High data rates require an exceedingly rapid switching of the magnetization direction of the storage medium by reversing the current circulating in the write head. One approach to alleviate the problem is to restrict the length of alternating sequences of 1’s and 0’s by using mapping techniques to translate data into coded sequences of bits where 1’s change to 0’s and back again less frequently. The fundamental challenge of this constrained-coding process is to design high-rate codes such that the performance degradation due to rate loss is minimized.

Therefore we design high-rate modulation codes for tape storage that satisfy the required constraints, have very low error propagation and minimize the modulation code overhead over previously used modulation codes. Furthermore, the code design needs to ensure that special patterns that are added in the tape format to allow separation of data sets and other functions do not appear in modulation encoded data.

Another research area is the construction of a new class of maximum-transition-run (MTR) block codes satisfying predetermined (j, k) constraints [2009-5],[2009-6], which is based on a novel enumerative encoding scheme. These MTR codes, which are a generalization of Fibonacci codes, are optimal in the following sense. For a given code word length of N bits, the new class of codes contains the maximum number of length-N codewords satisfying the (j, k) constraint.

Furthermore, by using a slight modification of the code construction, one obtains practical very-high-rate modulation codes that have a limited error propagation and a redundancy of less than 1%.

Ask the expert

Mark A. Lantz

Mark A. Lantz

IBM Research scientist