is added to the queue by rewriting the next pointer of the frame descriptor set to 0 to the address of the descriptor indicating the frame to be newly added, and setting the next pointer of this descriptor to 0. Further, when a frame is taken out from the queue, the value of the pointer indicating the beginning is rewritten to the value of the next pointer of the leading frame descriptor.

In a simple router, all frames are managed by one queue regardless of protocols such as TCP / IP, IPX and SNA. In this case, T
When another protocol frame such as SNA is registered in the queue while the CP / IP or IPX frame is already registered in the queue, this SNA frame is processed by the other protocol frame previously registered or all of them are processed. It is processed by waiting for it to be done. Here, when the number of other protocol frames registered before the SNA frame is large, this SNA frame is not processed easily and the relay delay time becomes long. If such a frame is relayed by multiple routers before reaching the final destination, this relay delay time occurs in each router, and in the worst case, the connection between end terminals times out. There was a problem.

In order to solve this problem, there is a router in which a plurality of queues are provided, the user specifies the priority for each protocol, and the frames can be managed by the queue for each priority. For example, priority levels 1, 2, and 3 can be designated for each protocol, and three queues are prepared. It is assumed that 1 is the highest priority and 3 is the lowest priority. For example, priority 1 is assigned to SNA, priority 2 is assigned to TCP / IP, and priority 3 is assigned to IPX. When relaying the frame received from the Ethernet to the dedicated line, the queue is checked according to the priority, and when all the frames registered in the queue are transmitted, the next queue is checked. That is, first, the queue of priority 1 is checked, all the registered frames are transmitted, then the queue of priority 2 is checked, and the registered frames are transmitted. Finally, the queue of priority 3 is checked and the registered frame is transmitted.

[0007]

However, in such a router, if a new frame is registered in the priority 1 queue while processing a priority 2 queue, the router returns to this queue to perform the processing. To do. For this reason, when a frame constantly arrives at the priority 1 or priority 2 queue, the frame registered in the priority 3 queue is not processed.

SUMMARY OF THE INVENTION It is therefore an object of the present invention to solve the above problems and provide a bandwidth reservation system for a network that can transmit every protocol frame at a constant rate.

[0009]

To achieve the above object, the present invention provides a multi-protocol router having a queue as a queue of frames for relaying between networks using a plurality of protocols. A queue is provided for each protocol, and the ratio of the available network bandwidth is assigned to each queue. Frames are processed in order from the queue with the largest allocated ratio, and when there are no more frames in that queue, the processing moves to the next queue. It processes up to the last queue and repeats the processing from the first queue again.

The predetermined unit time is divided by the ratio of the available bandwidth allocated to each of the queues, and the time occupied by the queue having the first processing order is used as the reference time, and the first queue is the reference time. Processing may be performed only from the next queue to the last queue for a time represented by the product of the bandwidth usage rate of the queue with respect to the first queue and the reference time.

If the reference time is not reached even when the frames in the queue become empty during the processing of the first queue, the time until the first queue becomes empty may be updated to the reference time. .

Each queue has a variable consisting of the sum of the sizes of the already transmitted frames, this variable is set to 0 before the processing of each queue is started, and this variable is set every time a frame is transmitted during the processing. The size of the frame is added to this variable, and if this variable exceeds the product of the network bandwidth and the ratio assigned to this queue, the remaining frames in that queue are not transmitted and processing is moved to the next queue. May be.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION According to the present invention, a plurality of queues are prepared in a router, and a packet (frame) received in each queue is divided according to a protocol and stored. When the user sets for each protocol and transmits a frame to the network, the process is performed in order from the queue with the largest assigned ratio, and when there are no more packets stored in the queue, the next queue is processed. After processing the last queue, return to the first queue and repeat the same processing. This process controls the number of frames to be transmitted by changing the frame transmission time depending on the queue. When a predetermined unit time, for example, 1 second, is divided at a rate assigned to each queue, the value occupied by the first queue is used as the reference time, and the frame of the queue to be processed first is sent out only for this reference time. Further, when all the frames in the queue are transmitted within the reference time, the transmission time until then is set as the reference time. The other queues transmit the frames in the queue for the time represented by the product of the bandwidth utilization rate of the queue with respect to the first queue and this reference time. This allows the network to be used according to the bandwidth reserved by the user. It also holds the sum of the frame sizes that have already been transmitted when processing a queue, and if this value exceeds the product of the network bandwidth and the available bandwidth allocated to that queue, there are remaining frames. Do not send, but process the next queue.

An embodiment of the present invention will be described with reference to the drawings.

A flow chart of the method of the present invention is shown in FIG. Here, n queues are provided for n types of protocols, and the ratio of usable network bandwidth is assigned to each queue. The procedure of FIG. 1 is such that all protocols are processed at a constant rate according to the rate assigned to each protocol.

In the network system here, as shown in FIG. 3, a router 103 is provided on the Ethernet 101 having terminals 111, 112, 113, and a router 123 is provided on the Ethernet 121 having terminals 121, 122. Are provided, and these router 103 and router 1
It is connected to 23 by a dedicated line 102. Below, Router 1
A case where 03 transmits the frame received from the Ethernet 101 side to the dedicated line 102 side will be described as an example.

The router 103, as shown in FIG.
A router operation control unit 3 including an Ethernet interface circuit 30, a dedicated line interface circuit 31, internal buses 32, 33 and 34, a microprocessor and a memory.
6, a buffer memory 37 for storing transmission / reception frames, signal lines 38, 39, etc.

In this embodiment, the router 103 is an SNA,
It is assumed that TCP / IP and IPX protocols are supported. Then, the bandwidth usage rate is reserved for each protocol. That is, 50% of bandwidth in SNA protocol, TCP /
30% of the bandwidth is allocated to the IP protocol and 20% of the bandwidth is allocated to the IPX protocol. Therefore, prepare three queues in the buffer memory, and use queue 1 for the SNA protocol.
, The queue 2 is used for the TCP / IP protocol, and the queue 3 is used for the IPX protocol. As a result, a queue is provided for each protocol and the network bandwidth that can be used for each queue is allocated at a ratio of 5: 3: 2.

The frame received by the router 103 from the Ethernet 101 is divided by protocol and is distributed to the above queues. Here, the frame of the SNA protocol assigned to the queue 1 is directly used for the dedicated line 10
The transmission time is set to the reference time t1. In this case, t1 is a product of 1 second and 50% so that the reserved bandwidth usage rate is obtained every unit time, for example, every 1 second.
00 ms is given. The time is the router operation control unit 3
Measured by the internal clock at 6. If queue 1 becomes empty within the reference time, transmission will stop at that point,
The time T from the start of transmission to the stop is substituted for t1. If the packet still exists in the queue 1, t1 is held as it is.

Since the line speed of the leased line is much lower than that of the Ethernet, the frames waiting to be sent are stored in the queues 2 and 3 at this point. Regarding the next TCP / IP packet, it is transmitted only for 3/5 of the previously obtained reference time t1. This three-fifth is the bandwidth usage rate of queue 2 with respect to queue 1. For example, if an SNA packet is sent for 100 ms, the TCP / IP packet will be 6
Send for 0 ms. Similarly, the IPX packet has a reference time t1.
3/5 time of transmission. When the processing of the queue 3 is completed, the process returns to the queue 1 and repeats relaying. With this method, the SNA protocol of queue 1 and the TCP of queue 2
/ IP protocol and queue 3 protocol IPX
The ratio of the amount of data sent per unit time is 5: 3:
2, and the bandwidth of the transmission path used is also a ratio of 5: 3: 2. If the packets in the queues 2 and 3 become empty during transmission, the transmission path is immediately used and the transmission is immediately moved to the next queue to continue the transmission.

Another embodiment according to the present invention will be described.

The flow chart here is as shown in FIG. The network system and the router are the same as those in the above-described embodiment shown in FIGS. 3 and 4, and the allocation for each protocol is also the same.
A case where the router 103 transmits a frame received from the Ethernet 101 side to the private line 102 side will be described as an example.

The bandwidth of the private line 102 is set to 64 kbit / se
c, the amount of data that can be transmitted per unit time in queue 1 is

[0024]

[Equation 1]

[0025] Similarly, the amount of data that can be transmitted in the queue 2 per unit time is

[0026]

[Equation 2]

The amount of data that can be transmitted in the queue 3 per unit time is

[0028]

(Equation 3)

It becomes This is called the allowable data amount of each queue. The router holds the sum of the sizes of the transmitted frames in the queue being processed as a variable, and sets this variable to 0 before starting the processing for each queue.
Then, the processing is started from the queue 1, and every time one frame is transmitted, the size of the transmitted frame is added to this variable and this is compared with the allowable data amount (4 kbytes) of the queue. When the sum of the transmitted data amounts stored in the variables does not exceed the allowable data amount, the data registered in the queue 1 is continuously transmitted. The sum of the amount of transmitted data stored in the variable is
If the allowable data amount is exceeded, or if there is no transmission data registered in the queue, the processing of queue 1 is ended and the processing of queue 2 is performed. For example, queue 1 to 256
If 20 bytes of data are registered, 16
After transmitting the number of frames, the process moves to the process of the queue 2 while leaving the four transmission data.

Similarly, the amount of transmitted data is set to 0 before the processing of the queue 2 is started, and the same processing as that of the queue 1 is performed.
When the sum of the data transmitted for the queue 2 exceeds the permissible data amount of the queue 2 (2.4 kbytes) or there is no data to be transmitted in the queue 2, the process of the queue 3 starts.

Similarly, before the processing of the queue 3 is started, the transmitted data amount is set to 0 and the same processing is performed. When the sum of the data transmitted for the queue 3 exceeds the permissible data amount of the queue 3 (2.4 kbytes) or the transmission data of the queue 3 is exhausted, the process of the queue 1 is performed again.

According to the above embodiment, all the protocols can be processed at a fixed rate according to the bandwidth allocated to each protocol.

[0033]

The present invention exhibits the following excellent effects.

(1) Any protocol frame can be transmitted at a fixed rate, and there is no possibility that only a specific protocol frame will not be processed.

[Brief description of drawings]

FIG. 1 is a flowchart according to an embodiment of a bandwidth reservation system of the present invention.

FIG. 2 is a flowchart according to another embodiment of the bandwidth reservation method of the present invention.

FIG. 3 is a configuration diagram showing a configuration example of a network system to which the present invention is applied.

FIG. 4 is a configuration diagram showing a configuration example of a router to which the present invention is applied.

FIG. 5 is a configuration diagram showing an implementation example of a queue in a router.

[Explanation of symbols]

101, 120 Ethernet 102 dedicated line 103, 123 router 111, 112, 113, 121, 122 Ethernet terminal 30 Ethernet interface circuit 31 dedicated line interface circuit 36 router operation control unit 37 buffer memory

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code Agency reference number FI Technical display location H04L 29/06

Claims (4)

[Claims] 1. In a multi-protocol router having a queue, which is a queue of frames for relaying between networks using a plurality of protocols, the queue is provided for each protocol and the bandwidth of the network that can be used for each queue is set. A percentage is assigned, frames are processed in order from the queue with the highest percentage assigned, and when there are no more frames in that queue, the process is moved to the next queue, processed to the last queue, and again from the first queue. A network bandwidth reservation method characterized by repetition. 2. A predetermined unit time is divided by a ratio of available bandwidth allocated to each of the queues, of which the time occupied by the queue having the first processing order is used as a reference time, and the first queue is Process only the reference time,
2. The bandwidth reservation method for a network according to claim 1, wherein processing is performed from the next queue to the last queue for a time represented by a product of a bandwidth usage rate of the queue with respect to the first queue and the reference time. 3. If during the processing of the first queue the frame in the queue becomes empty but the reference time is not reached,
The network bandwidth reservation method according to claim 2, wherein the time until the first queue becomes empty is updated to the reference time. 4. Each of the queues is provided with a variable consisting of the sum of the sizes of already transmitted frames, the variable is set to 0 before the processing of each queue is started, and each frame is transmitted during the processing. And add the size of this frame to this variable,
When the variable exceeds the product of the bandwidth of the network and the ratio assigned to the queue, the remaining frames of the queue are not transmitted and the process is moved to the next queue. Network bandwidth reservation method.