Providing enhanced tolerance of data loss in a disk array system

A device adapter for controlling access to an array of hard disk devices (HDDs) organized in a two-dimensional structure of rows and columns of HDDs, the device adapter comprising:

• access control logic to manage read requests and write requests to the HDDs in the array of HDDs, the access control logic mapping a full stride of data elements and checksum elements to the array of HDDs, wherein each element in the full stride resides on a physically independent HDD in the array of HDDs;
• checksum calculation logic to calculate a row checksum element and a column checksum element of the data elements in the full stride;
• HDD failure detection logic to identify a failed HDD in the array of HDDs;
• and
• correction logic utilizing at least one of the row checksum element and the column checksum element to correct an element of a full stride containing a failed HDD in response to the HDD failure detection logic identifying the failed HDD.

A method for providing enhanced tolerance to data loss on three hard disk devices (HDDs) in an array of HDDs, the method comprising:

• organizing the array of HDDs in a two-dimensional structure of rows and columns of HDDs;
• partitioning the rows and columns of HDDs into rows and columns of elements, each element having an element type, wherein a first element type is a data element and a second element type is a checksum element;
• creating a full stride from the rows and columns of elements, including an element from each HDD in the rows and columns of HDDs, each element in the full stride residing on a physically independent HDD;
• calculating a row checksum element as a checksum of all data elements in a row of the full stride;
• calculating a column checksum element as a checksum of all data elements in a column of the full stride;
• wherein the row checksum element calculation and the column checksum element calculation intersect at a single data element;
• and
• using the row checksum element and the column checksum element to correct one or more HDD failures in the array of HDDs.

A redundant array of independent disks (RAID) storage system comprising:

• more than one array of hard disk devices (HDDs), each array of HDDs organized in a two-dimensional structure of rows and columns of HDDs;
• each array of HDDs including more than one full stride, each full stride logically partitioning the rows and columns of HDDs into rows and columns of elements, each full stride including an element from each HDD in the rows and columns of HDDs, each element in each full stride residing on a physically independent HDD;
• each element having an element type, wherein a first element type is a data element and a second element type is a checksum element;
• row checksum elements calculated as a checksum of all data elements in a row of each full stride;
• column checksum elements calculated as a checksum of all data elements in a column of each full stride;
• and
• more than one device adapter for controlling access to the more than one array of HDDs, each device adapter calculating the row checksum elements and the column checksum elements, each device adapter operating in tandem providing redundant access control for the more than one array of HDDs.

A computer program product for providing enhanced tolerance to data loss in an array of HDDs, the computer program product comprising:

• a storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for facilitating a method comprising:
• organizing the array of HDDs in a two-dimensional structure of rows and columns of HDDs;
• partitioning the rows and columns of HDDs into rows and columns of elements, each element having an element type, wherein a first element type is a data element and a second element type is a checksum element;
• creating a full stride from the rows and columns of elements, including an element from each HDD in the rows and columns of HDDs, each element in the full stride residing on a physically independent HDD;
• calculating a row checksum element as a checksum of all data elements in a row of the full stride;
• calculating a column checksum element as a checksum of all data elements in a column of the full stride;
• wherein the row checksum element calculation and the column checksum element calculation intersect at a single data element;
• and
• using the row checksum element and the column checksum element to correct one or more HDD failures in the array of HDDs.