SBIR Phase I: Data Recovery from Low-Cost High-Capacity Backup/Archival Storage Systems for Small Enterprises

This Small Business Innovation Research (SBIR) Phase I project will investigate a novel approach for recovering mission-critical hard disk data from disk-based archival/backup systems that are unreadable due to hardware failures of the drive systems (such as a head crash or physical impact). Under these circumstances, read channel techniques like Partial Response Maximum Likelihood (PRML) often cannot correctly detect the data because prior knowledge of the write channel may not be available at the time of recovery. As such, this SBIR research is a new approach of data recovery using a drive-independent read channel technique that is insensitive to the specifications of the write channel. This technique is based on the "response function" characterization of giant-magnetoresistive heads of hard drives. It has the promise of effectively removing intersymbol interference (ISI) that in turn (after the appropriate decoding) will lead to the recovery of the otherwise unreadable backup/archival data. The anticipated result of this research will be a new data recovery technique and related software that will remove ISI without prior knowledge of how the data were written. In addition, this research may also lead to new designs of hard disk read channels that will significantly increase the data storage density of future backup/archival systems. It fills an existing void of recovering hard disk data from disk-based backup/archival systems that are unreadable by current data retrieval techniques. The commercial benefits of this research are promising due to the growing need for computer data recovery. From backup/archival databases in small enterprises to large mission-critical servers in corporate enterprises and federal agencies, this recovery technique potentially could save America hundreds of millions of dollars per year in data-loss-related costs. This research also has a broader societal impact in the intelligence, security, and law enforcement communities where, in the field of computer forensics, for example, the retrieved data may serve as digital evidence for criminal and terrorist prosecution, thus promoting homeland security. Furthermore, the successful implementation of this ISI-removal technique may substantially increase the capacity and reliability of future backup/archival hard drives. In addition, this technique also consumes less power and assumes a lower cost of implementation, making it suitable for mobile applications and a viable contender for read channel products in the multi-billion dollar hard disk market.