JPS62104245A - Data flow control system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to control of data flow between nodes of a data communication network.

(Conventional Technique? If) 2sil data and evening terminals each connected to two different nodes of a data communications network, in one typical configuration, are connected to a so-called virtual circuit data path that runs between 21 nodes. Say it all,'! ! Ta. Communication is accomplished through a series of intermediate nodes and physical links that complete the connection between the two data terminals. Each node or link handles a large number of paths. Paths are multiplexed on the links. Each node has a buffer that temporarily stores incoming data from one different path. The data is then sent out to the respective outgoing link or provided to a data terminal handled by that node. The average speed at which incoming data is received is
If the buffer does not overflow and data is not lost, the rate of outgoing data cannot be outpaced for a very long time.

Data flow in these data communication networks? One method of control is called an ARQ window. In such an arrangement, the node that takes all data over the path from the downstream node must periodically signal, or acknowledge, the receipt of a data block before the upstream node transmits it further. A maximum number of unacknowledged frames (ARQ window) can be on the path at the same time. A receiving node can adjust the amount of data it receives over a given path by acknowledging or not acknowledging incoming frames. In a type of communication network where a stream is sent from an upstream node to a downstream node via multiple data paths, the data flow can be controlled by having the downstream node specify the maximum average speed of the data stream on each path. The goal is to control.

Therefore, each node can directly dictate the rate of the data stream on the incoming data path,
Data loss is reduced.

Preferred embodiments of the invention include the following features. The downstream node temporarily stores the data stream 7171
means for evaluating whether the buffer means is currently in use and specifying a rate based thereon. A decision is made as to when a change in speed should be considered and the data path on which the speed should be changed is identified, thus making it possible to make the necessary changes in the overall incoming data g in the optimal path. Determining when the speed change itself should be considered depends on the usage state of the buffer and the length of time that has elapsed since the speed change was considered. That is, speed changes are only considered as frequently as is appropriate given how heavily the buffer is being used. Buffer means refers to a plurality of buffers, and buffer usage status is a simple and very appropriate criterion. Evaluated as percentage of buffer in use. Speed increases 710 or decreases are considered separately based on buffer usage and the length of time that has elapsed since the speed increase or decrease was considered. thus.

At the same time, the speed of one path is fully increased, and the speed of other paths is reduced, making it possible to efficiently handle the buffer usage status. An increase in speed M is taken into account more than a predetermined amount of time after a previous consideration of an increase or decrease in speed occurred. (The size of the predetermined time length varies directly with the buffer usage state in the case of a decision that considers an increase in speed, and changes inversely to the buffer usage state in the case of a decision to decrease the speed.) ) That is, speed increases are sometimes considered when buffer usage is high.

Speed reduction is sometimes considered when buffer summer conditions are low. Based on whether the ratio of the previously specified path's speed to the so-called ideal speed exceeds a high or low threshold, respectively, which varies with the buffer 7 usage status, the data path is subject to a speed increase or decrease. selected. Each threshold value increases as the condition of the pants becomes darker, so that the higher the buffer usage status becomes.

In order to start changing the speed, the ideal speed must be
The previously specified speed must be less. The threshold for a 4 degree decrease never exceeds 1. The threshold for speed increase is never less than one. That is, the speed decreases if its latest specified speed does not exceed the ideal speed.

It is never initiated against a path.

The new speed for the pass is equal to the ideal speed plus a factor not less than 1 gold, and the factor is a spit that changes inversely to the buff 7 usage state. As a result, if the buffer usage is small, the new rate is greater than the ideal rate. The ideal speed is the previous speed set for the path by further downstream nodes and the maximum available bandwidth for the path.

It is determined based on the usage status of buffers placed on a given path. If the buffer usage exceeds two minutes' worth of data bytes at the unadjusted ideal speed, the ideal speed decreases. The downstream node communicates the speed to the upstream node, thus allowing the downstream node to effectively control its own buffer usage.

Other advantages and features will be apparent from the following description of the preferred implementations and from the claims.

(Example of implementation) In FIG. 1, a data communication network 10 includes a data link 2
It includes nodes 12.14.16.18.20 connected to each other by 2.24.26.2&,30. Each node acts as one or more data terminals 62. Data flows from one terminal to another via a predefined virtual circuit data path. The data path refers to the coupling of nodes acting as two terminals and the intermediate links necessary to complete the physical connection between these two terminals, such as terminal 66. From terminal 65 to node 18. Link 26. Node 16. A path is established via link 28 node 20t. Datapath companies may be created or abolished as needed.

Multiple data paths are multiplexed on each oscilloscope data link.

In FIG. 2, to achieve this multiplexing.

Each link carries a series of scans. Each scan is a series of 8-bit bytes arranged in fields.

The field contains a scan initiator fieldro 4 which uniquely specifies the start of a scan. The remaining fields in the scan are delta 36 and slot 68.
It switches between two types called . Each scan automatically fills a slot 38 for each path currently configured on that link. In sequence, slot 1 of successive scans can create a particular path on the link and carry a stream of data bytes.

Each slot has a variable number of data bytes 2, so that
The relative maximum average rate that can be carried on the data bytes of the different paths is determined by the relative maximum number of data bytes in each slot (referred to as the slot weight). The absolute speed at which the data stream can be carried on the path also depends on the total length of each scan and the bit rate of the link. The rate at which various paths can carry data money can be varied by upstream nodes that reshape the relative slot weights.

Although a given path has a certain speed t at a particular i time, no terminal upstream of the path may be using the path at that one time when a scan is assembled for transmission. , if the configured path is not used, the slot corresponding to the path has no bandwidth? Omitted from scan to avoid large size.

@ In Figure 4, S/R 40 is connected to link port 44 and I
It is connected to node 42 via lo buffer 46. The scans are carried over link 4o using the HDLCI link layer protocol. The HDLC protocol is transparent to buffer 46 because the protocol is handled by link port 44 . The HDLC protocol only handles scan bytes, buffer 46.
acts as a temporary storage buffer for input and output data bytes. The node 42 has a buffer 46
The input/output processor (I
OP) 46, if the data byte of a slot, then the data pie If the data byte is a 62-byte long FIFO data buffer 48 allocated to the path corresponding to that slot.
Put it in.

2000, the fixed FIFO data buffers 48 are assembled to create a node-buffer pool 49 of finite capacity. A data buffer 148 is allocated to each pass request. If all buffers allocated to a given path are full when another data byte arrives on that path.

When another buffer is available, that buffer is assigned to the path. The addresses of buffers currently allocated to a path are maintained as a linked list.

IOPs periodically remove data bytes waiting in their respective data buffers. In the case of a data byte to be sent to another node, the IOP inserts this data byte into the appropriate slot of the scan and selects this scan? to the appropriate link port (only one of the multiple link ports is shown in FIG. 6). Data bytes to be sent to a terminal 50 handled by a node 42 are passed through a buffer 52 and a terminal port 54 which handles the protocol used between the node and the terminal. When the removal of data π)k from a given buffer 48 results in the buffer becoming free, the buffer is freed for allocation to another pass.

The flow of data bytes from the upstream node into node 42 via the link is regulated in a first-order manner to make effective use of the available capacity of buffer 48 without unnecessarily discarding data.

Referring to FIG. 4, the l0P46I/'i has a data buffer usage west monitor 58. The moni p58
The node buffer 48 is looked at every H2 seconds and the buffer usage status is determined as the percentage of the buffer 48 that is currently not empty. Based in part on this determination, the acceleration/deceleration evaluation 460
1 determines whether to examine individual paths and the number of paths for which incoming data bytes from upstream nodes should be accelerated, slowed down, or do nothing. Increase or decrease timer 62 and deceleration timer 64 each indicate the length of time since the last time an increase or decrease in the rate of incoming data on any path was considered? Chase.

The maximum interval table 66 in the evaluator 60 shows the maximum time that will elapse on both timers 62, 64 before speeding up or slowing down is considered.

represents the percentage of buffers 48 (not empty).

Referring to FIG. 5, the maximum interval table 66 includes maximum intervals for sequential ranges of 10 coefficients of buffer utilization that are used to analyze paths to determine whether an increase is a deceleration. 2 elapse on each timer 62.64 without.

Assuming a buffer usage rate of 65, an analysis is performed for each pass as to whether the incoming data rate in bytes per second should be increased when the value of the speed up timer exceeds 6 seconds. If both timers exceed the interval in their respective columns of table 66, both analyzes are done for each pass. The maximum interval for a speed-up timer increases as the buffer usage percentage increases, but the opposite is true for a speed-down timer. So the buffer:
- If the vehicle is full for one hour, the frequency of considering speed increase is relatively low, and the frequency of considering deceleration is relatively high.

Referring again to FIG. 4, once the decision to evaluate the path is made, the evaluator 60 uses the path analyzer 65kl.
I'm going to do J. Therefore, the path analyzer 65 considers each path for speed increase or deceleration based on the ideal speed of the path and the flow control wholesale table 68.

The path analyzer 65 includes an ideal velocity calculator 70.

Calculation 470 calculates the current specified (current) speed and maximum bandwidth buffer 77 for the path found in current speed buffer 71.
11111J derived from 5 and available for the path
Ideal speed for each path based on maximum bandwidth? calculate.

Calculator 70 is one of the two formulas for finding the ideal fourth degree.
applicable. The decision as to which formula to use is made based on whether more than two seconds worth of output data for that path is currently buffered. The ideal speed calculator 7U monitors the usage status of the buffer 48 of each path. While the path buffer 48 holds the output data below 2 seconds worth of μ at the unadjusted ideal speed, the ideal speed is updated to the maximum speed 1 (lower #, the highest speed at which data r can be transmitted to the measurement node). That is, the minimum value t! of ff1lf:, which is set before the ideal speed, is equal to the large bandwidth of the mouth.

Once more than 2 seconds worth of output data has been accumulated for a path (at the unadjusted ideal speed), the faster the node 42 retransmits it to a downstream node or provides it to the terminal 50, the faster the data is received. be done. Therefore, it is desirable to reduce the rate at which data reaches node 42 on this path by ik50<. In that case, ideally, the desired If67 is updated to be equal to the minimum value of the current rate, ie, 1/2 of the maximum bandwidth.

Once the ideal speed 67 is updated, the speed ratio 7 is calculated by the ratio of the fc ideal speed to the latest speed 72.
The calculator 70 calculates 6. When this ratio is less than 1.0, it indicates that data can be received and taken from the upstream node at a faster rate. If this ratio is greater than 1.0, it indicates that the rate at which one piece of data arrives is too high.

Referring to FIG. 6, the flow control table 68 is a static table with a speed ratio 76 and a buffer usage ratio 7.
7 and specifies whether the path associated with a given speed ratio should be selected for acceleration or deceleration, and if it is selected, a new input for that path arrives at 6 degrees. Gives a modification factor by which the imaginary velocity of this given path is multiplied to make it. Two columns 101 and 102 contain threshold ratios for use in making speed up or deceleration decisions. The lower threshold 100 is used when selecting a path to deceleration. For a given buffer usage percentage, the speed ratio 76 is equal to the corresponding lower threshold j[1
If it is less than 00.

The path is selected, and if it is large, the path is not selected. As the percentage of parafluid filling decreases,
The speed ratio must become progressively smaller to trigger an input data flow reduction; for example, a speed ratio of 0.9 will cause full selection of paths for deceleration if the paraphrase is used.

However, if only 8 buffers are used, will the speed ratio be reduced? Must be less than 0.05 to trigger.

Upper threshold 101 is applied when a path to be accelerated is selected. If a given speed ratio is greater than an upper threshold indicated for the current buffer usage percentage, then that path is selected. As the proportion for one box of buff increases.

The speed ratio must be significantly greater than 1.0 for the path to be selected.

Referring again to FIG. 4, to determine the appropriate modification factor, calculator 70 calculates the speed ratio 76? Applies to flow control table storage device 68. The memory device 68 then conveys to the new speed calculator 110 the value of zero or the indicated change factor a. The change factor indicates that the path currently being evaluated has been selected for speed change. A value of zero indicates that no path was selected. When the change coefficient a is received, this coefficient is multiplied by the ideal speed 67 to obtain the new speed 1
02 is found.

After a final test on each path, the path is selected to ensure it is active. The f-p buffer dipping cap monitor 58 is on another shift.

i.e. 60 bytes transmitted on the path and expressed in 7 seconds.
Gives an average of 104 seconds. A non-zero mean of 104 indicates that the path is active.

The new velocity calculator 110 examines the average 104 and, if the average 104 is not zero, selects the path by distributing the path identifier and the new velocity to the set velocity assembler 120.

In this way, valley paths are evaluated and a new speed is calculated for one selected path. These new rates, which are the input data rates for each pass, are communicated to upstream nodes by the assembler 120. The upstream node adjusts this new rate to the slot weights, ie.

This is done by adjusting the amount of data bytes contained in each pass slot to a large number.

To convey the desired transmission speed to the upstream node.

The set speed assembler 120 assembles the desired speed into a special message called the set speed.

In FIG. 7, set speed 122 includes a path identifier (path ID) 124 for each selected path and a new speed 102 for that path. Each set speed is decomposed into a field in which there are two types:
Path IDs 124 occupying the "2nd" type field and new IDs 2102 occupying the 2nd type field are arranged alternately.

The set rate is sent to the upstream node via a control path carried on the link.

Each upstream node receives the set speed IJj1
22 completely disassembled and its current speed (ha I fufu 71
stores the new speed and reshapes its outgoing scan.

Other implementations are within the scope of the claims.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a data communication network. FIG. 2 is a scan format diagram. FIG. 6C is a block diagram of a node. FIG. 4 is a block diagram of the input/output processor of the node. FIG. 5 is a table of maximum time intervals and cancers for various buffer usage states. FIG. 6 is a flow control table. FIG. 7 is a format diagram of a set speed message. 10...Data communication network, 12.14,16,18゜2
0... Node, 22, 24, 26.28.30...
Data link, 32.33.35...data terminal. 46... Input/output processor, 48'... Buffer. 49... FIFO data buffer, 58... Buffer usage status monitor, 60... Speed increase/deceleration evaluator, 62
...Speed up timer, 64-...Deceleration timer, 65...
Path analyzer, 66... Maximum interval table, 68...
Flow control table storage device, 70...Ideal speed calculator. 71...Current speed buffer, 75...Maximum bandwidth buffer 2 buffer, 110...New speed calculator, 122...
・Set speed.・-Ko agent Patent attorney Kyo Yuasa (5 people including Shikaku) FIG 5 Saitomo Iihaku j-pull / IG 6 Flow control tape, Shi / 77 too
80 101 Procedural Amendment, Patent Application No. 219109 of 1985 2, Name of the Invention T-Taflo Control System 3, Relationship with Person Making the Amendment 1'1 18' [Applicant Address Name Name Codex Corporation 4 , Agent U Address: Room 206, Room 5, Shin-Otemachi Building, 2-2-1 Hitemachi, Chiyoda-ku, Tokyo Subject of amendment

Claims (1)

[Claims] 1. A system for controlling the flow of data in a communication network in which a plurality of data streams are sent from an upstream node to a downstream node via a plurality of paths, each path comprising: A system according to claim 1, comprising designation means for said downstream node to designate a maximum average speed of said data stream on said downstream node. 2. The method according to claim 1, further comprising buffer means in the downstream node for temporarily storing the data stream, and wherein the maximum average speed is specified based on the usage state of the buffer means. system. 3. The designating means comprises determining means for determining when a change in the maximum average speed is taken into account, and identifying means for identifying the data path whose speed is to be changed. The system described in Scope 1. 4. The system of claim 3, wherein said determining means makes said determination based on said buffer usage and the length of time that has elapsed since the speed change was previously considered. 5. The system according to claim 2, wherein the buffer means includes a plurality of buffers, and the designation means includes measurement means for measuring the buffer usage status as a percentage of the buffers in use. 6. means for determining when a speed increase should be considered based on the buffer usage state and the length of time elapsed since the maximum average speed speed increase was previously considered; 4. The system of claim 3, comprising means for determining when deceleration should be considered based on usage conditions and the length of time that has elapsed since deceleration of the maximum average speed was previously considered. . 7. An increase or decrease in said maximum average speed is taken into account when the previously considered increase or decrease in speed, respectively, occurs at a time earlier than a predetermined length of time, the magnitude of which depends on said buffer usage state. , the system according to claim 6. 8. Claims in which the predetermined time length is longer than in a high buffer usage state in the case of a decision to consider speed increase, and shorter than in a high buffer usage state in the case of a decision to consider deceleration. The system according to paragraph 7. 9. The identifying means includes comparing means for comparing a previously specified speed of each of the paths with an ideal speed of the path, and determining a data path for speed change based on this comparison and the buffer usage state. 4. A system according to claim 3, comprising selection means for selecting. 10. The system of claim 9, wherein the comparison is a ratio, and the path is selected for speed change if the ratio exceeds a threshold that varies with usage of the buffer. . 11. The system of claim 10, wherein the threshold value increases with increasing usage. 12. The system of claim 10, wherein there is a first threshold value not exceeding a unit value for deceleration and a second threshold value not exceeding a unit value for speed increase. 13. The system of claim 9, wherein said designating means further comprises establishing means for establishing a new speed equal to said ideal speed and a factor not less than a unit value. 14. The system of claim 9, wherein said coefficient is dependent on buffer usage, with higher said coefficients being associated with lower said usage. 15. The comparing means determines the ideal speed based on a previous speed set for the path by a node further downstream from the downstream node and a maximum bandwidth available for the path. , the system according to claim 9. 16. further comprising a transmission means for transmitting the maximum average speed from the downstream node to the upstream node;
A system according to claim 1. 17. Claim 17, wherein the buffer means comprises a plurality of buffers allocated to each of the paths, and the comparison means determines the ideal speed based on the usage status of the buffers allocated to a given path. The system according to item 9. 18. Claim 17, wherein said ideal speed decreases as said allocated buffer exceeds a maximum value.
System described in section. 19. The method of claim 18, wherein the maximum value includes the allocated buffer containing two seconds worth of data bytes at the ideal speed before the ideal speed is reduced. system. 20. A method for controlling the flow of data in a communication network in which a plurality of data streams are transmitted from an upstream node to a downstream node via a plurality of paths, the downstream nodes A method characterized in that: specifying a maximum average speed of said data stream over a period of time.