Abstract: A data storage system architecture for accessing data, having a plurality of data storage subsystems coupling to a host for receiving an I/O request from the host, is provided. Each of the plurality of the data storage subsystems has: a redundant array of independent disks layer (RAID layer) for generating a plurality of media extents; a virtualization module for generating one or more virtual volumes (VVs); an on-board load balance (OBLB) module for offering a data access interface between the host and the data storage system architecture to parse the I/O request into at least one sub-I/O request; and a media extent server (ME server) module to receive the at least one sub-I/O request and to determine the validity of the at least one sub-I/O request. Accordingly, the host is able to issue I/O requests to any one of the data storage subsystems.

Abstract: A data storage system architecture for accessing data, having a plurality of data storage subsystems coupling to a host for receiving an I/O request from the host, is provided. Each of the plurality of the data storage subsystems has: a redundant array of independent disks layer (RAID layer) for generating a plurality of media extents; a virtualization module for generating one or more virtual volumes (VVs); an on-board load balance (OBLB) module for offering a data access interface between the host and the data storage system architecture to parse the I/O request into at least one sub-I/O request; and a media extent server (ME server) module to receive the at least one sub-I/O request and to determine the validity of the at least one sub-I/O request. Accordingly, the host is able to issue I/O requests to any one of the data storage subsystems.

Abstract: A method for redundant array of independent disks (RAID) consistency initialization. The method allows fast RAID creation of the RAID subsystem and the RAID is capable of accepting I/O access after completion of the RAID creation. The method comprises the steps of: setting a RAID configuration, creating an initialization progress table (IPT) for storing the progress states of the initialization of the RAID, and writing the IPT and the RAID configuration into a non-volatile memory device. After that, the RAID can accept I/O access and meantime a consistency initialization can be performed according to the IPT. In one example, the consistency initialization comprises a consecutive initialization mechanism consecutively progressing in background mode and an induced consistency initialization mechanism induced by an I/O accessing to the RAID.

Abstract: Data access methods and storage subsystems thereof for reading data from storage devices in a redundant array of independent disks (RAID) system are provided. After a controller receives a read request, the target data that the controller is about to read and the sub-stripe(s) where the target data is located are determined according to the logical address block information in the read request. The controller simultaneously issues a plurality of I/O requests to all storage devices in order to read the target data and the related data distributed on the same sub-stripe(s) at the same time. If there is any target data without responses for a certain time, it is able to use the related data, which is already responded to the controller, to generate the redundant data identical to the target data for responding the read request as soon as possible so as to reduce unnecessary waiting time.

Abstract: Method for updating control program of physical storage devices including the steps of: selecting PSD(s) to be updated with the control program; making the PSD(s) in an off-line state; and updating the control program of the PSD(s), wherein if it needs to issue a write request to the selected PSD(s) before completion of updating the control program, unmodified data segment(s) corresponding to the write request is marked as modified data segment(s). the selected PSD(s) performs partial rebuilding after completion of updating the control program, and if it needs to read out from or write to an unmodified data segment of the selected PSD(s) before completion of the partial rebuilding, a R/W request can be issued to the PSD directly without waiting for completion of the partial rebuilding of the PSD. Multipe parity data chunks can also be provided in the same data stripe in another embodiment.

Abstract: A structure of redundant array of independent disks (RAID) comprising multiple parity data is provided. A data protection field is attached after each basic data access unit of the parity data sequences and each subfields of the data protection field is defined according to different applications to protect the basic data access unit of the parity data or the data protection field of the payload data from errors incurring during data transmission.

Abstract: A method for committing cached data to storage media including a host entity and a physical storage device (PSD), comprising: the host entity issuing a write IO request to the PSD being in write-back caching mode to write data from a host memory to the PSD; the host entity receiving a successful IO completion response associated with the write IO request from the PSD and recording an entry of the write IO request to a selected list; the host entity issuing a synchronized cache IO request to the PSD; the PSD responding a successful IO completion response of the synchronized cache IO request to the host entity; the host entity extracting the entry of the write IO request from the selected list before issuing the synchronized cache IO request; and the host entity engaging completion processing of write IO request corresponding to the entry of the write IO request extracted.

Abstract: A structure of redundant array of independent disks (RAID) comprising multiple parity data is provided. A data protection field is attached after each basic data access unit of the parity data sequences and each subfields of the data protection field is defined according to different applications to protect the basic data access unit of the parity data or the data protection field of the payload data from errors incurring during data transmission.

Abstract: Method for updating control program of physical storage devices including the steps of: selecting PSD(s) to be updated with the control program; making the PSD(s) in an off-line state; and updating the control program of the PSD(s), wherein if it needs to issue a write request to the selected PSD(s) before completion of updating the control program, unmodified data segment(s) corresponding to the write request is marked as modified data segment(s). the selected PSD(s) performs partial rebuilding after completion of updating the control program, and if it needs to read out from or write to an unmodified data segment of the selected PSD(s) before completion of the partial rebuilding, a R/W request can be issued to the PSD directly without waiting for completion of the partial rebuilding of the PSD. Multiple parity data chunks can also be provided in the same data stripe in another embodiment.

Abstract: The invention relates to a method, a storage subsystem and system for improving data writing efficiency. When writing data, if a predetermined N number of data or fewer than the predetermined N number of data are yet written after waiting a specific time period, the data writing I/O request is regarded as completed. The corresponding resources are then released. The N number of yet written data and the related writing information are stored in a non-volatile memory (NVRAM), so that when there is a sudden power cut or power failure happens that interrupts the data writing, the process can be resumed after the power supply is turned back on.

Abstract: A storage system and method for handling bad storage device data therefor are described. The present invention uses a first mark and/or a second mark for the data processing of the redundant storage devices when the storage system is in the degraded mode together with a media error occurred or when in data writing, there is a reading abnormality in the storage devices so that new check data cannot be generated; and therefore, a better data protection method is provided.

Abstract: The present inveniton provides a method for improving data reading performance and a controller performing the same. After a read request is sent to a storage device, the processing time of the read request starts to be counted. If the read request does not successfully read data in a predetermined time period, redundant data identical to the read data is generated in order to achieve the action of reading data. The read request process is still ongoing while generating the redundant data. When the data is read or generated successfully by either of the read request process or the redundant data generating process, the data is returned to a request unit.

Abstract: A method for media scan operations for storage system is disclosed. The method comprises the steps of arranging a range of sections of media of PSDs to perform media scan operations; scheduling the media scan operations; selecting a section in the range; verifying media of the selected section; determining the status of selected section; if the status is not ok, responding by proceeding with the corrective action processes, otherwise responding by selecting another section in the range to proceed with the verifying step, the determining step, and this responding step, until no more sections in the range to be verified. A storage subsystem implementing the method, a computer system comprising such storage subsystem, and a storage media having machine-executable codes stored therein for performing the method are also disclosed.

Abstract: Data access methods and storage subsystems thereof for reading data from storage devices in a redundant array of independent disks (RAID) system are provided. After a controller receives a read request, the target data that the controller is about to read and the sub-stripe(s) where the target data is located are determined according to the logical address block information in the read request. The controller simultaneously issues a plurality of I/O requests to all storage devices in order to read the target data and the related data distributed on the same sub-stripe(s) at the same time. If there is any target data without responses for a certain time, it is able to use the related data, which is already responded to the controller, to generate the redundant data identical to the target data for responding the read request as soon as possible so as to reduce unnecessary waiting time.

Abstract: The invention relates to a method, a storage subsystem and system for improving data writing efficiency. When writing data, if a predetermined N number of data or fewer than the predetermined N number of data are yet written after waiting a specific time period, the data writing I/O request is regarded as completed. The corresponding resources are then released. The N number of yet written data and the related writing information are stored in a non-volatile memory (NVRAM), so that when there is a sudden power cut or power failure happens that interrupts the data writing, the process can be resumed after the power supply is turned back on.

Abstract: A storage system and method for handling bad storage device data therefor are described. The present invention uses a first mark and/or a second mark for the data processing of the redundant storage devices when the storage system is in the degraded mode together with a media error occurred or when in data writing, there is a reading abnormality in the storage devices so that new check data cannot be generated; and therefore, a better data protection method is provided.

Abstract: The present inveniton provides a method for improving data reading performance and a controller performing the same. After a read request is sent to a storage device, the processing time of the read request starts to be counted. If the read request does not successfully read data in a predetermined time period, redundant data identical to the read data is generated in order to achieve the action of reading data. The read request process is still ongoing while generating the redundant data. When the data is read or generated successfully by either of the read request process or the redundant data generating process, the data is returned to a request unit.

Abstract: A method for committing cached data to storage media including a host entity and a physical storage device (PSD), comprising: the host entity issuing a write IO request to the PSD being in write-back caching mode to write data from a host memory to the PSD; the host entity receiving a successful IO completion response associated with the write IO request from the PSD and recording an entry of the write IO request to a selected list; the host entity issuing a synchronized cache IO request to the PSD; the PSD responding a successful IO completion response of the synchronized cache IO request to the host entity; the host entity extracting the entry of the write IO request from the selected list before issuing the synchronized cache IO request; and the host entity engaging completion processing of write IO request corresponding to the entry of the write IO request extracted.

Abstract: A method for media scan operations for storage system is disclosed. The method comprises the steps of arranging a range of sections of media of PSDs to perform media scan operations; scheduling the media scan operations; selecting a section in the range; verifying media of the selected section; determining the status of selected section; if the status is not ok, responding by proceeding with the corrective action processes, otherwise responding by selecting another section in the range to proceed with the verifying step, the determining step, and this responding step, until no more sections in the range to be verified. A storage subsystem implementing the method, a computer system comprising such storage subsystem, and a storage media having machine-executable codes stored therein for performing the method are also disclosed.

Abstract: A method for redundant array of independent disks (RAID) consistency initialization. The method allows fast RAID creation of the RAID subsystem and the RAID is capable of accepting I/O access after completion of the RAID creation. The method comprises the steps of: setting a RAID configuration, creating an initialization progress table (IPT) for storing the progress states of the initialization of the RAID, and writing the IPT and the RAID configuration into a non-volatile memory device. After that, the RAID can accept I/O access and meantime a consistency initialization can be performed according to the IPT. In one example, the consistency initialization comprises a consecutive initialization mechanism consecutively progressing in background mode and an induced consistency initialization mechanism induced by an I/O accessing to the RAID.