JP3416498B2 - Server device, control method therefor, and recording medium storing server device control program - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a real-time server device for inputting and outputting digitized real-time data such as video and audio with a plurality of terminal devices and a control method thereof.

[0002]

2. Description of the Related Art In recent years, digital audiovisual equipment has become widespread, and the performance and cost of digital data recording devices have been increasing. Along with this, development and practical application of a real-time server device that stores digital data such as video and audio in digital form and transfers digital data such as video and audio in real time along a time axis in response to requests from multiple terminal devices. Is progressing.

As methods for transferring digital data, there are real-time transfer represented by transfer of video and audio data as described above and non-real-time transfer represented by transfer of static data such as a still image. These features will be described below.

In the transfer of real-time data, in order to faithfully reproduce video and audio along the time axis, it is necessary to transfer at a constant rate without delay or interruption. In order to realize this, the maximum bandwidth according to the processing capacity is usually set in the real-time server device, and bandwidth allocation control is performed when the connection is started between the server device and the terminal device. That is, the bandwidth that can be used by each terminal device at the time of connection is set so that the total load on the server device does not exceed the maximum bandwidth set in the server device. Further, the server device schedules and switches the transfer to each terminal device so as not to exceed the band set in the transfer with each terminal device, and allocates resources such as the processing capacity and memory of the server device. To do. As described above, the server device performs data transfer to the terminal device within the set maximum bandwidth, thereby performing processing with a margin for the processing capacity of the server device and causing delay or interruption due to overload. Data transfer with the terminal device.

On the other hand, static still images, data files, etc.
Strict bandwidth allocation is not required for non-real-time transfer in which delays and interruptions are allowed, and an algorithm is adopted that transfers as fast as possible with each terminal device. For example, a simple flow control algorithm corresponds to this. In the flow control algorithm, the transmission side performs data transmission with the maximum transfer capability within the range where the buffer provided on the data reception side does not overflow. As a result, the transfer capacities of the server device, the terminal device, and the communication medium connecting them can be matched, and the transfer capacity can be used without waste. However, on the other hand, when data is transferred to a plurality of terminal devices, delay or interruption may occur in another terminal device due to the influence of the load from one terminal device.

As described above, the real-time transfer and the non-real-time transfer are generally controlled by different algorithms. However, depending on the application, it may be desired to perform the real-time transfer and the non-real-time transfer at the same time by using the same server device. For example, real-time video transfer is performed with a certain terminal device, and real-time video display is not performed with another terminal device, but a digital video that has been created and edited by another device in advance is registered in the real-time server device. Such is the case. In the latter case, it is not necessary to perform real-time transfer, but rather non-real-time transfer that allows transfer and registration as fast as possible is desirable. In such a case, the server device ideally performs the real-time transfer to the terminal device requesting the real-time transfer, while performing the non-real-time transfer as fast as possible to the terminal device requesting the non-real-time transfer.

[0007] In order to meet the demand for such a mixture of real-time transfer and non-real-time transfer, and also to harmonize with the algorithm of non-real-time transfer so that there is no data delay or interruption in real-time transfer, there are some conventional means. was there.

As a simple solution, the server device is composed of a computer, each input / output is controlled by an individual process, and real-time processing is performed without arbitrating the bandwidth between the processes. It is conceivable that the parallel processing priority of the process may be simply set higher than that of the non-real-time processing process. There are several operating systems that can provide high priority to processes that perform real-time processing, and it was expected that resources for input / output will be preferentially allocated to execute real-time processing. However, according to this method, there is not enough certainty to guarantee that data will not be delayed or interrupted. This is because it is not always possible to unconditionally interrupt a process with a low priority even if a process with a high priority is prioritized. For example, when the data storage medium is a hard disk drive, it is not realistic to interrupt another transfer request during the transfer of one block, because a head seek or the like occurs and the transfer efficiency is significantly reduced. Therefore, the block transfer process of the hard disk drive is processed first, and a considerable waiting time is required to switch from the process of low priority to the process of high priority. If the processes by the simple priority management are sequentially executed under such a constraint, delays or interruptions may occur during the execution of the real-time process.

As a first example of solving the above problems, there is a technique disclosed in Japanese Patent Laid-Open No. 8-228200. That is, this is a method of scheduling transfer to each terminal device at a fixed cycle that is unified in the entire server device system, performing real-time transfer first, and then utilizing the remaining time of the cycle as transfer of non-real-time data.

As a second example, Japanese Patent Laid-Open No. 8-163
There is a technique disclosed in Japanese Patent Publication No. 072. That is, transfer is managed by time-divisional time slots, and first time slots are allocated for real-time transfer.
This is a method of utilizing an empty time slot that is not assigned to real-time transfer as a time slot for non-real-time data transfer.

In these methods, since the non-real-time processing is executed after reserving the processing amount required for the real-time processing in advance, the non-real-time processing is performed while performing the transfer without delay or interruption to the terminal device that requests the real-time transfer. High-speed transfer is performed to the terminal device that requests the transfer.

[0012]

On the other hand, however, the above-mentioned first and second conventional examples have a common drawback. According to these methods, data transfer scheduling is performed by one reference clock that is unified in the system. For this reason, when the terminal devices operate asynchronously with each other, it is difficult to perform scheduling in each terminal device because it is necessary to perform scheduling in accordance with the reference clock, and the degree of freedom of input / output timing decreases, and more than one It has been difficult to set a proper reference clock for the local clock of. Especially when implementing data input / output processing for various existing terminal devices,
Since the clock speed and the standard itself are significantly different,
Procedures such as clock synchronization and stuffing for adjusting speed are complicated, and it is difficult to perform operation arbitration by synchronizing. In addition, in the first example, scheduling is performed in accordance with a unique cycle which is a relatively long cycle, so that when a transfer request is generated on the terminal device side, a delay occurs due to synchronization with the unique cycle, such as an asynchronous operation. There are many disadvantages when trying to build a system with various terminal devices.

Further, in the above-mentioned first and second conventional examples, when this is realized by a computer program, the operations of both real-time transfer and non-real-time transfer processes are necessarily arbitrated by a single scheduler. Therefore, there is also a drawback that it is difficult to make the process into a divided structure. That is, it is difficult to separate the real-time transfer and the non-real-time transfer from each other, the program size increases, the structure becomes complicated, and the protocols for the real-time transfer and the non-real-time transfer cannot be easily changed.

In view of the problems of the prior art as described above, the present invention is capable of efficient processing even for data transfer to terminal devices operating asynchronously with each other, real-time transfer and non-real-time transfer. The arbitration between processes is simple and the program code is easy to be separated, and the real-time transfer is reliably performed to the terminal device requesting the real-time transfer without delay or interruption. An object of the present invention is to provide a server device capable of performing real-time transfer and a control method thereof.

[0015]

To solve the above problems, a real-time server device according to the present invention corresponds to a synchronous clock source and a real-time input / output control means corresponding to each real-time terminal device, and a non-real-time terminal device. A clock source, a non-real-time input / output control unit, and a storage medium are provided, and the real-time input / output control unit starts the use of the corresponding synchronous clock source when inputting / outputting data, and connects between the storage medium and the real-time terminal device. Data input / output is controlled, and when the data input / output is completed, the use of the synchronous clock source is stopped, and the non-real-time input / output control means, when data is input / output, the data transfer capability of the server device and the corresponding synchronization Real-time I / O control means that uses a clock source The data source input / output to / from the storage medium and the non-real-time data input / output to / from the non-real-time terminal device are performed by using the clock source at the maximum usable data rate calculated from the difference between the total tarates. And

With this configuration, a plurality of real-time input / output control means transfers data to the terminal device asynchronously with each other to eliminate delay or interruption of data to the terminal device for inputting / outputting data in real time, and non-real-time input / output control. By monitoring the clock usage status of each real-time input / output control means by means of the means and detecting the bandwidth currently used by each real-time input / output control means from the operating state, while limiting the bandwidth so that the load of the entire system is not excessive Non-real time input / output can be performed. Also, a single clock source is used as a rate-referenced clock to limit the bandwidth, and this clock source does not need to be associated with the overall system scheduling, so it can be used when arbitrating multiple terminals. Synchronous processing, stuffing processing, etc., which were required in the past, are not required, and a clock with a stable cycle is easily obtained, so that reliable rate control can be performed without complicated processing.

In order to solve the above-mentioned problems, a real-time server device according to the present invention includes a synchronous clock source and a real-time input / output control means corresponding to each real-time terminal device, and a clock source and a non-real-time terminal device corresponding to a non-real-time terminal device. Real-time input / output control means,
A storage medium and a band information management unit that manages band information used for data input / output of the server device, and the real-time input / output control unit registers band information in the band information management unit at the time of data input / output. Then, the use of the corresponding synchronous clock source is started to control the data input / output between the storage medium and the real-time terminal device, and when the data input / output ends, the band information registered in the band information management unit is stored. Deleting, stopping the use of the synchronous clock source, and the non-real-time input / output control means, when data is input / output, the data transfer capacity of the entire server device and the data rate based on the band information registered in the band information management unit. The maximum available bandwidth calculated from the difference between the total and the data input / output to / from the storage medium by using the clock source. And performing the non-real-time data input and output between the non-real-time terminal device.

The bandwidth information is speed information when data transfer is carried out, and typically means the bit transfer amount per unit time.

With this configuration, the non-real time input / output control means can easily and surely acquire the band information. Therefore, it is easy to functionally separate the control unit that performs real-time processing and the control unit that performs non-real-time processing.

Next, the server device uses, as a clock source corresponding to the non-real-time terminal device, any one of the synchronous clock sources used by the real-time input / output control means for inputting / outputting data. Is preferably used.

With this configuration, since the real-time input / output control means limits the clock according to any one clock source, the non-real-time input / output control means does not need to prepare an independent clock and needs to limit the band. In some cases there is definitely the effect that there is more than one clock available.

Next, in the server device, depending on whether the non-real time input / output control means has started or stopped using the corresponding synchronous clock source,
It is preferable that the real-time input / output control means has a function of determining whether or not real-time data input / output is being performed.

With such a configuration, since the clock is always used when the real-time input / output control means actually inputs / outputs data, the operation status can be judged easily and surely. Use of,
By adopting the method utilizing the non-use, it is reliable and there are few restrictions between the control means, and it is easy to separate the control means for performing real-time processing and the control means for performing non-real-time processing.

Next, in the server device, use start and stop of the synchronous clock source corresponding to each real-time input / output control means are performed by registering and deleting each clock interrupt, and the non-real-time input is performed. When selecting the synchronous clock source used by the output control means, it is preferable to determine whether or not the synchronous clock source is used by the real-time input / output control means by the registration status of the interrupt.

With this configuration, when selecting the clock source used by the non-real time input / output control means, it becomes easier and more reliable to determine whether the clock source is used by the real time input / output control means.

Next, in the method of controlling the server device, all or part of the functions of the real-time input / output control means and the non-real-time input / output control means of the server device are
It is characterized by being realized by CPU processes operating in parallel controlled by program steps.

Further, the recording medium recording the server device control program is characterized in that the program steps of the control method of the server device are recorded.

With this configuration, CPs operating in parallel are controlled by respective control means by respective program steps.
By using the U process, it is possible to provide a method of controlling a server device by a computer controlled by a CPU.

[0029]

BEST MODE FOR CARRYING OUT THE INVENTION Preferred embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a schematic diagram showing the operation of the real-time server device according to the embodiment of the present invention. The configuration and operation of the embodiment will be described below with reference to FIG.

The server device 101 performs input / output with the n-1 VTRs 111, which are real-time terminal devices, through the communication path, and at the same time, the tape data storage device 112, which is a non-real-time terminal device, and other server devices 113. , And also performs data input / output with the control computer 114. The server device 101 includes a recording medium 102, and records audiovisual data on the recording medium 102. Recording medium 102
Is a digital recording medium such as a hard disk drive that can be accessed from a plurality of streams via the bus 103. The maximum data input / output band is set in advance in the server device 101. The maximum bandwidth has a bandwidth that is equal to or greater than the sum of the maximum bandwidths allocated to each real-time terminal device, and is set so that even if each real-time terminal device inputs / outputs data up to the maximum bandwidth, no failure due to delay or interruption occurs. It is determined by the data input / output processing capacity of the recording medium 102 and the transfer processing capacity of the bus 103. The audiovisual data on the recording medium 102 is input to and output from the real-time terminal device 111 on the same time axis as the time axis specific to the video, and is treated as a real-time data stream 116. Further, the same audiovisual data is input / output to / from the non-real-time terminal devices 112 to 114 on an arbitrary time axis irrespective of the video-specific time axis, and is treated as a non-real-time data stream 117.

First, the data input / output operation in real time will be described. Since the VTR 111 records and reproduces video in real time, real-time data input / output is essential. Therefore, the control means 106 controls the input / output timing using the synchronous clock source 110. The control means 106 uses the buffer 104 to absorb and adjust the difference between the transfer processing speed and input / output timing on the internal bus 103 and the input / output processing speed and input / output timing of the real-time data stream for the VTR 111, thereby adjusting the video data. The data transfer process is performed with the real-time VTR device 111 without interruption.

The server device 101 detects the clock rate peculiar to each VTR of the real-time terminal device connected in advance, and sets the detected clock rate to the synchronous clock source 1 to clock source n-1.
Here, each of the control means 1 to n-1 uses the set synchronous clock sources 1 to n-1, and the control means n uses one clock arbitrarily selected from the synchronous clock sources 1 to n-1. As a result, the server device controls input / output by n control means that operate asynchronously with each other.

Of the n-1 VTRs 111, the VT
It is assumed that R1 records the digital data output from the server device on the tape. The synchronous clock source 1 in the server device 101 generates a clock that serves as a reference of the real-time data stream 116 for the VTR 1, and the control means 1 receives the periodic interrupt generated by the clock source 1 and the real-time property of the data stream 116. And as a time base to keep the transfer rate,
Real-time data stream 116 for VTR1
Generate and transfer. In the VTR 1, the built-in VTR specific clock 1 is synchronized with the clock source 1, and the VTR 1 receives the real-time data stream 116 and records it on the tape.

Next, it is assumed that the VTRn-1 is reproducing the tape. Most VTR devices except products used for broadcasting stations and business editing applications are not equipped with a function to perform playback in synchronization with an externally supplied clock from the viewpoint of cost, and playback is performed using a clock built into the VTR. It will be done on a standard basis. Therefore, the synchronous clock source n-1 in the server device 101 which is set in advance to be synchronized with the VTR built-in clock source n-1 of the VTR n-1 is used. The control means n-1 is a synchronous clock source n-1.
Is used as a time reference for maintaining the real-time property and bit rate of the data stream, and the real-time data stream 116 for VTRn-1 is generated and transferred.

As shown in the present embodiment, since the present invention does not require special synchronization for unifying the entire system, it is possible to use an existing device that does not have a function of synchronizing with an externally supplied clock as described above. There is an advantage that it can be used as a terminal device without special modification.

Next, the real-time input / output control means 1-n-
The operation 1 will be described with reference to the flowchart of FIG. Of course, according to the program based on this flowchart, this control means can be realized as a process on the CPU.

Now, VTRm which is a real-time terminal
(VTRm is a VTR of VTR1 to VTRn-1)
Performs data transfer with the server device 101.
The control means m (the control means m is a control means of the control means 1 to n-1) is assigned for the transfer processing. Before starting the data input / output, the control means m
It is determined whether there is a usable band. That is,
The total bandwidth already registered by the other control means 1 to n-1 and the total bandwidth to be registered by the control means m are calculated, and the result and the maximum bandwidth preset for the server device 101 are calculated. If it does not exceed, it is determined to be usable. This registration management is performed by providing a memory area for registration management in a memory connected to the bus 103, for example. If the result of determination is that there is no usable band, data input / output is not performed. In this case, it is preferable to report to the control computer 114 and VTRm of the server device that the transfer process is busy. Next, if there is a usable band, the band information of the data input / output used by the control means m is registered, and then the clock source m specific to the control means m (the clock source m is the clock source of the VTRm). Pre-configured to sync) to start using interrupts.

When the use of the interrupt is started, the recording medium 102 is limited until the data input / output is completed by using the interrupt to limit the input / output band to the registered band or less.
And data transfer between VTRm and VTRm. When the data input / output is completed, the use of the unique clock source m interrupt is stopped, and the data input / output bandwidth information registered by the control means m is deleted.

Next, the data input / output operation in non-real time will be described. In FIG. 1, a tape data storage device 112, another server device 113, a control computer 1
14 are devices for inputting / outputting data in non-real time, and these are for inputting / removing images regardless of real time for the purpose of fetching stored images, registration of encoded images, high-speed transfer for editing or viewing. Output. For these terminal devices, the control means n (107) controls the input / output timing.

The control means n according to this embodiment is composed of sub-control means n0 and n1 to n3. Here, the connected non-real time devices are 112 to 11
The corresponding sub-control means will be described as n1 to n3, but the number of connecting devices and the number of sub-control means are examples and are not limited to three.

An important point in the configuration of the control means n is that it has two types of sub-control means n0 and n1 to n3, and the sub-control means n0 coordinates with other real-time terminal equipment processes. In order to limit the bandwidth, data input / output with the storage medium 102 is performed, and the sub control means n
1 to n3 enable high-speed data transfer processing with the non-real-time terminal device as much as possible. This makes it possible to reliably perform real-time data transfer with the real-time terminal device and realize high-speed data transfer with the non-real-time terminal device while considering the data input / output processing efficiency of the entire system.

The sub-control means n0 controls the data transfer between the recording medium 102 and the buffer 105, and controls means n1-n1.
n3 is the buffer 105 and each terminal device 112-1.
Control transfer between 14 The buffer 105 has a FIFO structure and can independently perform input and output. Therefore, the control means n0 and the control means n1 to n3 can perform the transfer independently of each other. According to the configuration of the control means n as described above, since the sub-control means n1 to n3 can operate independently, it is possible to use the conventional control means that does not perform band management as it is for n1 to n3, and n1 to n3. It is possible to flexibly operate such that each of them may use different algorithms and protocols, and it is a particularly suitable configuration for non-real time high-speed transfer. Each of the control means n1 to n3 performs input / output between the buffer 105 and each non-real-time terminal device according to any general transfer algorithm such as conventional flow control. Since these operations are well-known data transfer means, description thereof is omitted here.

Next, the control means n0 for executing the data input / output to / from the storage medium 102 by limiting the band in the non-real time input / output control means n will be described.

The operation of the control means n0 will be described with reference to the flowchart of FIG. The control means n0 starts real-time input / output control means 1-n- before starting the transfer.
It is determined whether 1 is using interrupt. If the result of determination is that interrupts have not been used, data transfer is performed between the recording medium 102 and the buffer 105 at the maximum allowable speed. This makes it possible to make full use of the full capabilities of the server device. As a result of the interrupt use determination, when the interrupt is used by the other control means 1 to n-1, the control means n0 first calculates the usable band. The bandwidth that can be used by the control means n0 is calculated by subtracting the total bandwidth that has already been registered by the control means 1 to n-1 from the preset maximum bandwidth of the server device 101 in advance.

Next, of the synchronous clock sources 1 to n-1, an arbitrary one is selected from among those in which an interrupt is used, and the interrupt is shared so that the control means n0 can also use the interrupt. Select multiple clocks here, or not a clock synthesized from multiple clock sources,
The reason for selecting any one of the already used clocks is that there is a great advantage in sharing processing with the clocks of other processes, rather than using a completely independent clock, because the processing complexity can be reduced. Because there is. It is not always necessary to share the clock source with another process, and the control means n may use an individual clock source n. In this case, in particular, even if the first to (n-1) th clock sources can be changed due to the convenience of the terminal device, the clock for non-real time transfer is not affected, so that a clock with a stable cycle can be obtained. it can.

When the sharing of the interrupt is completed, the control means n uses the interrupt to calculate the input / output band (the total of the three bands corresponding to the respective non-real-time terminal devices 112 to 114), which is available in the previous step. Data transfer is performed between the recording medium 102 and the buffer 105 while limiting the bandwidth to a certain band or less. Here, the control means n0 performs data input / output with the recording medium 102 by adjusting so as not to collide with the process of the control means using the shared interrupt. It should be noted that the control unit n0 returns to the first step and performs the update process each time the interrupt status changes, such as an increase or decrease in the number of terminal devices that perform real-time data input / output.

The connection stream 116 between the server device and the terminal device shown as a stream in FIG. 1 shows a data flow having logically consistent continuity, and a physical communication line is used. It is not a representation.
That is, even when the server device and the terminal device are individually connected by communication lines, the server device and the terminal device are physically connected by a network that shares one communication line, and a plurality of real-time devices are connected on the network. Even if the server device and the terminal device are connected by a non-real time stream, the effect of the present invention is the same.

As described above, each control means of the real time system can notify other control means of the current operating state through the use of the interrupt. Since the use of interrupts is a basic operation for generating a time base based on hardware, there are few errors as an indicator of the operating status, and bandwidth information can be updated at the same time by registering bandwidth information at the same time. Is. By using the interrupt, the non-real-time control means can appropriately grasp the current operating conditions of the other control means.

By using such a notification method using interrupts, it is possible to construct a server device control means which operates asynchronously but surely in cooperation with each other. According to the configuration of the present invention, even when the control means that operate in parallel with each other are distributedly processed by a plurality of CPUs, communication between the CPUs is not necessary for adjusting the bandwidth, so that independence can be achieved. Higher operation is possible. In addition, the real-time control means 106
Even if the number of operations of is not determined, the control unit 107 can perform control so that the band of the entire system does not exceed the maximum band by detecting an interrupt.
Very flexible for configuration changes. Further, since the control means n does not use an individual interrupt, there is an effect that the interrupt resource in the server device can be saved.

[0051]

The real-time server device according to the method of the present invention is suitable for data transfer to terminal devices that are asynchronous with each other, and arbitration between program codes for real-time data transfer and non-real-time data transfer is simple and programs are separated. Easy, the program configuration can be easily changed, real-time transfer can be reliably performed without delay or interruption to the terminal device requesting real-time transfer, and it is efficient for the terminal device requesting non-real-time transfer. High-speed transfer in non-real time is possible.

[Brief description of drawings]

FIG. 1 is a schematic diagram of a server device according to an embodiment of the present invention.

FIG. 2 is a flowchart of processing by a real-time input / output control unit of the server device according to the embodiment of the present invention.

FIG. 3 is a flowchart of processing by a non-real-time input / output control unit of the server device according to the embodiment of the present invention.

[Explanation of symbols]

101 server device 102 recording medium 103 bus 104, 105 buffer 106 control means 1 to n-1 107 control means n 108 sub control means n0 109 sub control means n1 to n3 110 synchronous clock source 111 VTR 112 Tape data storage device 113 Other server devices 114 control computer 115 interrupt 116 Real-time data stream 117 Non-real time data stream

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H04L 29/08 G06F 13/00 353

Claims (7)

(57) [Claims] 1. Inputting and outputting data in real time 1
A server device capable of inputting / outputting data to / from a plurality of real-time terminal devices and performing non-real-time data input / output to / from a plurality of real-time terminal devices, the server device corresponding to each of the real-time terminal devices. A synchronous clock source, real-time input / output control means, a clock source corresponding to the non-real-time terminal device, non-real-time input / output control means, and a storage medium are provided, and the real-time input / output control means handles the data input / output. Control of the data input / output between the storage medium and the real-time terminal device is started, and the use of the synchronous clock source is stopped when the data input / output ends, and the non-real-time input / output control is performed. The means transfers the data of the server device when inputting / outputting the data. Capacity and the maximum available data rate calculated from the difference between the sum of the data rates of the real-time input / output control means using the corresponding synchronous clock source, and the data input to the storage medium using the clock source. A server device for performing output and non-real time data input / output with the non-real time terminal device. 2. A device for inputting and outputting data in real time 1
A server device capable of inputting / outputting data to / from a plurality of real-time terminal devices and performing non-real-time data input / output to / from a plurality of real-time terminal devices, the server device corresponding to each of the real-time terminal devices. Synchronous clock source and real-time input / output control means, clock source and non-real-time input / output control means corresponding to the non-real-time terminal device, storage medium, and bandwidth for managing bandwidth information used for data input / output of the server device. An information management unit, the real-time input / output control unit registers band information in the band information management unit at the time of data input / output, starts using the corresponding synchronous clock source, and stores the storage medium and the real-time data. When data input / output between terminal devices is controlled and data input / output is completed, Remove the band information registered in the frequency information management unit, to stop using the synchronous clock source,
The non-real-time input / output control means uses the maximum available bandwidth calculated from the difference between the data transfer capability of the entire server device and the total data rate based on the bandwidth information registered in the bandwidth information management unit during data input / output. According to the server device, the clock source is used to perform data input / output with the storage medium and non-real time data input / output with the non-real time terminal device. 3. The clock source corresponding to the non-real-time terminal device is any one of the synchronous clock sources used by the real-time input / output control means for inputting / outputting data. Alternatively, the server device according to item 2. 4. The non-real time input / output control means,
3. The method according to claim 1, further comprising a function of determining whether the real-time input / output control unit is executing real-time data input / output depending on whether the use of the corresponding synchronous clock source is started or stopped. The server device described. 5. The non-real-time input / output control means uses the synchronous clock source corresponding to each real-time input / output control means, which is started and stopped by registering and deleting each clock interrupt. 4. The server device according to claim 3, wherein, in selecting a synchronous clock source, whether or not the synchronous clock source is used by the real-time input / output control unit is determined based on a registration status of the interrupt. 6. The whole or a part of the functions of the real-time input / output control means and the non-real-time input / output control means of the server device according to claim 1.
A method for controlling a server device, which is realized by CPU processes operating in parallel controlled by program steps. 7. A storage medium on which the program steps according to claim 6 are recorded.