JP4510785B2 - Database server - Google Patents

Description

  The present invention can store a large amount of medical image information taken by a modality device such as a CT apparatus or an MRI apparatus as electronic information, and can search an image database system that contributes to diagnosis or a large amount of document files as electronic information. The present invention relates to a database server incorporated in a document database system stored in the database.

  Conventionally, in a medical image database system, in order to permanently store image data, image data temporarily stored on a hard disk device has been moved to a removable medium such as a DVD or an MO. The transfer of medical records to removable media is dependent on human resources, resulting in inconveniences such as increased costs and loss of image data due to human error. Network connection means, a hard disk as first image storage means for temporarily storing image files, removable media as second image storage means for permanently storing image files, and storage of image files received via the network Control means for controlling processing, and in response to receiving an image file to which a recording instruction is given, the control means sends an image to the designated one of the first or second image storage means. An image management system for storing files has been proposed.

JP 2002-111987

  However, according to the conventional image management system described above, a large amount of permanent storage media is required by permanently storing a large amount of image data in a medium different from the primary storage medium. In addition to requiring management, there is an essential problem that image data once stored on a permanent storage DVD or MO cannot be quickly accessed on the network when necessary.

  In view of the above-described problems, the present invention provides a database that can be quickly accessed on a network without using a medium different from the temporarily accumulated medium, and can safely and stably back up data without human intervention. The object is to provide a database server used in the system.

In order to achieve the above object, a first characteristic configuration of the database system according to the present invention includes a first switch for processing a job related to an external request, as described in claim 1 of the claims. a first network through, more second network via a second switch that handles the job related to the internal request, load balancer to load balance control, and a network connection to the storage unit and each other, the storage Stores records of attribute information including key information associated with documents or image data stored in the unit, and stores addresses of documents or image data stored in the storage unit with externally connected devices the a database server that relays treatment, versus the first network In addition to the data update port and the data search port, the update process is performed on the database by adding the update date and time to the update record for the data update process performed via the first network. And an update processing means for updating the database of another image database server by adding an update date and time to the update record via the second network, and an update port and a search port of itself when recovering from a failure Close both of them, obtain records from the other database server from the last update date and time until the recovery start date and time updated by itself through the second network, update their own database and set the recovery start time Update process to make the last update date and time of own database, the last update date and the current date and time are within the predetermined time Until the last update date and time and the current date and time are within the specified time, another database server is requested to update the database after its own last update date and time, and the database update processing after the last update date and time is performed A recovery processing means for opening a data update port and a data search port later is provided.

  According to the above-described configuration, the database server that recovers from the failure updates the own database by acquiring records from other database servers until the last update date and time and the current date and time are within a predetermined time, Thereafter, a record within a predetermined time is updated from the database server that is operating normally.

As described in claim 2 of the same column, the second feature configuration closes its own data update port in response to a request from a database server that is recovering from a failure in addition to the first feature configuration described above. Then, after performing a database update process after the last update date and time of the database server being recovered from the failure via the second network, a recovery processing means is provided for opening a port for updating the data itself . It is in.

  As described above, according to the present invention, a database that can be quickly accessed on a network without using media different from the primary storage media and can safely and stably back up data without human intervention. A database server used in the system can be provided.

  An embodiment of an image database system according to the present invention will be described below. As shown in FIG. 1, the image database system includes a basic frame (basic minimum tower) and a plurality of expansion frames (additional towers).

  The basic frame is stored in each storage unit 1, a storage unit group consisting of four storage units 1 that store medical image data, a control unit group consisting of two control units 2 that manage each storage unit 1, and Image database server comprising two image database servers 3 that store attribute information including key information associated with the medical image data and relay the medical image data with the externally connected viewer 11 And a DICOM gateway group including two DICOM gateways 4 for relaying medical image data by the DICOM protocol between the image database server 3 and a plurality of modality apparatuses 10 externally connected to the image database server 3. Connected to each other via network Configured Te, provided with a two load balancers 7 for load balancing control them per group based on the header information of the request over the network, and provided with power supply unit 8 and the display switch 9. Here, the number of each unit or server is merely an example, and may be more than or less than the number.

  The expansion frame includes a power supply unit, a plurality of storage units, and a switch, and is configured to be expandable so that large-capacity image data can be stored.

  The storage unit 1, the control unit 2, the image database server 3, and the DICOM gateway 4 are connected to the first network via the first switch 5 that processes jobs related to external requests from the viewer 11 or the modality device 10. N1 and the second network N2 through a second switch 6 for processing jobs related to internal requests are connected to each other via a network. The networks N1 and N2 are OSI reference standard models, and the physical layer and data link layer of the network are configured with 1000BASE-T Ethernet (registered trademark), the transport layer and the network layer are configured with TCP / IP, and more than the session layer. The upper layer is configured to operate with the DICOM protocol, which is a diagnostic imaging standard in the medical field, or other local upper protocols.

  The two load balancers 7 are connected by a local communication line (RS232-C), and normally one unit operates. For example, if there is no difference in the performance of the components of each group, there is a difference in round robin method and performance. For example, a load balancing control is performed so that the load of each component is equalized by a weighted round robin method, response time algorithm, etc., and when a failure is detected via the local communication line, a backup load balancer is activated. It is configured. Here, as a local communication line, it is possible to adopt a communication line according to another communication standard other than the one using RS232-C, and further, local communication is performed via the first or second network described above. It may be a thing.

  Hereinafter, the operation of the present system will be described by taking, for example, an image storage request from the modality device, an image read request, a response operation to an image browsing request from the viewer, and the like.

  The modality device 10 starts an association by transmitting an image accumulation request by the DICOM protocol at a virtual IP address (VIP) assigned to a group constituted by two DICOM gateways 4. The load balancer 7 selects the DICOM gateway 4 that operates normally and has the smallest load, intercepts the packet having the virtual IP address as the destination address, and rewrites the actual IP address of the selected DICOM gateway 4. The image accumulation request is received by the selected DICOM gateway 4 by the above processing. The return packet can be communicated by the reverse operation. Here, the load of the DICOM gateway 4 is distributed by the load balancer 7 in units of one association.

  As shown in FIG. 2 and FIG. 3, the DICOM gateway 4 that made the request uses the received image information as an examination example UID (consisting of a vendor prefix code, product unique number, serial number, etc.), patient name, patient ID. The image data is divided into attribute information and image information including the receipt number, examination date, examination time, and date of birth of the patient, and a predetermined lossless compression algorithm (here, PNG (Portable Network Graphics) is used) A plurality of compressed images with different compression ratios (in this case, 512 × 512 pixel original image, 256 × 256 pixel 1/4 image, 128 × 128 pixel 1/16 image) Four types of images of 1/64 images of 64 × 64 pixels used as thumbnails are generated and stored. Are not limited to.) Of compresses each, transferred to the control unit 2.

  The control unit 2 receives the transferred image data, stores the image data according to an algorithm described later, and returns the storage address to the DICOM gateway 4. The DICOM gateway 4 transfers the attribute information to the image database server 3 together with the storage address of the image data returned from the control unit 2, and instructs data update processing.

  The control unit 2 and the image database server 3 involved in the above-described operation are communicated via the first network N1 and are load-balanced by the load balancer 7 in the same manner as the DICOM gateway 4. Each operation described below is basically load-balanced by the load balancer 7 in the same manner as described above, but the description of the operation of the load balancer is omitted hereinafter.

  Next, when the DICOM gateway 4 receives an image provision request from the modality device 10, as shown in FIGS. 4 and 5, the DICOM gateway 4 receives from the image database server 3 based on any key information of the attribute information that identifies the request image. Acquire attribute information and storage address, acquire a compressed image reversibly compressed from the storage unit 1 based on the storage address, decompress the compressed image, and send it back to the modality device 10 by the DICOM protocol together with the attribute information To do.

  As shown in FIG. 6 and FIG. 7, when there is an image data browsing request from the viewer 11 to the image database server 3, the image database server 3 receives the key information of any attribute information that identifies the request image. The attribute information and storage address are acquired and returned to the viewer 11. Thereafter, the viewer 11 accesses the storage unit 1 based on the received storage address and acquires desired image data. Here, if the viewer 11 is a DICOM viewer, the operation is performed through the DICOM gateway 4 described above. However, although the DICOM protocol is not used, if a local protocol by TCP / IP is supported, an arbitrary upper protocol can be selected. Can be used.

  When the power is turned on, the control unit 2 does not go through the load balancer 7 (external access to the control unit 2 is load-balanced by the load balancer 7). Operation status inspection means for executing the operation status inspection via the operation status table and updating and storing the result in the operation status table of the image database server 3, and based on the update history of the operation status table data by the other control unit 2. A failure diagnosis means for performing a failure diagnosis of the other control unit is provided.

  More specifically, the operation status checking means is installed in units of frames and in order with respect to the storage unit 1 that is mounted on the frame and includes several hard disks, that is, the storage unit installation status, that is, Inquires about the installation status of whether the unit is removed or installed, the remaining disk capacity for each drive, and the total remaining capacity of the unit when installed, and writing any file to each drive, Reading is performed, the disk failure of each drive is investigated, and the result is updated and stored in the operation status table shown in FIG. Although not explicitly shown in FIG. 13, management is possible in units of frames by adding a frame ID to the upper data of the storage unit ID.

  In response to a data storage request to the storage unit 1, the control unit 2 selects and stores the data storage destination from the normally operating storage units identified based on the operation status table for the number of redundancy. Means are provided. More specifically, the redundant storage means examines the remaining disk capacity for each drive based on the operation status table, and selects and stores a drive having a large remaining disk capacity in different storage units by the number of redundancy. Here, the redundancy number is set to two. Accordingly, the same data always exists in different storage units 1 in the system (here, it is needless to say that the redundancy number may be set to 2 or more).

  According to another method, the remaining disk capacity is investigated for each storage unit, and two storage units 1 having a larger remaining disk capacity are selected, and a drive having a larger remaining disk capacity is selected from the selected storage units 1. You may select and store data. The data storage address is returned to the transmission source of the storage request. The image data storage request is returned to the DICOM gateway 4 and stored in the database server 3 together with or as attribute information of the image data by the DICOM gateway 4.

  According to still another method, the redundant storage means selects a plurality of storage units 1 having a remaining capacity of a predetermined capacity or more as candidates based on the operation status table, and selects a target based on the random number generated by the random number generation means. Alternatively, two storage units 1 may be selected. For example, a number of storage units greater than the number of redundancy is selected as candidates, random numbers 1 to 9 are sequentially generated by the random number generation means and assigned to each candidate storage unit, and the generated random number value is This can be achieved by selecting as many storage units as possible from the largest. In this case, it is possible to prevent the load from concentrating on the newly added storage unit.

  According to still another method, the redundant storage means is configured to execute the redundant storage processing by any of the methods described above for the storage units mounted in different frames based on the operation status table. May be. In this case, the influence from disasters such as fires, earthquakes, and floods can be kept low by making the locations and buildings where the frames are installed different.

  In addition, when the redundant storage means checks the remaining disk capacity based on the operation status table, it searches for a larger remaining capacity than the maximum capacity assumed for the stored image data. On the contrary, when the remaining capacity is smaller than this value, data indicating that the drive is full is stored in the remaining capacity data of the operation status table, and when all the drives of the storage unit 1 are full, the storage unit By configuring 1 to be excluded from the selection target, the time for investigating the remaining disk capacity can be shortened.

  Further, if there is a storage unit 1 or a drive that is determined to be abnormal by the operation status inspection unit, the control unit 2 is determined to be abnormal based on the storage address stored in the database server 3 in order to recover the data. Obtain the data storage address of another storage unit that stores the same data as the data stored in the storage unit or drive, and store the data stored in that data storage address as the same data. A storage unit 1 other than the storage unit 1 selected is selected from the normally operating storage units 1 identified based on the operation status table and copied. In this way, even if any storage unit fails or is removed, the image data is automatically duplicated, and if a new storage unit is added, operation for the added storage unit is performed. Since the status table is automatically generated, maintenance work such as adding or removing storage units can be performed without turning off the system power.

  The failure diagnosis means of the control unit 2 accesses the operation status table regularly or irregularly, and based on the update history of the operation status table data by other control units 2, is performed within a predetermined time in the past from the time of access. When there is no update history, it is determined that the other control unit has failed. When a spare control unit is installed, one spare control unit is started up.

  Here, the operation state inspection means may be operated by one of the two operating control units, and the failure diagnosis means may be operated by the other control unit. In this case, a flag indicating the completion of the operation of the failure diagnosis means by the other control unit and the operation time data are recorded in the operation status table, and the one control unit checks the flag and the operation time data. Thus, the failure of the other control unit can be detected.

  The control unit 2 may be integrally configured on the same substrate as the storage unit 1. In this case, for example, two storage units can be set to be active, and the spare control unit can be switched to active when one failure is detected by the failure diagnosis means.

  Hereinafter, update processing and failure recovery processing of the image database server 3 will be described. The image database server 3 updates its own database in response to the data update processing performed via the first network N1, and also stores the database of the other image database server 3 via the second network N2. Update processing means for performing update processing is provided.

  The image database server 3 is provided with an update port and a search port for the load balancer 7, and both ports are open when normal. Further, the update processing means of one image database server 3 (DB1 in the figure) manages the final operation time of the update processing means of the other image database server 3 (DB2 in the figure), and all the tables are all stored. The last update date / time is written and managed in the record.

  As shown in FIGS. 8 and 9, the update processing means receives the storage request for the attribute information and image data storage address from the DICOM gateway 4 via the load balancer 7, and the operation status table from the control unit 2. When there is a data update request such as a storage request (circled step 1 in FIG. 8), it updates its own database (circled step 2 in FIG. 8), and then others via the second network 2 The database of the image database server 3 is updated (step 3 in FIG. 8). When this operation is completed, an acknowledgment for the request is returned (step 4 in FIG. 8).

  The image database server is provided with a data update port and a data search port for the first network N1, and is used for data update at the time of recovery processing for the database of another image database server via the second network N2. Recovery processing means for closing the port is provided.

  This will be described in detail below. When the other image database server 3 is out of order, both the update port and the search port are closed, and the load balancer 7 requests processing from only one image database server 3. As shown in FIG. 10, in this case, only the update process of its own database is performed for the update request, and an acknowledge is returned.

  At the time of recovery from the above-mentioned failure state, one image database server 3 performs normal operation, and both ports of the other image database server 3 are closed. As shown in FIGS. 11 and 12, the other image database server 3 that recovers from the failure acquires the last update date and time updated by the other image database server 3 (step 1 in FIG. 11), The recovery start time is recorded (step 2 in FIG. 11). Next, the record of the update date and time acquired in the first step is acquired from one image database server 3, and its own database is updated (step 3 in FIG. 11). The recovery start time recorded in step 2 is used as the last update date and time of its own database (circled step 4 in FIG. 11). Repeat steps. Here, n is on the order of several seconds. If it is within n seconds, a restoration operation is requested to one image database server 3 (step 5 in FIG. 11). On the other hand, one image database server 3 closes the update port (step 6 in FIG. 11), and performs update processing of the database after the last update date and time of the other image database server 3 (in FIG. 11). Step 7), the update port is opened after completion of the update process (Step 8 in FIG. 11), and an end notification is transmitted to the other image database server 3. The other image database server 3 opens both ports and returns to normal operation (steps 9 and 10 in FIG. 11). In the above description, the update port and the search port are separately configured. However, the update port and the search port may be configured as the same port.

  Another embodiment will be described below. In the above embodiment, a system has been described in which the upper layer above the session layer operates with the DICOM protocol, which is a diagnostic imaging standard in the medical field, or other local protocols. However, a terminal such as a viewer is compatible with WEB. It is also possible to configure. In this case, for example, a WEB server group composed of a plurality of WEB servers may be provided in the basic frame described above, and configured to operate under the management of the load balancer 7.

  In the above-described embodiment, the storage unit is configured as a hard disk. However, the storage unit is not limited to the hard disk, and may be a nonvolatile semiconductor memory or other media. .

  In the above-described system, any or all of the storage unit, control unit, image database server 3, and DICOM gateway are configured to be hot swappable on a unit basis, so that maintenance can be performed while maintaining the operating state of the system.

  In the embodiment described above, the image database system for accumulating and managing medical image data has been described, but it goes without saying that the present system is applicable to a database system for accumulating and managing various document data other than images.

  The above-described embodiment is merely an example of the present invention. As described in the section for solving the problem, a plurality of database systems are constructed by combining each component.

Configuration diagram of an image database system according to the present invention Illustration of image storage procedure Flow chart of image storage procedure Illustration of image provision procedure Flow chart of image provision procedure Explanatory diagram of image search procedure by viewer Flow chart of image search procedure by viewer Operation explanatory diagram of image database server Flow chart of image database server Operation explanatory diagram of image database server Operation explanatory diagram of image database server Flow chart of image database server Explanatory drawing of operating status table

1: Storage unit 2: Control unit 3: Image database server 4: DICOM gateway 5, 6: Switch 7: Load balancer 10: Modality device 11: Viewer

Claims (2)

A first network via a first switch that handles the job related to external requests, and more second network via a second switch that handles the job related to the internal request to the load balancing control load Stores records consisting of attribute information including key information associated with the document or image data stored in the storage unit and connected to the balancer and the storage unit, and between externally connected devices. A database server that relays a storage address of documents or image data stored in the storage unit ;
A data update port and a data search port for the first network, and an update date is added to an update record for the data update processing performed via the first network. Update processing means for updating the database of another image database server by adding an update date and time to the update record via the second network,
Closes both its own update port and search port when recovering from a failure, and acquires records from the other database server from the last update date and time until the recovery start date and time updated by itself via the second network Update the own database and repeat the update process with the recovery start time as the last update date and time of the own database until the last update date and the current date and time are within a predetermined time. When the time is up, the database update process is requested to the other database server after its last update date and time, and the data update port and data search port after the database update process after the last update date and time is performed A database server equipped with a recovery processing means for opening a database. In response to a request from a database server that is recovering from a failure , the database update processing after the last update date and time of the database server that is recovering from the failure is closed via the second network by closing its own data update port. 2. The database server according to claim 1, further comprising recovery processing means for opening a port for updating the data of the data after performing the processing.

Images