Showing papers by "Evangelos Eleftheriou published in 2008"
TL;DR: This is the first time a scanning-probe recording technology has reached this level of technical maturity, demonstrating the joint operation of all building blocks of a storage device.
Abstract: Ultrahigh storage densities can be achieved by using a thermomechanical scanning-probe-based data-storage approach to write, read back, and erase data in very thin polymer films. High data rates are achieved by parallel operation of large two-dimensional arrays of cantilevers that can be batch fabricated by silicon-surface micromachining techniques. The very high precision required to navigate the storage medium relative to the array of probes is achieved by microelectromechanical system (MEMS)- based x and y actuators. The ultrahigh storage densities offered by probe-storage devices pose a significant challenge in terms of both control design for nanoscale positioning and read-channel design for reliable signal detection. Moreover, the high parallelism necessitates new dataflow architectures to ensure high performance and reliability of the system. In this paper, we present a small-scale prototype system of a storage device that we built based on scanning-probe technology. Experimental results of multiple sectors, recorded using multiple levers at 840 Gb/in2 and read back without errors, demonstrate the functionality of the prototype system. This is the first time a scanning-probe recording technology has reached this level of technical maturity, demonstrating the joint operation of all building blocks of a storage device.
139 citations
TL;DR: An efficient intradisk redundancy scheme based on an interleaved parity-check coding scheme provides the same reliability as the optimum, albeit more complex, Reed--Solomon coding scheme, which incurs only negligible I/O performance degradation.
Abstract: Today's data storage systems are increasingly adopting low-cost disk drives that have higher capacity but lower reliability, leading to more frequent rebuilds and to a higher risk of unrecoverable media errors. We propose an efficient intradisk redundancy scheme to enhance the reliability of RAID systems. This scheme introduces an additional level of redundancy inside each disk, on top of the RAID redundancy across multiple disks. The RAID parity provides protection against disk failures, whereas the proposed scheme aims to protect against media-related unrecoverable errors. In particular, we consider an intradisk redundancy architecture that is based on an interleaved parity-check coding scheme, which incurs only negligible I/O performance degradation. A comparison between this coding scheme and schemes based on traditional Reed--Solomon codes and single-parity-check codes is conducted by analytical means. A new model is developed to capture the effect of correlated unrecoverable sector errors. The probability of an unrecoverable failure associated with these schemes is derived for the new correlated model, as well as for the simpler independent error model. We also derive closed-form expressions for the mean time to data loss of RAID-5 and RAID-6 systems in the presence of unrecoverable errors and disk failures. We then combine these results to characterize the reliability of RAID systems that incorporate the intradisk redundancy scheme. Our results show that in the practical case of correlated errors, the interleaved parity-check scheme provides the same reliability as the optimum, albeit more complex, Reed--Solomon coding scheme. Finally, the I/O and throughput performances are evaluated by means of analysis and event-driven simulation.
117 citations
IBM1
TL;DR: The results demonstrate that the reliability improvement due to disk scrubbing depends on the scrubbing frequency and the workload of the system, and may not reach the reliability level achieved by a simple IPC-based intra-disk redundancy scheme, which is insensitive to the workload.
Abstract: Two schemes proposed to cope with unrecoverable or latent media errors and enhance the reliability of RAID systems are examined. The first scheme is the established, widely used disk scrubbing scheme, which operates by periodically accessing disk drives to detect media-related unrecoverable errors. These errors are subsequently corrected by rebuilding the sectors affected. The second scheme is the recently proposed intradisk redundancy scheme which uses a further level of redundancy inside each disk, in addition to the RAID redundancy across multiple disks. Analytic results are obtained assuming Poisson arrivals of random I/O requests. Our results demonstrate that the reliability improvement due to disk scrubbing depends on the scrubbing frequency and the workload of the system, and may not reach the reliability level achieved by a simple IPC-based intra-disk redundancy scheme, which is insensitive to the workload. In fact, the IPC-based intra-disk redundancy scheme achieves essentially the same reliability as that of a system operating without unrecoverable sector errors. For heavy workloads, the reliability achieved by the scrubbing scheme can be orders of magnitude less than that of the intra-disk redundancy scheme.
69 citations
IBM1
TL;DR: It is argued that controlling the head-tape interaction is key to achieving high linear density, whereas track-following and reel-to-reel servomechanisms as well as transverse dimensional stability are key for achieving high track density.
Abstract: We examine the issue of scaling magnetic tape-recording to higher areal densities, focusing on the challenges of achieving 100 Gb/in2 in the linear tape format. The current highest achieved areal density demonstrations of 6.7 Gb/in2 in the linear tape and 23.0 Gb/in2 in the helical scan format provide a reference for this assessment. We argue that controlling the head-tape interaction is key to achieving high linear density, whereas track-following and reel-to-reel servomechanisms as well as transverse dimensional stability are key for achieving high track density. We envision that advancements in media, data-detection techniques, reel-to-reel control, and lateral motion control will enable much higher areal densities. An achievable goal is a linear density of 800 Kb/in and a track pitch of 0.2 µm, resulting in an areal density of 100 Gb/in2.
48 citations
Journal Article•
36 citations
Patent•
IBM1
TL;DR: In this article, a method for reducing data loss comprising a first computing step for computing an intermediate result for each redundancy information entity of a redundancy set by processing respectively associated data information entities of a given data set on at least two main diagonals of a parity check matrix representing an error correction coding scheme.
Abstract: The present invention relates to a method for reducing data loss comprising a first computing step for computing an intermediate result for each redundancy information entity of a redundancy set by processing respectively associated data information entities of a given data set on at least two main diagonals of a parity check matrix representing an error correction coding scheme. The method further comprises a second computing step for computing the information content of the respective redundancy information entity dependent on the respective intermediate result.
33 citations
Patent•
IBM1
TL;DR: In this article, a servo reader is configured to read servo bursts from a data tape, and an anti-aliasing filter is used to output a bandlimited signal.
Abstract: Provided are techniques for synchronous servo channel for a data tape drive. A servo reader is configured to read servo bursts from a data tape. An anti-aliasing filter is operable to output a bandlimited signal and is coupled to receive a servo channel signal comprising servo bursts from the servo reader. An analog-to-digital converter (ADC) is operable to output signal samples and is coupled to receive the bandlimited signal from the anti-aliasing filter. An interpolation and control unit is operable to output interpolated signal samples and control signals, has a first input coupled to receive the signal samples from the ADC, and has a second input coupled to receive the correlation signal samples from the dibit correlator. A dibit correlator is operable to output correlation signal samples and is coupled to receive the interpolated signal samples and control signals from the interpolation and control unit.
16 citations
TL;DR: In this article, the authors proposed a control architecture based on resonant controllers where the impact of measurement noise on positioning is minimal while providing sufficient damping and hence satisfactory tracking performance.
12 citations
Patent•
IBM1
TL;DR: In this paper, a digital signal sample having a data-dependent noise component is received and a gain value, stored in a location in a gain table, is selected in a datadependent manner.
Abstract: Method, apparatus and computer program product adjust gain in a read channel of a magnetic media data storage device. A digital signal sample having a data-dependent noise component is received. A gain value, stored in a location in a gain table, is selected in a data-dependent manner. The gain of the signal sample is adjusted in response to the selected gain value. A bit pattern is detected from the gain-adjusted signal sample and a data output signal is output based upon the detected bit pattern.
11 citations
IBM1
TL;DR: A novel RC architecture for product codes is presented and illustrated by an example based on the linear tape open standard, generation 4 (LTO-4), which achieves 5.7% improvement in rate while maintaining the same interleaved I = 11 modulation constraint, but at the cost of a slight weakening of the G-constraint.
Abstract: Reverse concatenation (RC) architectures, which recently have been deployed in hard-disk-drive (HDD) products, offer crucial advantages in coding such as (i) avoiding error propagation through the modulation decoder, (ii) allowing the use of efficient high-rate modulation codes, and (iii) passing of soft information from the detector to the decoder, which facilitates parity-post processing and iterative coding schemes. In HDDs, error-correcting codes essentially consist of a single high-rate Reed-Solomon code, whereas in tape recording, large product codes are used that require a new RC architecture. Such a novel RC architecture for product codes is presented and illustrated by an example based on the linear tape open standard, generation 4 (LTO-4). Compared with the rate-16/17 modulation code of the LTO-4 standard, the proposed RC scheme has a modulation scheme of rate 0.9951, i.e., achieves 5.7% improvement in rate while maintaining the same interleaved I = 11 modulation constraint, but at the cost of a slight weakening of the G-constraint.
10 citations
IBM1
TL;DR: Significant improvements in both of the read-lifetime and the rate of signal amplitude-loss were achieved using the intermittent-contact read method compared with the contact read method because of substantial reduction in tip wear rate.
Abstract: Ultrahigh data storage densities on the order of 1 Tb/in2 or higher can be achieved by using scanning-probe microscopy techniques to write, read back and erase data on very thin polymer films. The written information is usually read back in contact mode because it is simple to implement and analyze. However, the physical contact between the probe-tip and the media leads to wear of the tip due to abrasive and adhesive forces. Tip wear has adverse effects on the read and write performance at ultrahigh data storage densities. In this paper, an intermittent-contact mode read operation is presented that has been developed to improve the durability of probe-based devices by significantly reducing tip wear. The electrostatic pull-in force between the cantilever and the silicon substrate underneath the polymer medium is used as the actuating force for intermittent-contact operation. The cantilever response is highly nonlinear because of the large adhesion between the tip and the soft polymer medium, which forces the cantilever to remain in contact with the medium for significant fraction of its periodic orbit. For fast and noninvasive reading a feedback controller is designed to ensure reliable small-amplitude operation. Significant improvements in both of the read-lifetime and the rate of signal amplitude-loss were achieved using the intermittent-contact read method compared with the contact read method because of substantial reduction in tip wear rate.
Patent•
IBM1
TL;DR: In this article, a longitudinal position (LPOS) word L(n) is encoded with error correction capability, which includes a plurality of LPOS symbols L0 through Lk(n), calculated as (Formula I) and representing the longitudinal position of a magnetic tape relative to a tape head in a tape storage system.
Abstract: A longitudinal position (LPOS) word L(n) is encoded with error correction capability. The LPOS word includes a plurality of LPOS symbols L0(n) through Lk(n) calculated as (Formula I) and representing a longitudinal position of a magnetic tape relative to a tape head in a tape storage system. A word type is determined in response to at least one of the plurality of LPOS symbols and, in response to the determined word type, at least one formatted symbol F(n) is generated from the plurality of LPOS symbols L(n). At least one parity symbol P(n) is generated from the formatted symbol F(n). The formatted and parity symbols are communicated to a servo channel of the tape storage system to be recorded onto a servo track of the magnetic tape. Adding redundancy to LPOS words provides the capability of correcting multiple bit errors without increasing the LPOS word length.
Patent•
IBM1
TL;DR: In this article, an intermediate bit sequence of the identified most likely path memory state is obtained, the intermediate bits extending from an initiation point of the path memory which is intermediate the output and the input of the PRML Viterbi detector.
Abstract: Magnetic tape read channel signal values are developed employing intermediate bits of the path memory of a PRML Viterbi detector. Identification logic identifies a most likely path memory state of the PRML Viterbi detector from the path metrics of the PRML Viterbi detector. An intermediate bit sequence of the identified most likely path memory state is obtained, the intermediate bit sequence extending from an initiation point of the path memory which is intermediate the output and the input of the PRML Viterbi detector. A sample value is determined which corresponds to the obtained intermediate bit sequence.
Patent•
IBM1
TL;DR: In this article, a servo pattern, including stripes arranged in servo bursts for use in position error signal (PES) generation, is provided in which a stripe width is narrower than 1.7 μm and in which the stripes are oriented at an azimuth angle which in absolute value is equal to or larger than 6 degrees.
Abstract: A servo pattern, including stripes arranged in servo bursts for use in position error signal (PES) generation, is provided in which a stripe width is narrower than 1.7 μm and in which the stripes are oriented at an azimuth angle which in absolute value is equal to or larger than 6 degrees.
Patent•
IBM1
TL;DR: In this article, a method for controlling the movement of a positioner of a micro-scanner is described. But this method is limited to the case of a single positioner and requires the positioner's vibration resonance frequency ranges to be determined.
Abstract: The invention relates to a method of controlling movements of a positioner of a micro-scanner, the method comprising: determining the vibration resonance frequency ranges of the positioner, and performing a main scan by a controlled movement of the positioner.
Patent•
19 Jun 2008TL;DR: In this paper, a method and apparatus for providing error correction capability to longitudinal position data are disclosed, where data are encoded via a set of even and odd LPOS words and the encoded data are then decoded by generating a subset of syndrome bits for each LPOS word.
Abstract: A method and apparatus for providing error correction capability to longitudinal position data are disclosed. Initially, data are encoded via a set of even LPOS words and a set of odd LPOS words. The encoded data are then decoded by generating a set of syndrome bits for each of the LPOS words. A determination is then made as to whether or not there is an error within one of the LPOS words based on its corresponding syndrome bits.
IBM1
TL;DR: In this paper, the dynamics of an axially-moving flexible medium are examined in the context of an application where the medium is partially supported by a frictional surface, that actively orients itself relative to the direction of transport.
Abstract: The dynamics of an axially-moving flexible medium are examined in the context of an application where the medium is partially supported by a frictional surface, that actively-orients itself relative to the direction of transport. The stability and motion of the medium are of interest in a magnetic tape data storage application where the orientation of a sensing surface is continuously altered in order to ‘follow’ the medium’s motion. Moving media that are in contact with such guiding surfaces experience friction excitations induced by the relative motion in addition to what is observed with a stationary guiding surface. Friction-induced bending moments, as well as tension fluctuation beyond the permissible limits for the flexible material can erode the potential benefits of such active positioning. This paper describes some of these dynamic phenomena using the simplified example of a planar guiding surface whose orientation is dynamically altered relative to the moving medium. A physical model for the friction-induced excitation of the moving medium is developed, and the dynamics are analyzed for their effect on critical design parameters such as the achievable bandwidth of the active control algorithm, as well as with respect to constraints on the geometry and positioning of the guiding surface.Copyright © 2008 by ASME
Patent•
IBM1
TL;DR: In this paper, a method for receiving a DSS sequence and a readback sequence, which is a function of a channel processing of the sequence by a read channel, is presented.
Abstract: Provided is a method for receiving a DSS sequence and a DSS readback sequence, which is a function of a channel processing of the DSS sequence by a read channel. A coefficient cyclic equalizer vector is generated as a function of the DSS sequence and the DSS readback sequence. An error signal is generated as a function of a comparison of the DSS sequence and an equalization of the DSS readback sequence based on the coefficient cyclic equalizer vector. An unacceptable error signal indicates a need to adjust the coefficient cyclic equalizer vector to yield an acceptable comparison of the DSS sequence and an equalization of the DSS readback sequence based on the coefficient cyclic equalizer vector.
Patent•
14 Feb 2008
TL;DR: In this paper, the authors proposed a fully synchronous servo channel for a data tape drive, which includes the initial acquisition of synchronous channel parameters, generation of a timing basis for signal interpolation, generating of a tape velocity estimate and a y-position estimate, and optimum detection of longitudinal position symbols.
Abstract: PROBLEM TO BE SOLVED: To provide a fully synchronous servo channel for a data tape drive. SOLUTION: The fully synchronous servo channel for a data tape drive includes the initial acquisition of synchronous servo channel parameters, generation of a timing basis for signal interpolation, generation of a tape velocity estimate and a y-position estimate, and optimum detection of longitudinal position (LPOS) symbols. COPYRIGHT: (C)2008,JPO&INPIT
Patent•
12 Jun 2008
TL;DR: In this paper, a probe-based data storage device is proposed, where information is stored in the form of grooves of variable length separated by lands of variable lengths by applying a series of write signals at respective probe-positions on the storage surface.
Abstract: PROBLEM TO BE SOLVED: To provide methods and apparatus for recording/reproducing data in a device in which application of a write signal causes formation of an indentation on a storage surface by a probe of the probe-based data storage device SOLUTION: Information is stored in the form of grooves of variable length separated by lands of variable length By applying a series of write signals at respective probe-positions on the storage surface, a first value (generally "1") which is a series of n>1 successive bits sequence in a recording signal is recoded These probe-positions are spaced at w COPYRIGHT: (C)2008,JPO&INPIT
IBM1
TL;DR: In this paper, a track-follow control of a MEMS-based micro-scanner for a high-density probe-based storage device is presented. But the authors focus on the positioning requirements and controllers designed for tracking performance as well as controllers using sensor fusion.