JP2000148639A - Data transmission control method and recording medium with data transmission control program recorded therein - Google Patents

Abstract

PROBLEM TO BE SOLVED: To maintain a prescribed transfer rate, and to effectively use limited computer resources by properly controlling the timing of the reading of transmission data from a disk and the transmission of data to a network in a system for transmitting video data or the like to plural terminals. SOLUTION: A time T1 required for the reading of data from an auxiliary storage device is measured every time, when data are read from the auxiliary storage device, and a time T4 required for pausing a process is calculated, based on the measured time T1 or based on the time T1 and the time T3 required for the transmission of data which is measured at data transmission. During the calculated process pausing time T4, the process is turned into a paused state, and when the process returns to active state after the lapse of the pausing time T4, the transmission of data is conducted, and the process is turned into a paused state again.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data transmission control method and a recording medium on which a data transmission control program is recorded, and more particularly, to reading data from a medium of an auxiliary storage device such as a magnetic disk at a fixed speed, and reading the data. A technology for transmitting data over a network, which is used for a video server or the like that transmits a plurality of videos at a high speed at the same time.

[0002]

2. Description of the Related Art The processing performed by a conventional video server is as follows.
The process can be divided into two processes: a process of reading data from a magnetic disk in which a video is stored, and a process of transmitting the data to a network. A typical method of reading data from a disk is a time slot method.

FIG. 5 is an explanatory diagram of a conventional time slot system. In this method, a time slot is created by equally dividing a fixed time as shown in FIG. 5, a time slot is assigned to each client that transmits video data, read from a disk by time sharing, and stored in a shared memory. To buffer. The communication microprocessor transmits the data on the shared memory to the network in the order in which the data was read.

[0004]

The conventional method using a time slot can be implemented by one process (task). However, it can be adjusted according to the performance of a disk to be used, a disk interface, a controlling CPU, and a memory. Therefore, tuning work was required to determine the length or number of time slots. Therefore, it was difficult to operate the same software on various types of general-purpose personal computers and workstations.

It is an object of the present invention to read out data from a disk and to read out the read data so that any personal computer or workstation can operate as a video server if it has a certain level of performance. To provide transmission technology.

[0006]

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and in a system such as a VOD (Video on Demand) operated mainly by a transmission side, a transmission data from an auxiliary storage device such as a disk. By properly controlling the timing of reading data and transmitting data to the network, while maintaining a predetermined transfer rate,
It is intended to make effective use of limited computer resources.

Therefore, the present invention measures the time required to read data from the auxiliary storage device every time data is read from the auxiliary storage device, and determines the time to suspend the process based on the measured time. calculate. The process is put into a sleep state during the calculated process pause time, and when the process returns to the active state after the lapse of the pause time, data is transmitted to the network and the process is put into the sleep state again. Here, the time required for data transmission at the time of data transmission may be further recorded, and the time for suspending the process may be calculated from the time required for data transmission and the time required for reading data from the auxiliary storage device.

[0008] Instead of performing transmission to the network, writing to a memory for storing data to be transmitted to the network is performed, and transmission of data to the network is performed by another processor reading from the memory. Can also.

Specifically, in the present invention, for example, the following processing is performed. First, a process is started for each client. In each process, video data of a predetermined length (T0 seconds) is read from the disk at a time. At that time,
The time required for reading is measured (T1 second). And
The time T2 available for data transmission is calculated by subtracting T1 from T0. Similarly, the time T3 required for transmitting the video data for T0 seconds is measured. Then, T3 is subtracted from T2, and the value is divided by the sum of T3 and T1 to obtain the number of divisions N.

A timer mounted on an operating system (OS) is set to a time T1 + T3, and the process is temporarily stopped. After the time T1 + T3, the process is started by the OS, transmits only the data obtained by dividing the video data by the division number N, sets the timer again, and suspends the process. Until the read video data runs out,
Repeat the process of transmission, timer setting, and pause. When there is no more video data read, the process is executed again from reading from the disk.

By constantly measuring the time required to read data from the disk and the time required to transmit data,
No prior tuning is required. Also, since the calculated number of divisions N + 1 is the maximum number of clients that can be transmitted by the personal computer or workstation, admission control is also possible. In addition, by suspending its own process, it becomes possible to execute data reading and transmission processing for another client, and it is possible to transmit data to a plurality of clients.

In a system in which a disk reading processor and a transmission processing processor are composed of different processors, the time T3 required for data transmission is stored in a shared memory between the disk reading processor and the transmission processing processor. Is the writing time.

In the above processing, time T3 is equal to time T1.
When the time T3 is very small as compared with the time T3, the process pause time may be calculated by setting the time T3 to a predetermined value close to 0 or 0 instead of measuring the time T3.

By using the present invention, a general-purpose personal computer or workstation can be used as a video server without any tuning.

A program for causing a computer to execute the above processing can be stored in an appropriate recording medium such as a portable medium memory, a semiconductor memory, and a hard disk which can be read by the computer.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS As an embodiment of the present invention, an example in which the present invention is applied to a video server will be described. FIG. 1 is a processing flowchart of a video server according to an embodiment of the present invention, and FIG. 2 is a diagram illustrating an example of a system of the video server.

In FIG. 2, a data transmitting device 1
A video server based on a personal computer or a workstation including a U and a memory, and has an operating system that supports multi-processes (tasks). The disk storage device 2 (hereinafter, referred to as a disk) is an auxiliary storage device that stores video data to be distributed. Client terminals 4-1 to 4-n receiving distribution of video data (hereinafter, referred to as clients)
Are connected to the data transmission device 1 via the network 3.

The data transmission device 1 operates as follows. First, when the program (the first process P1) is started, it enters a state of waiting for a connection request from a client. When there is a connection request from the client, the client is connected to the client, and it is determined whether or not the connection is possible. If the connection is possible, another process (process P2) is newly created to prepare for the connection from the next client. In addition, the number of the process in which the user has started (process serial number) is recorded (step 1).

The process connected to the client reads out T0 seconds of the video data requested by the client from the disk, and reads the time T1 required for the reading process.
Is recorded (step 2).

Next, a time T2 that can be used for transmission is obtained by subtracting T1 from T0. Then, the sum of the time T3 and T1 required for the previous transmission processing is obtained, and the value obtained by subtracting T3 from T2 is divided by the sum of T1 and T3 to obtain the number of divisions N (step 3).

Then, the number of divisions N is notified to the process P2 generated in step 1. By receiving the notification of the number of divisions N, it knows what process it is in step 1, so that when a connection request is received from a client, it is possible to connect to another new client. Can be determined. If the total number of processes including its own process exceeds a value obtained by adding 1 to the number of divisions N, a new process is not created and connection from a new client is not accepted. T3
Is not measured at the time of the first transmission, but it is possible to use a default value empirically obtained or to use a value obtained when the program was executed last time.

Next, a timer value T4 is calculated (step 4). T4 can be obtained simply by the sum of T1 and T3. With the timer set for T4, the process changes itself to a suspended state and releases CPU resources for other processes (step 5).

After a time T4, the timer causes the process to be in an execution state and starts transmitting data. The amount of data transmitted at this time is only a value obtained by dividing the read data amount by N. At the same time, the time required for transmission is measured, and the value is multiplied by N to obtain T3 to be used next. Then, the timer for the time T4 is set again, and the self-suspended state is set to release the CPU resources (step 6).

After Steps 5 and 6 are repeated N times, the process of reading data from the disk is performed again, and the processes of Steps 2 to 6 are similarly repeated until all data has been transmitted.

Next, a description will be given of scheduling when a request is made from another client. In FIG.
7 shows an example of a change over time in the states of three processes when one client is connected. The horizontal axis indicates elapsed time, the thick line indicates that the process is active (running), and the thin line indicates that the process is suspended (paused).
It is in the state.

First, at time S1, process P1
Reads data from the disk in order to transmit data to the client C1. During the time when reading from this disk,
Assuming that a connection request from another client (client C2) has arrived, the process P1 performs steps 2 to 3 in FIG.
Step 5 is executed, and a suspend state is set at time S2. In this example, it is assumed that the number of divisions N calculated by the process P1 is 2. In this case, since N + 1 = 3, a maximum of three clients can be connected.

Process P2 (already the client C1
Becomes active at time S2 because the process P1 is suspended, accepts a connection from the client C2, and creates another process P3 in preparation for the next request. . Then, reading from the disk is performed, scheduling is determined, and the system enters the suspend state at time S3.

At time S4 after the timer time T4 elapses from time S2, the process P1 is activated by the OS and becomes active, the process P1 performs transmission processing, and enters the suspend state again.

Next, assuming that a connection request from the client C3 arrives at the time S5, there is no process active at the time S5, and the process P3 immediately becomes active. The process P3 is the third process, and the process P3 indicates that the number of divisions N = 2.
2, the next process is not created. As in the process P2, reading from the disk is performed, scheduling is performed, and the own process is changed to a suspended state.

If the timer of the process P2 expires at time S6 during this process, that is, if the CPU resource cannot be allocated at the time when data is to be transmitted, the process P3 is suspended. The process P2 waits until time S7, and transmission of data to the client C2 starts from time S7. The delay due to this process conflict can be solved by providing a buffer on the terminal side. The process P1 is similarly delayed, and after time S7, the entire schedule is shifted backward. Further, after time S8,
It is as if the processes are switched in round robin.

In the example of FIG. 3, it is assumed that process switching does not occur during all processes. This assumption is easily achieved by using an exclusive control function such as a semaphore provided by the OS. It is possible to realize it.

As a transmission protocol, for example, UD
When a protocol with a short processing time such as P / IP is used, the time required for the transmission processing is extremely shorter than the time required for the processing for reading from the disk. The value can be determined.

Conventionally, the schedule switching timing has been fixedly determined by measuring and calculating the required time in advance from the data transfer speed, and in the present embodiment,
By appropriately setting the schedule, the timing value is dynamically set so that other work can be performed during the idle time of the CPU. Further, in the present embodiment, the value of “how many more client terminals can be sent data” is calculated from two values, ie, the time for reading data from the disk and the time for data transfer, and the calculation result , The number of divisions N (the number of times data read from the disk is divided and transmitted) is determined, and "N + 1" is set to the number of client terminals capable of simultaneously transmitting data.

This has the following advantages. (a) Length T0 of video data read at a time from disk
Is maintained at a relatively large value, it is possible to reduce the overhead due to reading from the disk. (b) CPU resources can be utilized without waste. (c) Even if the data read speed from the disk and the data transfer speed to the network change, the upper limit speed and the lower limit speed are generally within a certain range, so that the transfer buffer size is also fixed. It is enough to prepare (d) Further, it is not necessary for the client terminal to notify the data transmitting device of the free / full state of the buffer on the terminal side.

The reason why data transmission can be simultaneously performed to N + 1 client terminals by determining the number of divisions N as described in the present embodiment will be described.

T0: Length of video data to be read from the disk at one time (reproduction time) T1: Time required to read T0 seconds of video data from the disk T2: Available for transmission of T0 seconds of video data Time T3: Time required for transmitting video data for T0 seconds T4: Time during which the process is paused In the present embodiment, the number of divisions N is determined by the following equation.

N = (T2−T3) ÷ (T1 + T3) = {(T0−T1) −T3} (T1 + T3) (Equation 1) Data of length T0 seconds is transmitted to N + 1 client terminals. Assuming that the total processing time for transmission is T _total , T _total is T _total = (N + 1) × (T1 + T3). By substituting the above [Equation 1] into this, T _total = [{(T0−T1) −T3} (T1 + T3) +1] × (T1 + T3) = {(T0−T1) −T3} + (T1 + T3) = T0 That is, the total processing time T _total is equal to the length T 0 of the video data read at one time from the disk. In other words,
It is possible to transmit video data for T0 seconds to each of the N + 1 client terminals within the time T0. If data is sent to more client terminals than the division number N + 1, the reproduction of video on the client terminal side will be interrupted, causing a problem. Further, if the number of client terminals capable of simultaneous transmission is set to a value smaller than the number of divisions N + 1, CPU resources will be wasted. According to the present embodiment, such a problem does not occur, and effective use of CPU resources becomes possible. In the calculation of the number of divisions N, N = (T2-T
3) When a fraction occurs in the value of ÷ (T1 + T3), the fraction should be truncated.

The derivation of the time T4 for pausing the process will be described. T4 is a value obtained by subtracting the time (T1 + T3) required to execute the processing for reading and sending the video data from the length T0 of the video data read at one time from the disk, and dividing the result by the division number N of T0. is there. Therefore, T4 = {T0- (T1 + T3)} N = {T0- (T1 + T3)} (T2-T3)} (T1 + T3)} = (T0-T1-T3) * (T1 + T3)} (T2-T3) ) = (T0−T1−T3) × (T1 + T3) ÷ (T0−T1−T3) = T1 + T3 From this, the time T4 during which the process is suspended is T1 + T3.
It can be understood that it should be the time of.

In the above example, the case where the microprocessor for communication and the microprocessor for reading the disk are the same has been described. FIG. 4 shows an example of a data transmission apparatus having a configuration in which a disk reading processor and a transmission processing processor are separated as another configuration example of the system. 4 includes a disk reading processor 11, a separate transmission processing processor 13, and a shared memory 1 commonly accessible from these processors.
2 and a bus 14 for connecting these.

In the case of this configuration, the transmission data is read from the disk storage device 2 by the disk reading processor 11 and written to the shared memory 12. The transmission processor 13 detects that data has been written to the shared memory 12 and performs transmission processing to the requesting client via the network 3.

Therefore, when the process scheduling for data transmission according to the present invention is applied to the disk reading processor 11, the network 3
In step 6, only the process of writing data to the shared memory 12 is performed. Since the processing time required for the processing of writing to the shared memory 12 is extremely shorter than the processing time of reading from the disk storage device 2, the time T3 is set to a small fixed value, or T3 is set to 0 and only the time T1 is set. Number of divisions used N
Can be determined.

[0042]

By using the present invention, a general-purpose personal computer or workstation equipped with a disk, disk interface, or microprocessor with unknown performance can be used as a data transmission device for video data without prior tuning. It can be used.

[Brief description of the drawings]

FIG. 1 is a processing flowchart according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a system to which the present invention is applied.

FIG. 3 is a diagram illustrating an example of a time change of a process state when three clients are connected according to an embodiment of the present invention.

FIG. 4 is a diagram showing an example of a second system to which the present invention is applied.

FIG. 5 is an explanatory diagram of a conventional time slot system.

[Explanation of symbols]

 DESCRIPTION OF REFERENCE NUMERALS 1 data transmission device 2 disk storage device 3 network 4-1..., 4-n client terminal 11 disk reading processor 12 shared memory 13 transmission processing processor 14 bus

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference)

Claims (6)

[Claims] 1. A method for reading data in an auxiliary storage device using a computer system equipped with an auxiliary storage device, a network interface, and an operating system that supports multi-process, by a process corresponding to each destination, and transmitting the data to a network. Measuring a time required to read data from the auxiliary storage device each time data is read from the auxiliary storage device, and a time period for suspending a process based on the measured time. Transmitting the data to the network when the process returns to the active state after the elapse of the pause time.
Putting the process back to a sleep state again. 2. The data transmission control method according to claim 1, further comprising the step of recording the time required for data transmission at the time of data transmission; A data transmission control method, wherein a time for suspending a process is obtained from a time required for reading data from the auxiliary storage device. 3. The data transmission control method according to claim 1, wherein writing to a memory for storing data to be transmitted to the network is performed instead of performing transmission to the network. Data transmission control method. 4. A program for reading data in the auxiliary storage device by a process corresponding to each transmission destination in a computer system equipped with an auxiliary storage device, a network interface, and an operating system supporting multi-process, and transmitting the data to a network. A process of measuring the time required to read data from the auxiliary storage device each time data is read from the auxiliary storage device, and suspending the process based on the measured time. A process for determining the time to perform the process, a process for suspending the process during the determined process pause time, and transmitting the data to the network when the process returns to an active state after the lapse of the pause time. The process of suspending a process
A recording medium on which a data transmission control program is recorded, wherein a program to be executed by a computer is recorded. 5. A recording medium on which the data transmission control program according to claim 4 is recorded, wherein the program further comprises a process for recording a time required for data transmission at the time of data transmission, and a process for obtaining a time for pausing the process. ,
Recording a data transmission control program for causing a computer to execute a process of obtaining a time period for suspending a process from a time period required for the data transmission and a time period required for reading data from the auxiliary storage device. Medium. 6. A recording medium storing a data transmission control program according to claim 4 or 5, wherein said program stores data to be transmitted to a network instead of to said network. A recording medium storing a data transmission control program for causing a computer to execute a process of writing data to a computer.