JP3024686B2 - Storage subsystem - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the field of information processing and, more particularly, to a storage subsystem for storing long and continuous data by combining a plurality of storage devices.

[0002]

2. Description of the Related Art Conventionally, as a means for increasing the capacity of an external storage device and increasing the data transfer speed, a method of connecting a plurality of storage devices to a main unit and performing data transfer in parallel has been known. Are known.

As one example, a storage subsystem as shown in FIG. 2 has been proposed for continuously storing and reproducing continuous data. In FIG. 2, reference numeral 1 denotes a storage device unit in which a plurality (four in FIG. 2) of storage devices # 1, # 2, # 3, and # 4 are connected in parallel to a higher-level device (not shown).

In this storage subsystem, a storage unit 1
A synchronization control unit 2 for synchronizing the read and write start times between the respective storage devices # 1 to # 4, and a storage device for selecting the storage devices # 1 to # 4 for reading / writing and switching the connection state to the higher-level device A switching unit 3 is added to calculate the holding position of the next data block to be accessed next, and to position a read / write mechanism (not shown) at that position.

With such a configuration, as shown in FIG. 3, the seek operation time and the rotation waiting time accompanying the seek operation are hidden, and the seek operation is conventionally performed during the write / read operation. Necessary long data can be continuously written / read.

In this system, unlike a case where long data is stored in one storage device, a plurality of storage devices # 1 to # 1 are stored.
Data blocks are distributed and held in # 4, and data is continuously reproduced while the storage devices # 1 to # 4 are sequentially switched by the storage device switching unit 3. Storage devices # 1 to # 4
The switching procedure is constant regardless of the type of data.
The storage devices # 1 to # 4 are sequentially selected in a cycle that goes through all the storage devices # 1 to # 4, and the data of each of the storage devices # 1 to # 4 is selected.
Block is accessed.

Therefore, if the same storage device is not accessed at the same timing, the access is not duplicated in other storage devices, and a characteristic is that a plurality of continuous data read requests can be accepted.

For example, in FIG. 3, from the point A where the reading of the first data block from the storage device # 1 is completed and the access for reproducing the fifth data block is completed, the fifth data block is obtained. In the period up to the point B at which the reading of the data block is started, the first storage device # 1 is, for example, the first data block or the first storage device (not shown in FIG. 3). An access request to read the (4N-3) th (N is a natural number) data block held in # 1 can be permitted. The same applies to the remaining storage devices No. 2 to No. 4.

Therefore, the storage subsystem constructed as described above has a plurality of users without interrupting a long data.
There is a feature that can be provided to the.

[0010]

However, in the above-mentioned conventional storage subsystem, if an access request to the same storage device is generated at the same timing due to a plurality of continuous data read requests, this is not considered. There was a problem that it was not acceptable. That is, the multiple access capability of the storage subsystem can be effectively used only when the access is not duplicated in the same storage device.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and an object of the present invention is to permit a case where accesses to the same storage device are duplicated at the same time by a plurality of continuous data read requests. It is an object of the present invention to provide a storage subsystem which does not cause a wait state for a plurality of continuous data read requests.

[0012]

In order to achieve the above object, according to the present invention, a plurality of storage devices capable of independently performing read / write operations are combined, and a plurality of storage devices are accessed while switching the storage device to be accessed. In a storage subsystem that reads and writes continuous data consisting of a plurality of data blocks, a data block accessed next in one storage device by a continuous data read request is stored in the storage subsystem. There is provided a reading unit for reading in advance using the idle time of access to the apparatus, and an auxiliary storage device for holding the previously read data and reading out the held data by the next access.

According to the second aspect of the present invention, in the read unit, when allocating the execution timing of access to each of the storage devices within a repetitive read cycle that cycles through all the storage devices, one of the storage devices is accessed next. If there is a request to read another continuous data at the same time as the timing at which the data block is read in advance using the idle time of access to the one storage device and held in the auxiliary storage device, the other continuous data is The access related to the read request is assigned in preference to the access for holding the data block in the auxiliary storage device.

[0014]

According to the first aspect, when a request for reading continuous data stored as a plurality of data blocks across a plurality of storage devices is issued, for example, the current access to a certain storage device is performed. The data block accessed next to the data block is read by the reading unit using the idle time of the access in which the access by the host device is switched from this storage device to another storage device,
The read data is stored and held in the auxiliary storage device 5. The held data is read by the next access.

Therefore, even if accesses to the same storage device are duplicated at the same time due to a plurality of continuous data read requests, a wait state for the read requests does not occur. According to the second aspect, when allocating the execution timing of the access to each storage device within the repetitive read cycle that cycles through all the storage devices, the data block to be accessed next in one storage device is assigned to the one storage device. When there is a request to read another continuous data at the same time as the timing of reading the data in advance using the idle time of the storage device and holding it in the auxiliary storage device, the access related to the read request of the other continuous data is made. Are allocated in preference to the access for holding the data block in the auxiliary storage device.

[0016]

FIG. 1 is a block diagram showing one embodiment of a storage subsystem according to the present invention. The same components as those in FIG. 2 showing a conventional example are denoted by the same reference numerals. That is, 1 is a storage device unit, # 1 to # 4 are storage devices, 2 is a synchronization control unit, 3 is a storage device switching unit, 4 is a reading unit, and 5 is an auxiliary storage device. In the present embodiment, long data is continuously recorded / reproduced by four storage devices # 1 to # 4.

The read unit 4 receives a read request from a higher-level device (not shown) via the storage device switching unit 3, and stores a data block accessed next to the data block in the currently accessed storage device. The data is read out using the idle time of the access in which the access by the host device is switched from this storage device to another storage device, and the read data is stored in the auxiliary storage device 5.

Next, the operation principle of the above configuration will be described with reference to FIG.

Here, for ease of explanation, the storage device #
Paying attention to No. 1, the vacant time immediately after reproducing the first data block shown in FIG.
The third data block is read in advance, and the read data is stored in the auxiliary storage device 5.

The fifth data block held in the auxiliary storage device 5 is reproduced from the auxiliary storage device 5 at the timing shown in FIG. As a result,
Since the storage device # 1 can process the read request of another continuous data generated at the timing of ', the reading of the continuous data generated in the same storage device at the same timing can be permitted.

Needless to say, if there is a read request for other continuous data at the same time as the timing when the fifth data block is read and held in the auxiliary storage device 5, this access is prioritized, and thereafter, In addition, the switching control is also sequentially performed on the auxiliary storage device 5 according to a predetermined procedure. This procedure is exactly the same as normal access.

Thus, the storage subsystem of FIG.
This allows a plurality of successive data read requests to be sequentially generated at any timing.

Next, the relationship among the number of readouts of multiplexed continuous data, the number of storage devices, and the required storage capacity of the auxiliary storage device of the storage subsystem according to the present invention will be described with reference to FIG.

As is apparent from FIG. 5, the multiple access number Na of the storage subsystem is determined by the cycle of the storage device,
In other words, it means how many times data processing including access can be executed within a repetitive read cycle Tc in which one storage device repeatedly reproduces data blocks. Therefore, assuming that the number of storage devices is n, the data transfer time is Tt, and the maximum access time of the storage device is Ta, the multiple access number Na of the storage subsystem of the present invention is expressed by the following equation (1). n is represented by equation (2).

Number of multiple accesses Na = (repetitive read cycle Tc / data processing time Ts) = nTt / (Tt + Ta) (1) Number of storage devices n = Na (Tt + Ta) / Tt (2) Equation (1) From (2), to increase the number of multiple accesses Na, it is necessary to configure a storage subsystem with small-capacity storage devices and increase the number n of storage devices, and to transfer data. It can be seen that if the data transfer time Tt is selected such that the maximum access time Ta of the storage device becomes relatively shorter than the time Tt, the required number of storage devices can be reduced. Finally, if Tt >> Ta,
The number of multiple accesses Na approaches the number n of storage devices.

On the other hand, the storage capacity Ms of the auxiliary storage device is obtained by the following equation (3), where Mb is the storage capacity of the data block.

The storage capacity of the auxiliary storage device Ms = n · Na · Mb (3) In the formula (3), n is n data corresponding to the number of storage devices in order to reproduce continuous data. The memory of the block is required, and Na means that it corresponds to the read request of the continuous data of Na set. Substituting equation (2) into equation (3), equation (3) can be rewritten as equation (4) below.

Ms = Na ² (Mb / Tt) (Tt + Ta) = Na ² (Mb / Tt) Ts (4) In equation (4), Mb / Tt is the data transfer rate of the storage device; Ts is a data processing time required to read one data block. Data transfer speed Mb
/ Tt and the maximum access time Ta are constants determined by the form of the storage device, and significant improvements cannot be expected. Therefore, to reduce the capacity Ms of the auxiliary storage device,
Data transfer time T by reducing the capacity Mb of the data block.
It can be seen that t should be reduced.

However, if the data transfer time Tt is reduced, it is necessary to increase the number of storage devices according to the equation (2), which leads to an increase in cost and space. Therefore, the number of storage devices required for the storage subsystem and the capacity of the auxiliary storage device are determined by their respective balances.

Next, an example of a memory configuration of the auxiliary storage device will be described with reference to FIG. In the example of FIG. 6, the auxiliary storage unit connects the continuous data divided into the seven storage units # 1 to # 7 and records them as a data block having a size corresponding to the repetition period Tc. Using a data read pointer PT that can be set at an arbitrary timing, data can be reproduced, for example, immediately after writing from the start point STP or after, for example, Tc time. Further, in response to a read request for four sets of continuous data, data is simultaneously recorded and reproduced.

In this example, the storage subsystem consisting of seven storage devices # 1 to # 7 is sequentially accessed by a continuous data read request, and the address numbers stored in the storage subsystem are in the 000's. The figure shows a case where four consecutive data starting from the 100s, the 200s, and the 300s are being read.

Access requests 1, 2, and 4 are generated at the same timing in the storage device # 1, and in the access request 1 generated from the earliest continuous data read request, the storage device originally has the data. The data is read at the point that is the most retroactive from the position from which the data is read. Therefore, access request 1
Is located farthest from the start point STP. Next, the generated access request 2 originally follows the access request 1,
Since access request 3 generated thereafter occurs at a timing without duplication, priority is given to access request 2 and access request 2
After. Then, the last access request 4 which is generated almost immediately receives data transfer directly from the storage device.

The capacity of the auxiliary storage device shown in the equation (4) indicates the total capacity of the memory configuration of FIG. 6, and in practice, the start point ST in the memory configuration of FIG.
It is sufficient if a storage capacity corresponding to the position of the pointer PT from P is secured. If the read request for the continuous data is generated with the timing shifted, the start point and the position of the pointer almost coincide with each other as in the case of the access request 3, so that the auxiliary storage device requires almost no capacity. On the other hand, the maximum storage capacity is required when read requests for continuous data are concentrated at the same timing.

In this case, by the operation of the reading section 4,
The positions of the pointers are determined in a state where the pointers are dispersed at substantially equal pitches from the start point. Therefore, it can be seen that, in practice, the storage capacity of the auxiliary storage device 5 is only half the value obtained by the equation (4).

More specifically, about 100 two-hour programs are stored in a magnetic disk drive, and at the same time, 100 users are served.
Let's find the capacity of the auxiliary storage device assuming the service.

Maximum access time Ta of the magnetic disk drive
= 35 ms, video transfer rate = 24 Mit / s, data transfer time 15 ms, and allowable multiple access number Na = 100. If the total storage capacity is 20 Tbits, the storage subsystem is 334 units from equation (2). According to the equation (4), the required capacity of the auxiliary storage device 5 is 12 Gbits.
About 6GBits (1 of total capacity)
/ 3000).

Further, in order to reduce the storage capacity, the start point SPT and the pointer PT
It is only necessary to reduce the distance between them. For example, a request for reading continuous data may not be accepted at the same timing within a cycle Tc of one cycle through the storage device, and a slight restriction may be added to the request for reading continuous data.

Since the auxiliary storage device 5 needs to be able to read data at an arbitrary timing as described above, a FIFO (First In First Out) type semiconductor memory is suitable.

Further, in the above embodiment, the example in which the synchronization control unit 2 for synchronizing the storage devices is provided is shown. However, since the synchronization deviation between the storage devices can be absorbed by the auxiliary storage device 5, the synchronization control unit 2 can be used. The control unit 2 is not essential for implementing the present invention.

It is desirable that the start positions of a plurality of continuous data are assigned to the respective storage devices as # 1, # 2, # 3,....

[0041]

As described above, according to the first aspect of the present invention, in a storage subsystem in which long data is distributed and held in a plurality of storage devices, a plurality of successive data generated sequentially are read. The request can be permitted at an arbitrary timing and up to the maximum capacity of the storage subsystem [(transfer time of one data block × number of storage devices) / (data transfer time + maximum access time)]. Therefore, a wait state does not occur for a plurality of continuous data read requests as compared with the conventional storage subsystem.

According to the second aspect, in addition to the effect of the first aspect, the capacity of the auxiliary storage device can be reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing one embodiment of a storage subsystem according to the present invention.

FIG. 2 is a configuration diagram of a conventional storage subsystem.

FIG. 3 is a diagram illustrating the principle of a conventional configuration.

FIG. 4 is a diagram illustrating the principle of a storage subsystem according to the present invention.

FIG. 5 is a diagram illustrating the principle of a storage subsystem according to the present invention.

FIG. 6 is a diagram showing a memory configuration example of an auxiliary storage device according to the present invention.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 storage device unit 2 synchronization control unit 3 storage device switching unit 4 readout unit 5 auxiliary storage device

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Yutake Sato 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation (72) Inventor Toruichi Arai 1-16-1 Uchisaiwaicho, Chiyoda-ku, Tokyo Within Nippon Telegraph and Telephone Corporation (72) Inventor Junichi Kishigami 1-6-1 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation (72) Inventor Norihiko Sakurai 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation Telephone Co., Ltd. (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 3/06 301

Claims (2)

(57) [Claims] 1. A method of reading a continuous data consisting of a plurality of data blocks across a plurality of storage devices while combining a plurality of storage devices capable of independently performing read and write operations and switching storage devices to be accessed. In a storage subsystem that performs writing, data to be accessed next in one storage device by a continuous data read request.
A reading unit that reads out the data block in advance using the idle time of access to the storage device; and an auxiliary storage device that holds the previously read data and reads the held data by the next access. A storage subsystem, comprising: 2. The read unit, when allocating the execution timing of access to each storage device within a repetitive read cycle in which all the storage devices are cycled, includes:
The next data block to be accessed in the storage
Utilizing the free time of access to the one storage device
Timing of reading in advance and holding it in the auxiliary storage device
At the same time, another continuous data read request is issued
Is an access related to the other continuous data read request.
Is stored in the auxiliary storage device.
The storage subsystem according to claim 1, wherein the storage subsystem is assigned prior to the access to the storage subsystem.