Method and system for mitigating adjacent track erasure in hard disk drives

Abstract

A method, system, and computer program product for mitigating adjacent track erasures in hard disks, includes: determining input / output (I / O) characteristics for a plurality of blocks on a hard disk; assigning the plurality of blocks to a plurality of categories of I / O characteristics by the processor; and clustering content of the blocks assigned to the same category in one or more continuous tracks on the hard disk. Each block is assigned to one category. Blocks with similar I / O characteristics are clustered on one or more continuous tracks. By performing this clustering, blocks with a high number of I / O operations are grouped and stored on fewer tracks than if they were scattered across numerous tracks. This reduces the number of tracks experiencing a high number of I / O operations, and in turn, the amount of refreshing of adjacent tracks is reduced.

Description

BACKGROUND[0001]Secure deletion of data on non-volatile magnetic data storage, such as a hard disk, is an act of securely purging the data such that there are no residuals of the data on the hard disk. Regulations exist which mandate the need for secure deletion according to various standards, such as that published by the U.S. Department of Defense. A hard disk drive includes one or more hard disks clamped to a rotating spindle and at least one head for reading or writing data on the disk. Storage on a hard disk is divided into “blocks”, which are in turn grouped into “tracks”. A disk head uses magnetism to read or write data onto the tracks. Secure deletion involves the overwriting of blocks on a track by the disk head numerous times, depending on the standard being applied. Thus, secure deletion involves multiple input / output (I / O) operations on the same blocks of a track.[0002]As a track on the hard disk is written, adjacent tracks may be overwritten by the magnetic field genera...

AI Technical Summary

Benefits of technology

[0004]According to one embodiment of the present invention, a method for mitigating adjacent track erasures in hard disks, the computer operationally coupled to a processor, the method comprising: determining input / output (I / O) characteristics for a plurality of blocks on a hard disk by the processor; assigning the plurality of blocks to a plurality of categories of I / O characteristics by the processor; and clustering content of the blocks assigned to the same category in one or more continuous tracks on the hard disk by the processor.

[0005]In one aspect of the present invention, the determining the I / O characteristics for the plurality of blocks on the hard disk by the processor comprises determining secure deletion requirements for the plurality of blocks on the hard disk by the processor, the secure deletion requirements comprising overwriting rates for the plurality of blocks; the assigning the plurality of blocks to the plurality of categories of I / O characteristics by the processor comprises assigning the plurality of blocks to a plurality of categories associated with the overwriting rates; and the clustering the content of the blocks assigned to the same category in the one or more continuous tracks on the hard disk by the processor comprises clustering the content of the blocks assigned to the same overwriting rate category in the one or more continuous tracks on the hard disk by the processor.

[0006]In one aspect of the present invention, the determining the I / O characteristics for the plurality of blocks on the hard disk by the processor comprises determining expected access frequencies for a given time period for the plurality of blocks on the hard disk by the processor; the assigning the plurality of blocks to the plurality of categories of I / O characteristics by the processor comprises: assigning the plurality of blocks to a plurality of categories of expected access frequencies for the given time period by the processor; and the clustering the content of the blocks assigned to the same category in the one or more continuous tracks on the hard disk by the processor comprises: clustering the content of the blocks assigned to the same category of expected access frequency in the one or more continuous tracks on the hard disk by the processor.

[0007]In one aspect of the present invention, the determining the I / O characteristics for the plurality of blocks on the hard disk by the processor comprises assigning a plurality of counters for the plurality of blocks of the hard disk by the processor, and for each of the plurality of blocks, incrementing by the processor the counter for the block each time a read operation or a write operation is performed on the block; the assigning the plurality of blocks to the plurality of categories of I / O characteristics by the processor comprises assigning one of the plurality of categories to each of the plurality of blocks by the processor, wherein the plurality of categories are defined based on thresholds of the plurality of counter values; the clustering the content of the blocks assigned to the same category in the one or more continuous tracks on the hard disk by the processor comprises clustering the content of the blocks assigned the same counter value category in the one or more continuous tracks on the hard disk by the processor.

[0008]In one aspect of the present invention, the determining the I / O characteristics for the plurality of blocks on the hard disk by the processor comprises maintaining a table comprising a number of accesses and remanence states of the plurality of blocks on the hard disk by the processor; the assigning the plurality of blocks to the plurality of categories of I / O characteristics by the processor comprises defining a space by the processor representing the plurality of blocks with attribute dimensions comprising the number of accesses, time, and the remanence states, placing objects representing the plurality of blocks within the space by the processor, and applying a K-means clustering algorithm on the objects by the processor using a plurality of categories of the attribute dimensions as classes; the clustering the content of the blocks assigned to the same category in the one or more continuous tracks on the hard disk by the processor comprises: storing the content of the plurality of blocks in each cluster in the space in the one or more continuous tracks on the hard disk by the processor.

Problems solved by technology

As a track on the hard disk is written, adjacent tracks may be overwritten by the magnetic field generated in the disk head, and data recorded in the adjacent tracks may be erased or corrupted as a result.

However, the I / O operations required in the refreshing of the adjacent tracks impose a performance penalty on the hard disk drive.

With the aggravation of the ATE and ATI phenomenon due to secure deletion requirements, the need to refresh adjacent tracks become more frequent, imposing an even high performance penalty on the hard disk drive.

Images