Rewrite-efficient ECC/interleaving for multi-track recording on magnetic tape

A data storage tape device, comprising:

• a host interface through which a stream of user data symbols comprising a data set is received;
• a segmenting module operable to segment the data set into a plurality S of unencoded subdata sets, each subdata set comprising a array having K2 rows and K1 columns;
• a C1 encoder operable to generate N1-K1 C1-parity bytes for each row of a subdata set and append the C1-parity bytes to the end of the row to form an encoded C1 codeword having a length N1;
• a C2 encoder operable to generate N2-K2 C2-parity bytes for each column of the subdata set and append the C2-parity bytes to the end of the column to form an encoded C2 codeword having a length N2, whereby an encoded subdata set is generated having N2 C1 codewords;
• a partial codeword object (PCO) formatter operable to form a plurality of (S×N2) PCOs, each comprising a header and a C1 codeword, from the S encoded subdata sets;
• a PCO interleaver operable to map each PCO onto a logical data track according to information within the header of the PCO;
• a codeword object formatter operable on each logical data track, to merge adjacent PCOs into COs, each comprising at least two adjacent PCOs;
• a modulation encoder operable to encode the COs into synchronized COs;
• and
• a write channel, including a write head, operable to write T synchronized COs simultaneously to the tape, where T equals the number of concurrent active tracks on a data storage tape.

A computer program product of a computer readable medium usable with a programmable computer, the computer program product having computer-readable code embodied therein for writing data to a multi-track data tape, the computer-readable code comprising instructions for:

• receiving a stream of user data symbols, the stream comprising at least one data set;
• segmenting each data set into a plurality S of unencoded subdata sets, each subdata set comprising an array having K2 rows and K1 columns;
• for each unencoded subdata set:
• generating N1-K1 C1-parity bytes for each row of a subdata set and appending the C1-parity bytes to the end of the row to form an encoded C1 codeword having a length N1;
• and
• generating N2-K2 C2-parity bytes for each column of the subdata set and appending the C2-parity bytes to the end of the column to form an encoded C2 codeword having a length N2, whereby an encoded subdata set is generated having N2 C1 codewords;
• from the S encoded subdata sets, forming a plurality of (S×N2) partial codeword objects (PCOs), each comprising a header and a C1 codeword;
• mapping each PCO onto a logical data track according to information within the header of the PCO;
• on each logical data track, merging adjacent PCOs to form codeword objects (COs), each comprising at least two adjacent PCOs;
• modulation encoding the COs and adding synchronization patterns to obtain T synchronized COs, where T is the number of concurrent active tracks on the data tape;
• and
• writing T synchronized COs simultaneously to the data tape.

An apparatus for encoding a stream of user data symbols comprising a data set, comprising:

• a segmenting module operable to segment the data set into a plurality S of unencoded subdata sets, each subdata set comprising a array having K2 rows and K1 columns;
• a C1 encoder operable to generate N1-K1 C1-parity bytes for each row of a subdata set and append the C1-parity bytes to the end of the row to form an encoded C1 codeword having a length N1;
• a C2 encoder operable to generate N2-K2 C2-parity bytes for each column of the subdata set and append the C2-parity bytes to the end of the column to form an encoded C2 codeword having a length N2, whereby an encoded subdata set is generated having N2 C1 codewords;
• a partial codeword object interleaver operable to map each PCO onto a logical data track according to information within the header of the PCO;
• a codeword object formatter operable on each logical data track, to merge adjacent PCOs into COs, each comprising at least two adjacent PCOs;
• and
• a modulation encoder operable to encode the COs into synchronized COs.

A method for deploying computing infrastructure, comprising integrating computer readable code into a computing system, wherein the code, in combination with the computing system, is capable of performing the following:

• receiving a stream of user data symbols, the stream comprising at least one data set;
• segmenting each data set into a plurality S of unencoded subdata sets, each subdata set comprising an array having K2 rows and K1 columns;
• for each unencoded subdata set:
• generating N1-K1 C1-parity bytes for each row of a subdata set and appending the C1-parity bytes to the end of the row to form an encoded C1 codeword having a length N1;
• and
• generating N2-K2 C2-parity bytes for each column of the subdata set and appending the C2-parity bytes to the end of the column to form an encoded C2 codeword having a length N2, whereby an encoded subdata set is generated having N2 C1 codewords;
• from the S encoded subdata sets, forming a plurality of (S×N2) partial codeword objects (PCOs), each comprising a header and a C1 codeword;
• mapping each PCO onto a logical data track according to information within the header of the PCO;
• on each logical data track, merging adjacent PCOs to form codeword objects (COs), each comprising at least two adjacent PCOs;
• modulation encoding the COs and adding synchronization patterns to obtain T synchronized COs, where T is the number of concurrent active tracks on the data tape;
• and
• writing T synchronized COs simultaneously to the data tape.