Multi-path redundant storage system architecture and method

The present invention relates generally to data storage systems, especially to system and methods of interconnecting storage components in fault-tolerant data storage systems.

A data storage system may consist of one or more disk drives connected to one or more disk controllers and one or more disk controllers connected to a host or network interface. Each component of the storage system, such as disk drives, controllers, connectors and wiring, is a potential failure point in the system. Some systems, such as personal computers, can become inaccessible to data in the event of a controller, bus or connector failure. Access to the data may require repairing or replacing the failed component, or installing a disk drive on another system to access the data. Disk drive failures usually lead to the loss of stored data. Large-scale storage systems can also employ redundant methods such as RAID, which distributes data across multiple drives so that data is not lost in the event of a drive failure. In a RAID system, the data from the failed drive may be copied from the mirror drive or reconstructed from the data and parity information on the operating drive. After a disk or disk controller failure, the system often operates under degraded performance conditions until the failed component is replaced or repaired. Bus failures may require the drive to be removed and installed in other fixtures or systems in order to access data.

Fault tolerance levels, storage capacity, operating life and data availability are important determinants of the value of a storage system. Fault tolerance may be expressed in relation to the number of disk, controller, and bus failures (sequential and simultaneous) that occur while maintaining data integrity and data access. The storage capacity reflects the number of disk drives, the capacity of each drive, and the data encoding method used. As the number of drives increases, so does the number of interconnects and the likelihood of failure. The operating life of a storage system is reflected in the life of the component and the fault tolerance level of the system. Spare disk drives may be employed to store copied or reconstructed data to extend the operation of the system after a disk drive failure. Data availability may be expressed in terms of data transfer rate, fault tolerance, and system performance after one or more component failures.

The commercial viability of the storage system reflects the configuration decisions and component selections made by the designer to obtain the desired level of fault tolerance, storage capacity, operating life and data availability. are doing. Components with a very long mean time between failure (MTBF) rate can be counterproductive to system cost.

<p> The embodiments of the present invention provide fault tolerance in a data storage system and provide redundant storage system configurations and isolation methods that can be employed to eliminate single point failures.</p>

<p> Accordingly, embodiments of the present invention include a plurality of data configurations that are internally located, provide a plurality of signals, and have at least one connector having at least one independent signal for each data storage of the plurality of data storage devices. A large number of disk assemblies, including storage devices, and a large number of disk assembly receptacles designed to accept an assembly with a mounting connector that engages at least one connector, at least one disk controller, and a fabric. Is in the first configuration when at least one independent signal for each data storage device in multiple data storage devices can be selectively connected to the disk controller and the structure is in another configuration. May include a data storage system with at least one structure that can be configured to selectively disconnect at least one independent signal.</p><p> Further, an embodiment of the present invention includes a plurality of data storage devices arranged in a disk assembly, a connector for communicating a signal from the disk assembly to a fixture adapted to accept the disk assembly, and a disk. A structure that is located in the assembly and communicates with the connector so that it can be constructed to selectively connect and disconnect at least one of the data storage devices for at least one signal of the connector. It may further include a large number of disk assemblies that it has.</p><p> Further, in the embodiment of the present invention, a plurality of data storage devices arranged as a plurality of pairs arranged in an assembly having at least two pairs of data storage devices, and each data of the plurality of data storage devices. It may further include a removable data storage assembly with a connector that provides external communication for at least one independent signal with respect to the storage device.</p><p> Further, the embodiments of the present invention include a plurality of dual port data storage devices, have at least one connector for communicating at least two independent signals to the fixture, and each data of the plurality of data storage devices. It has a first structure that can be constructed to connect the first port of the storage device to the first signal of at least two independent signals, and the first of each data storage device of the data storage device. To accept a large number of disk assemblies with a second structure that can be constructed to connect two ports to the second signal of at least two independent signals, and an assembly with a mounting connector that engages at least one connector. A data storage system with a large number of disk assembly receptacles and at least one disk controller that can access at least one of the data storage devices through the mounting connector. Further may be provided.</p><p> Further, the embodiments of the present invention are connected to a large number of disk assemblies including multiple data storage devices installed in at least one structure connected to a large number of disk assembly receptacles and assemblies. A method of constructing a data storage system having at least one structure can be further provided, the method of detecting an error in the data storage system and being included in the assembly as inoperable. It comprises a step of recognizing one of the data storage devices and constructing at least one structure to disconnect at least one data storage device.</p><p> In addition, embodiments of the present invention include a plurality of data storage devices and a number of disks having connectors that provide at least one individual signal line for each pair of data storage devices of the plurality of data storage devices. It has an assembly and has a fixture connected to a host system with a disk controller and internally placed structures, the fixture having a large number of disk assembly receptacles adapted to accept the assembly. Structures to have and communicate signals between them, detect errors in the storage system, recognize inoperable data storage equipment in the assembly, and further disconnect the inoperable data storage equipment. A data storage system with a computer program that can operate to build a data storage system may be provided.</p><p> Further, the embodiments of the present invention can be constructed so as to connect and disconnect each data storage device of the data storage to at least one signal of a plurality of data storage devices and a connector for communicating signals outside the assembly. A fixture with a large number of disk assemblies containing at least one structure and a large number of disk assemblies that have an internally located disk controller and that are adapted to accept and communicate with the assembly. Computer program code that detects errors in the storage system, recognizes the inoperable data storage device in the assembly, and builds at least one structure to disconnect the inoperable data storage device. An additional data storage system may be provided.</p>

The embodiments of the present invention, along with redundant components and data paths, provide failure point isolation within the storage subsystem so that data access can be maintained after a bus or component fails. Failures can occur most often in connectors and components that have moving parts, such as disk drives. In general, electronic components such as integrated circuits are likely to exhibit lower failure rates than connectors or disk drives.

Examples of the present invention include different storage systems that employ a single or multiple disk arrays installed in a cabinet fixture and a system that employs a large number of removable disk assemblies. Applicable to architecture. A large number of disk assemblies are defined as installable units that allow the removal of a given size, shape and connector configuration that may include different internal data storage devices, components and configurations. In one embodiment, multiple disc assemblies may include a first number of 3.5 inch (8.88 mm) discs, while other embodiments have different numbers of 2.5 inch (6.) discs. 35 mm) discs may be provided. A large number of different disk assembly examples may be installed in a single fixture design. This makes it possible to create systems with different storage capacities, data speeds and processing powers using a single fixture (cabinet, shelves, etc.). Examples of multiple disk assemblies include multiple disks, one or more structures, one or more disk controllers, and one or more interface controllers from multiple units containing only multiple disks and multiple connectors. It may be changed to a composite structure within the range of a plurality of provided units. Many disk assembly examples employ interfaces such as Fiber Channel that enable devices ranging from simple multiple storage devices to intelligent controllers, for example, while adopting the same connector. You may. The computer program code that runs on the host and other systems reflects the complexity of many disk assemblies. A large number of disk assemblies may simplify and upgrade simple system assemblies and may reduce the possibility of radio frequency emission. A large number of disk assembly receptacles are defined as receptacles in shelves, racks, enclosures or other fixtures, which remove a large number of individual disk assemblies that are mutable within the internal architecture. Can be installed as possible. The embodiments of the present invention may be employed to form a storage system in which a large number of disk assemblies can be considered as "maintenance-free" storage devices. Many disk assembly examples have one or more spares so that they can operate for an extended period of time without user intervention, even after the bus, controller, and / or one or more drives have failed. It can provide drives, multiple buses and spare controller capacities. In the embodiments of the present invention, after a component failure

Figure 1 shows the configuration of a single-port data storage system. System 100 includes host 102, disk array controller A 104, disk array controller B 106, bus A 108, bus B 110, A drive array 112, and B. "Equipped with a drive array 114. Drive arrays are described as having 5 drives each other. The A drive array 112 and the B drive array 114 are single ports and provide a single interface to bus A 108 or bus B 110. The disk array controller A 104 and the disk array controller B 106 are connected to host 102 by one or more buses and are dual-port, which are two disk drive buses. Provide each interface. The interface of each drive array controller is constructed so that if one of the controllers fails, one of the controllers can support communication on both bus "A" 108 and bus "B" 110. The controller. Depending on the number of disk drives in each array and the data transfer rate of the drives in the array, the system may operate at a slower data rate after one of the controllers fails. Failure of either bus "A" 108 or bus "B" 110, defective bus signals due to associated connectors, or connected components, provides complete access to the data stored in the array attached to the bus. Ban. Such bus A 108, bus B 110, and related connectors and mounting components that may damage the bus represent a single point of failure. Reproduction of stored data requires repairing the bus or removing the disk drive and installing a functioning bus on the fixture. From the point of view of data availability, the architecture in Figure 1 is a controller failure.

Figure 2 shows the architecture of a dual-port data storage system. System 200 includes host 202, disk array controller A 204, disk array controller B 206, bus A 208, bus B 210, and B drive array 212. There is. The disks in the "B" drive array 212 are dual-port, providing dual ports that provide a single interface for bus "A" 208 and bus "B" 210, respectively. The disk array controller A 204 and the disk array controller B 206 are connected to the host 202 by at least one bus and, in a preferred embodiment, by at least two buses. The disk array controller A 204 and the disk array controller B 206 are dual-port and each provide two disk drive bus interfaces. The interface of each drive array controller is built to support communication on both bus "A" 208 and bus "B" 210 and operates continuously in the event of one of the controllers failing. I will provide a. The dual port characteristic of the "B" drive array 212 is that its availability in the array allows it to communicate with either drive array controller. The system continues to provide data access in the event of a bus or controller failure. Access may slow down according to the transfer rate and the number of drives in the array. The architecture shown in Figure 2 has the advantage of maintaining continued data availability after a bus failure compared to the system in Figure 1, but at the cost of using dual-port disk drives. There is. The architectures in Figures 1 and 2 include parallel or series bus interfaces such as, for example, SCSI, series SCSI, series ATA, or fiber optic channels.

Figure 3 shows the configuration of the loop storage system. System 300 includes a host 302, a drive array controller 304, a bus 306, and a drive array 308. The drive array controller 304 is connected to the host 302 by one or more buses. Bus 306 interconnects the drives of the drive array controller 304 and the drives of the drive array 308 in the loop in series. The drive array controller 304 and the drive array 308 connect the input port and the output port so as to form a loop of the bus 306. The system in Figure 3 can continue to operate in the event of a disk failure and does not affect bus operation. A bus, controller failure, or disk failure that disrupts the operation of the bus results in a loss of data availability, repairing the bus, controller or disk drive, or other fixtures to access the data. Requires drive installation.

Figure 4 shows the configuration of a storage system with interchangeable single-port disk drives. System 400 includes host 402, disk controller "A" 404, disk controller "B" 406, switch control 408, bus "A" 410, bus "B" 412, disk drives 414-422 and switching device 424. It has ~ 432. The disk controller "A" 404 and the disk controller "B" 406 are dual-port type, which are connected to the host 402 by one or more buses and each provide a two-disk drive bus. Bus A 410 and bus B 412 are connected to both disk controller A 404 and disk controller B 406. In another embodiment (not shown), two single-port disk controllers in which the first disk controller has bus "A" 410 and the second disk controller has bus "B" 412. May be used. Switching devices 424 to 432 are controlled by switch control 408 and have disk drives 414 to 422 subordinately connected to bus "A" 410 or bus "B" 412. The switching devices 424 to 432 include, but are not limited to, a switching device of any type, a port multiplexer, or the like. The switch control may be provided with one or more buses connecting the switching devices 424 to 432 to the host 402, and may be provided with an I2C bus, an RSC232, or another series or parallel bus. Alternatively, the switching device may be controlled by disk controller "A" 404, disk controller "B" 406, or both. In other embodiments, the switch control may employ bus "A" 410 and / or bus "B" 412. Therefore, the switching device may be bus "A" 410 or by host 402. It may be controlled directly by the host 402 via the bus B 412, or may be controlled by the bus A 410 or the bus B 412. The architecture of FIG. 4 may employ a large number of discs and a switching device than shown. The configuration of FIG. 4 may employ more disks and switching devices than shown.

The switching device can be constructed individually for each drive so that each drive employs either bus "A" 410 or bus "B" 412. This makes it possible to maintain communication in the event of a bus failure and balance the load between the buses. The configuration of FIG. 4 provides continuous operation in the event of a bus, disk or controller failure. In addition, switching devices 424-432 allow the disk drive to be detached from both buses. In the event of a disk failure, or a disk failure that causes the bus to malfunction, the associated switching device may be constructed to disconnect the drive from both buses. The switching method shown in FIG. 4 may be applied to a dual-port drive in which each port of each drive can be selectively connected to or detached from both buses "A" 410 and "B" 412. .. Alternatively, a third bus may be employed to provide even higher transfer rates in the event of a bus failure.

Figure 5 shows the configuration of a storage system with interchangeable single-port disk drives. System 500 includes host 502, disk controller "A" 504, disk controller "B" 506, disk controller "C" 508, switch control 510, bus "A" 520, bus "B" 522, bus " It has multiple drive / switching units starting with C "524 and drive / switching unit 512 and ending with drive / switching unit 526. These examples are not limited to a particular number of drive / switching units. The drive / switching unit 512 is connected to the dual port drive 514, the first switching device 516 connected to the first port of the dual port drive 514, and the second port of the dual port drive 514. It is equipped with a second switching device 518. The storage device 516 makes the first port of the drive 514 connectable to the bus "A520", the bus "B" 522 or the bus "C" 524. Similarly, the switching device 518 makes the first port of the drive 514 connectable to bus "A" 520, bus "B" 522 or bus "C" 524. These switching devices are controlled via control logic, such as a bus interface such as 12C, or switch control 510, which may be another circuit that allows host 502 to control the functionality of each switching device. .. Alternatively, the switch control 510 may be connected to one or more disk controllers, or one or more buses. Disk controller A 504, disk controller B 506 and disk controller C 508 are dual, connected to host 502 by one or more buses and each providing two disk drive buses. It is a port type. Bus A 520 to 524 malfunctions 1 In case of disk controller failure In addition, all buses remain mobile and are connected to two ports of different disk controllers of disk controllers "A" 504 to 508, respectively. In another embodiment of the configuration of FIG. 5, the switching device connected to the first port of each disk drive is controlled by the first switch control and is connected to the second port of each drive. The device is connected to a second switch control. The first and second switch controls can be controlled directly by the host, controlled by the host through one or more disk controllers connected to the switch control, or by one or more disk controllers. It is possible to connect. The switching device may be employed to connect multiple drive ports to one of the buses, or to separate the ports from all buses. The switching device may include any device that can be constructed to provide the aforementioned functions, including switching, multiplexers, port controllers, cross point switches, structures and the like. It may be adopted to separate the cells. The switching device may include any device that can be constructed to provide the aforementioned functions, including switching, multiplexers, port controllers, cross point switches, structures and the like. It may be adopted to separate the cells. The switching device may include any device that can be constructed to provide the aforementioned functions, including switching, multiplexers, port controllers, cross point switches, structures and the like.

The configuration in Figure 5 allows system operation to continue after one or more disk controller failures. In addition, the configuration in Figure 5 allows the data load to be distributed among multiple disk controllers for optimal performance. According to the number of disk drives and the data rates of the disk drives, buses and disk controllers, the configuration in Figure 5 provides near optimum performance after a disk drive, bus or disk controller failure. can do. Therefore, the above configuration may be adopted in a system in which high performance is desired continuously after a failure of the disk controller bus.

Figure 6 shows the configuration of a loop bypass storage system. The system 600 includes a host 602, a disk controller 604, a switch control 606, drives 608 to 616, switching devices 618 to 626 and bus 630. Disk controller 604 is connected to host 602 by one or more buses. The bus 630 connects the associated drives in series with the bus 630, or connects the disk controller 604 in series to each of the switching devices 618 to 626 that bypass the drives. When all switching devices are enabled, all drives are connected in series. These switching devices may be controlled by the host 602 or by the disk controller 604 via switch control 606. The configuration shown in Figure 6 allows individual disk connections to be bypassed so that in the event of a disk failure or a disk failure that affects bus operation, the failed drive can be bypassed or the system continues to operate. .. The switching devices 618 to 626 may be of any type capable of connecting or bypassing a plurality of disks in series. The switching devices 618 to 626 and the switch control 606 may be implemented as a single unit. The switching devices 618 to 626 and the switch control 606 may include a port bypass controller.

The loop bypass method may be employed to disconnect one or more drives. One or more drives may be connected to each port on the port bypass controller. Figure 7 shows a loop bypass storage system with two drives connected to each bypass controller port. System 700 includes a host 702, a disk controller 704, disk drives 706-724, a bypass controller 726 and a bus 728. For these drives, drives 706, 708 are connected to port 1 of port bypass controller 726, drives 710, 712 are connected to port 2, drives 714, 716 are connected to port 2, and drive 718. , 720 are connected to further ports, and drives 722, 724 are arranged in multiple pairs so that they are connected to further ports. Bus 728 connects disk controller 704 to port bypass controller 726. In another embodiment, the two buses may connect a disk controller and a port bypass controller to provide redundancy in the event of a bus failure. The port of any or port bypass controller 726 can be forwarded to or ported through two drives to which the signal is connected so that it can be disconnected in the event of a drive failure or a drive failure that causes the bus to fail. May be constructed to be bypassable. Although FIG. 7 shows two drives connected to each port of the port bypass controller 726, more than one drive may be connected within the scope of the present invention. Although FIG. 7 employs a port bypass controller, any device and configuration that produces the functions described may be employed.

The loop bypass configuration can employ multiple drives connected to each port if each drive is dual port. Figure 8 shows a loop bypass system with two dual-port drives connected to each port. The system 800 includes a host 802, a disk controller 804, a disk controller 806, a port bypass controller 808, a bus 810, a port bypass controller 812, a bus 814, and disk drives 816 to 824. The disk controller 804 and the disk controller 806 are each connected to the host 802 via one or more buses. The disk controller 804 is connected to the port bypass controller 808 via bus 810. Disk controller 806 is connected to port bypass controller 812 via bus b. In another embodiment, one or more buses may connect the disk controller 804 to the port bypass controller 808, and one or more buses connect the disk controller 806 to the port bypass controller 812. You may connect. In another embodiment, each disk controller may connect both port bypass controllers. The disk drives 816 to 824 are dual-port, and each drive has a first port connected to the port bypass controller 808 and a second port connected to the port bypass controller 812. Has. Thus, each disk drive may be individually constructed to connect to a loop formed by bus 810 on the first port, or bus 814 on the second port of the drive, or both buses. In case of disk failure or disk failure that impairs the bus signal, disconnect the drive by configuring port bypass controller 808 or port bypass controller 812, or by configuring both port bypass controllers. You may. In case of disk controller, bus failure, connector failure, or port bypass controller failure, data can be accessed from the driver using a functioning disk controller, bus, or port bypass controller. Good.

Two or more dual-port disk drives may be connected to each port of the port bypass controller. Figure 9 shows a loop bypass storage system with two dual port devices connected to a port bypass controller. System 900 includes a host 902, a disk controller 904, a bus 906, a port bypass controller 908, disk drives 910-928, a disk controller 930, a bus 932, and a port bypass controller 934. The disk controller 904 and the disk controller 930 are connected to the host 902 by one or more buses. The disk controller 904 is connected to the port bypass controller 934 via bus 906. The disk controller 930 is connected to the port bypass controller 934 via bus 932. Disk drives 910-928 are dual-port, and each driver has a first port connected to port bypass controller 908 and a second port connected to port bypass controller 934. .. In another embodiment, the disk controller 904 is further connected to the port bypass controller 934, and the disk controller 930 is also connected to the port bypass controller 908. Port bypass controllers 908 and 934 can be individually constructed to establish a connection to a disk drive port or to bypass a connection to a disk drive, resulting in a disk failure or failure that compromises the port connection. In the case of, each drive can be separated. The disk drive is dual-port and employs a two-port bypass controller, so in the event of a disk controller failure, the system in Figure 9 will have continuous operation.

Figure 10 shows a redundant storage system with multiple paths. System 1000 includes host 1002, host bus "A" 1004, host bus "B" 1006, disk controller "A" 1008, disk controller "B" 1010, structure bus "A" 1012, structure. -Equipped with bus "B" 1014, structure "A" 1016, structure "B" 1018, and disk drives 1020-1028. The disk controller A 1008 and the disk controller B 1010 are both connected to the host 1002 by the host bus A 1004 and the host bus B 1006. The disk drives 1020 to 1028 are dual port type by the first port connected to the structure "A" 1016 and the second port connected to the structure "B" 1018. Structure A 1016 and structure B 1018 may include any and all switch types and methods including Fiber Channel structures, switches, multiplexers, cross point switches and the like. .. Structures are address mapped It may have control) and may be controlled by host 1002 via disk controller A 1008 or disk controller B 1010. Alternatively, individual buses, or multiple buses (not shown), such as I2C, may transfer control and configuration information from host 1002 to structure "A" 1016 and structure "B" 1018. .. Further, the structure "A" 1016 and the structure "B" 1018 may be controlled and constructed in whole or in part by the disk controller "A" 1008 and / or the disk controller "B" 1010. Good. Configuration and control tasks may be shared between host 1002 and disk controller A 1008 and / or disk controller B 1010.

Figure 11 shows another multipath redundant storage system. System 1000 includes system interface 1102, system bus "A" 1104, system bus "B" 1106, interface controller "A" 1108, interface controller "B" 1110, interface bus "A" 1112, interface. -Bus "B" 1114, disk controller "A" 1116, disk controller "B" 1118, structure bus "A" 1120, structure bus "B" 1122, structure "A" 1124, structure "B" It is equipped with "1126", structure control bus "A" 1128, structure control bus "B" 1130, and drive groups 1132 to 1140. The interface controller A 1108 and the interface controller B 1110 are connected to the system bus A 1104 and the system bus B 1106. The two system buses include redundant communication paths, allowing continuous communication with both interface controllers in the event of one of the multiple system buses failing. The interface controller A 1108 and the interface controller B 1110 are connected to the disk controller A 1116 and the disk controller B 1118 via the interface bus A 1112 and the interface bus B 1114. Allows continuous communication between one of the interface controllers and one of the disk controllers if one of the interface buses fails. Disk controller A 1116 and disk controller B 1118 are connected to structure A 1124 and structure B 1126 via structure bus A 1120 and A drive array 1122. If one of the multiple structure buses fails, one of the structure controllers and the Izu Provides continuous communication with that structure. The structure control bus A 1128 and the structure control bus B 1130 are from the interface controller A 1108 to the interface controller B 1110 to the A drive array 1124 and the structure B 1126. A redundant control path is provided, and if one of the structure control buses fails, either of the interface controllers enables the construction of any of the structures. Structure A 1124 is connected to each drive group in drive groups 1132 to 1140 by individual connections. The drives in the drive group are dual port type. Structure B 1126 is connected to each group in drive groups 1132 to 1140 by individual connections. The structure "A" 1124 is connected to a dual-port drive, or the first port of multiple drives with each drive group, and the "B" drive array 1126 is a dual-port drive. , Or is connected to the second port of multiple drives with each drive group. The duality of system buses, interface buses, structure buses, structure control buses, and drive group connections provides independence or redundant paths for any data path in the system. The duality of the interface controller, disk controller, and structure related to the duality of the bus provides continuous operation in the event of an interface controller failure. Therefore, the system shown in FIG. 11 has no single point failure with respect to the bus, controller or structure. If the object control bus fails, either interface controller allows the construction of either structure. Structure A 1124 is connected to each drive group in drive groups 1132 to 1140 by individual connections. The drives in the drive group are dual port type. Structure B 1126 is connected to each group in drive groups 1132 to 1140 by individual connections. The structure "A" 1124 is connected to a dual-port drive, or the first port of multiple drives with each drive group, and the "B" drive array 1126 is a dual-port drive. , Or is connected to the second port of multiple drives with each drive group. The duality of system buses, interface buses, structure buses, structure control buses, and drive group connections provides independence or redundant paths for any data path in the system. The duality of the interface controller, disk controller, and structure related to the duality of the bus provides continuous operation in the event of an interface controller failure. Therefore, the system shown in FIG. 11 has no single point failure with respect to the bus, controller or structure. If the object control bus fails, either interface controller allows the construction of either structure. Structure A 1124 is connected to each drive group in drive groups 1132 to 1140 by individual connections. The drives in the drive group are dual port type. Structure B 1126 is connected to each group in drive groups 1132 to 1140 by individual connections. The structure "A" 1124 is connected to a dual-port drive, or the first port of multiple drives with each drive group, and the "B" drive array 1126 is a dual-port drive. , Or is connected to the second port of multiple drives with each drive group. The duality of system buses, interface buses, structure buses, structure control buses, and drive group connections provides independence or redundant paths for any data path in the system. The duality of the interface controller, disk controller, and structure related to the duality of the bus provides continuous operation in the event of an interface controller failure. Therefore, the system shown in FIG. 11 has no single point failure with respect to the bus, controller or structure. Connected to the second port of multiple drives with a drive group. The duality of system buses, interface buses, structure buses, structure control buses, and drive group connections provides independence or redundant paths for any data path in the system. The duality of the interface controller, disk controller, and structure related to the duality of the bus provides continuous operation in the event of an interface controller failure. Therefore, the system shown in FIG. 11 has no single point failure with respect to the bus, controller or structure. Connected to the second port of multiple drives with a drive group. The duality of system buses, interface buses, structure buses, structure control buses, and drive group connections provides independence or redundant paths for any data path in the system. The duality of the interface controller, disk controller, and structure related to the duality of the bus provides continuous operation in the event of an interface controller failure. Therefore, the system shown in FIG. 11 has no single point failure with respect to the bus, controller or structure.

In addition to buses, connectors, disk drives, structures and controllers, isolation and redundancy methods apply to power distribution in storage systems so that there is no single point of failure that could render the system inoperable. May be done. Figure 12 shows the power distribution of a redundant multi-pass storage system. Power is supplied via connector 1202. Alternatively, one or more connectors may be employed. More than one contact pin in the connector may provide the same voltage, and path duality is obtained if one pin fails to connect or is higher than the desired resistance. The power bus A 1204 powers the local regulator 1208, and the local regulator 1212 powers one or more additional local regulators, as optionally indicated by the local regulator 1216. You may. Local regulator 1208 powers structure A 1206. The local regulator 1212 powers the structure B 1210. The optional local regulator 1216 may power the disk controller 1214. Other local regulators (not shown) may power additional disk controllers and interface controllers, discrete circuits, or other circuits such as environmental monitors. Local regulators may be employed to deliver regulated power to the desired voltage for components such as integrated circuits that consume relatively low power compared to disk drives. Systems, disk controllers and structures with redundant interface controllers can employ redundant components to access data, so local regulators can also be employed for each component and stand alone. Continuous system operation is obtained in the event of a regulator failure. In addition, the connector 1202 in FIG. 12 is one or more connected to the power bus B 1218. Has a pin. The power bus B 1218 powers the voltage regulators 1220 and 1222. The regulators 1220 and 1222 may be connected so that they can be powered by either regulator and include a disconnect circuit such as a diode or other component. Alternatively, regulators 1220 and 1222 may include input signals that can enable or disable each regulator. The regulator may be controlled by writable registers, I2C buses, or other signal lines. The voltage regulators 1220 and 1222 provide regulated power to one or more additional controls represented by control 1224, control 1228 and optionally control 1232. Control 1224 controls power for disk group 1226. Control 1228 controls power for disk group 1230. Control 1232 controls power for disk group 1234. Additional control units (not shown) may power control additional disk groups or other components such as environmental monitors and fans. Controls 1224, 1228, 1232 and other controls selectively power switches, fuses, circuit breakers, transistors (field effect transistors, SCRs (Silicon Control Transistors) or disk groups or other components). It may be equipped with any other device adopted). The control may include detection of current and voltage or either, and may operate in response to an automatic method or control signal. Figure 12 wastes power in the system so that the power redundancy and disconnection method remains available after a regulator failure and cuts off power for one or more disk drives in the group that includes the failed drive. Show that it is applicable to the components of the data storage system so as to disconnect the components that do not or carry excessive power. As mentioned above In addition, data from a failed drive or disk group can be copied or rebuilt using spare capacity and saved. Accordingly, embodiments of the present invention can provide a data storage system that is free of single point failures resulting in data loss. Insert term structure refers to any form that can provide a constructable connection between a data storage device and a disk controller, and is a Fiber Channel structure, switch, cross point switch, multiplexer, A port bypass controller and other equipment shall be included. The structure may replace the switch, switching device, or port bypass controller shown in the figure.

The above figure includes switches, switching devices, port bypass switches and structures to provide a buildable connection between the data storage device and the disk controller.

The embodiments of the present invention can be conveniently adopted in a large number of disk assemblies (MDAs) that include multiple switching devices and are inserted or removed from a cabinet or other fixture as a single unit. MDA can include switching equipment, can include switching equipment and structures, can include switching equipment, structures and disk controllers, or can include data storage equipment, structures and disk controllers. It can include, or can include data storage devices, structures, disk controllers and interface controllers. In other words, the embodiments of the invention illustrated in the figure may be divided between the components placed in the MDA and the components placed in a cabinet, shelf or other fixture. Such divisions can reflect the size of the MDA, the number of connectors, the interface type, the driving force of the bus signal and other factors. In some embodiments, the MDA may employ a laterally mounted storage system, the device having at least one longest axis of the storage device body making the MDA into a cabinet, shelf or other fixture. It is mounted orthogonal to the insertion direction. These embodiments allow the connectors of storage devices, such as disk drives, to engage directly with connectors located on the backplane, resulting in intermediate connectors, cables, etc., and failures caused by intermediate connections. Eliminate potential additional points.

Computer program code running on a host system and / or one or more interface controllers, or one or more disk controllers, may be employed to build the structures of the invention. The structure may be controlled by computer program code running on one or more host computers. Such program code may include performance monitoring and load balancing functions. Structure construction, as a result of detected failures, or as a result of detected failures, such as transaction processing or video streams, or load, data format, data size, data storage format, It may be executed in response to other conditions including a desired response time and the like. One or more disk controllers may control multiple structures. The computer program code running on the disk controller may build the structure in response to a failure or other condition. The construction of multiple structures is shared between one or more host computers and one or more disk controllers. As mentioned above, the switch control may employ one or more control buses such as I2C, one or more disk buses, or both. Structures may be mapped as devices on one or more disk arrays buses, and control signals for one or more structures may be carried via disk arrays or buses. Some figures show individual switch control blocks. In some embodiments, the switch control block may be a pair of structures.

Figure 13 shows the steps performed by the system configuration computer program code running on the host and / or disk controller. The process of FIG. 13 is applicable to the system as shown in FIGS. 10 and 11. Process 1300 starts at step 1300 and performs a check to determine if an error condition exists. The error condition may include an error such as a read or write error detected by a disk drive, disk controller or host system, for example. When an error is detected by the disk drive, the error may be reported to the disk controller and checked by the disk controller and forwarded to the host system or either. When the disk controller detects an error, it may check for the error and transfer it to the host system or either. Alternatively, the error may be detected by the host system. At step 1304, tests may be run to determine if the host can communicate with interface controller A using system bus A. At step 1306, tests may be run to determine if the host can communicate with interface controller A using system bus B. At step 1308, tests may be run to determine if the host can communicate with interface controller B using system bus A. At step 1310, tests may be run to determine if the host can communicate with interface controller B using system bus A. Steps 1304 to 1310 determine whether the system bus "A" and the system bus "B" can be used to communicate with the interface controller "A" and the interface controller "B". In step 1312, step Report errors detected by 1304 to 1310 to the host or other system. At step 1314, a check is performed, for example to recheck the reported error, to determine if the host or other system can communicate with at least one interface controller. If the host or other system is unable to communicate with at least one interface controller, the process ends in step 1316. If the check performed in step 1314 determines that the host or other system can communicate with at least one interface controller, continue processing to step 1318, run the test, and press disk controller "A". Use to determine if disk controller A can be accessed. At step 1320, a test is performed to determine if the interface bus B can be used to access the disk controller A. In step 1322, a test is performed to determine if the interface bus A can be used to access the disk controller B. At step 1324, a test is run to determine if the interface bus B can be used to access the disk controller B. In step 1326, report the error detected by steps 1318-1324. At step 1328, the test results are checked to determine if at least one disk controller is accessible. If no disk controller is accessible, the process ends in step 1330. When at least one disk controller is accessible, the process goes to step 1332, runs a test, and uses the disk controller "A" to determine if the structure "A" can be accessed. At step 1334, a test is performed to determine if structure B can be used to access structure A. Ste At 1336, a test is performed to determine if structure A can be used to access structure B. At step 1338, a test is performed to determine if structure B can be used to access structure B. In step 1340, report the error detected in steps 1332 to 1338. In step 1342, the test results are checked to determine if at least one structure is accessible. If none of the structures are accessible, the process ends in step 1344. Continue processing to step 1346 when at least one structure is accessible. At step 1346, a test is performed to determine if structure "A" has access to all installed drives. Such tests may include reading and writing drive registers and / or reading data in drive media and / or reading data in drive media. If all drives are inaccessible or do not work properly, structure A is constructed to disconnect one or more drives in step 1348, and then processing continues in step 1350. At step 1346, if the test is run and it is determined that all drives are accessible and working properly, then the process continues to step 1350. At step 1350, a test is performed to determine if all drives with structure "B" are accessible. If some drives are inaccessible or do not work properly, structure "B" is constructed to detach one or more drives in step 1352, and then processing is continued in step 1354. In step 1354, data from an inaccessible or failed drive may be reconstructed or copied and stored in another device, or stored in another system for fault tolerance. .. I / O commands are remapped and functioning The interface controller, disk controller or structure may be utilized as recognized by previous tests. Next, the process ends in step 1356. If the test is performed in step 1350 and it is determined that all drives are accessible and working properly, the process is continued in step 1356. If the test is performed in step 1350 and it is determined that all the drives are accessible and working properly, the process ends in step 1356. In addition, the results of the tests performed may be employed to build, for example, the power circuit shown in FIG. 12 so that power is not supplied to the failed component. The tests performed, the order of the tests performed, the structure of the structure, the reconstruction of the data, and the remapping of the inputs and outputs include the number of disk drives installed in the port of the structure or multiple structures. May be modified according to the configuration of the storage system, including the number of host buses, interface controllers, disk controllers, number and type of structures, and number of disk drives. The form of the reported error may be used to select a test or test set. Alternatively, following the reported error, the scope of the test may be run to determine the overall conditions of the storage subsystem. There may be a hierarchical order of tests in which the actions of the various system components are performed in a predetermined order. The tests performed in FIG. 13 may be performed by the host or other system, or by components within the storage subsystem. The computer program code that performs the tests may be present in individual components of the system, or may be transferred from other systems or components. The test may include the execution of self-testing computer program code in the component. For example, a disk drive is powered on

Embodiments of the present invention may be employed to provide a large number of maintenance-free disk storage assemblies that can be installed and removed from fixtures such as storage cabinets, dividers, shelves, and the like. A large number of interface controllers, disk controllers, buses and structures allow continuous operation after a disk, disk controller, interface controller, connector or bus failure. Systems with a large number of drives may employ a third bus, as shown in Figure 5, so that system performance can be maintained high after a bus or disk controller failure. In the spirit of the present invention, various substitutions of the disclosed embodiments are possible, including the number of disk drives, the types of disk controllers, interface controllers, buses, switching devices and their controls.

In the above description, various explanations have been made by adopting a disk drive and a disk controller in order to explain an embodiment of the present invention. The embodiments of the present invention are not limited to a specific number of storage devices, and are not limited to data storage devices including the format of the storage medium and the format of the bus. The disk controller needs to be associated with any form of controller adopted to access the data from the storage device. In addition, the disk controller may provide fault-tolerant data formatting capabilities such as RAID, ECC or other formats. The data storage device is not limited to any form including an electrical, magnetic, optical or chemical data storage device including a RAM drive including a hard disk, an optical drive, a solid-state memory device, etc. It may be equipped with a data storage device and may include a combination thereof, and may further include a combination of volatile and non-volatile data storage devices. A structure or multiple structures that interconnect one or more disk controllers and one or more storage devices allow any constructable connection between multiple disk controllers and multiple switch controls. Device or a plurality of devices, and may include an interface format and a data format conversion. For example, the structure may convert series-mounted SCSI storage devices and interface signals into Fiber Channel signals that are communicated to the controller. The interface controller may provide interface formats and data format conversions, and may further execute computer program code to build one or more structures.

The above description of the present invention has been provided for purposes of illustration and description. This is not intended to be exhaustive or limited to the form itself in which the present invention is disclosed, and other modifications and modifications are possible in the light of the above techniques. The examples have been selected and described to best explain the principles of the invention and its practical applications, and are therefore considered suitable for the particular use intended by those skilled in the art. The present invention is best utilized in various embodiments and modifications. The appended claims are intended to be construed as including other alternative embodiments of the invention.

<figref num="1">The figure which shows the structure of the data storage system by a single port.</figref><figref num="2">The figure which shows the structure of the dual port type data storage system.</figref><figref num="3">The figure which shows the structure of the loop storage system.</figref><figref num="4">The figure which shows the structure of the storage system which adopted the disk drive by the interchangeable single port.</figref><figref num="5">The figure which shows the structure of the storage system which adopted the disk drive by the interchangeable single port.</figref><figref num="6">The figure which shows the Example of the loop bypass storage system.</figref><figref num="7">Diagram showing a loop bypass storage system with two drives connected to each bypass controller port.</figref><figref num="8">The figure which shows the loop bypass storage system which has two dual port devices connected to each port.</figref><figref num="9">Diagram showing a loop bypass storage system with two dual port devices connected to a port bypass controller.</figref><figref num="10">The figure which shows the redundant storage system of a large number of paths.</figref><figref num="11">The figure which shows the redundant storage system of other many paths.</figref><figref num="12">The figure which shows the power distribution of the redundant storage system of a large number of paths.</figref><figref num="13">A system configuration diagram running on a host and / or disk controller showing the steps performed by computer program code.</figref>

Code description

1002 Host 1004 Host Bus A 1006 Host Bus B 1008 Drive Array Controller A 1010 Drive Array Controller B 1012 Structure Bus A 1014 Structure Bus B 1016, 1018 Structure 1020 , 1022, 1024, 1026, 1028 drives