Thin film media composed of Te-Te02 (TeOx) containing Pd additive was found to have a feasibility of being applied to high reliability data file optical disks. The carrier to noise ratio (C.N.R.) exceeds 56 dB at 1,200 r.p.m. disk rotation. The response time of saturation amplitude after a pulsed laser power irradiation was less than 300 nsec and is able to apply to high speed read verify systems. From acceleration environmental stress test results, no defect variation was observed. By the Arrhenius plot method, the disk life has been estimated to be more than 10 years at the storage environment 32 degree C, 80 %RH. The recording and degradation mechanism will be also discussed.

We examine the principal systems of Error Detection and Correction (EDAC) which have been recently proposed as U.S. standards for optical disks and discuss the the two principal methodologies employed: Reed-Solomon Codes and Product Codes, and describe the variations in their operating characteristics and their overhead in disk space. We then present current knowledge of the nature of defect distributions on optical media including bit error rates, the incidence and extents of clustered errors and burst errors, and the controversial aspects of correlation between these forms of error. We show that if such forms are correlated then stronger EDAC systems are needed than if they are not. We discuss the nature of defect growth over time and its likely causes, and present the differing views on the growth of burst errors including nucleation and incubation effects which are not detectable in new media. We exhibit a mathematical model of a currently proposed end-of-life defect distribution for write once media and discuss its implications in EDAC selection. We show that standardization of an EDAC system unifies the data recording process and is permissive to data interchange, but that enhancements in EDAC computation during reading can achieve higher than normal EDAC performance, though sometimes at the expense of decoding time. Finally we examine vendor estimates of disk longevity and possible means of life extension where archival recording is desired.

Multi-layered Bi/Se/Bi/Se... films were prepared by thermal evaporation. Optical writing experiments on these films showed an exceptionally high sensitivity for the following reason: the absorbed light heats the film until it begins to melt, at which point an exothermic reaction between the Bi and the Se releases more energy. The bits written on the multi-layered Bi/Se films are very clean and comparable to those of the best known optical recording media. Intrin-sic S/N, corrected for the substrate defects on these films, is the same or slightly better than that of 300A pure Te thin film on glass substrate. It should also be noted that an earlier study by AT&T Bell Laboratories showed an exceptionally good environmental stability of the Bi/Se bilayers. In fact, some Bi/Se bilayer films made more than 10 years ago still show little signs of degradation. For write° pulse energies below that required to form a hole, the write beam caused the reflectivity of the film to increase. The reflectivity increase effect could be used to write format information, while hole formation could then be used to record user data. Salient features of the optical recording characteristics of these media, as well as the material preparation technique, will be presented.

Dynamic GaAlAs diode laser marking experiments on an optical data storage medium are reported. The medium comprises thin films of a proprietary organic recording layer that are spin coated onto both glass and plastic disk substrates. These amorphous films with high substrate incident reflectivities (R - 30%) are stable to both air and moisture. Writing experiments were performed in the substrate incident mode. Readable marks were made with 2 to 15mW of laser power at pulse widths down to 50 ns. The marks are pits with well-defined rims. The dependence of laser recording quality on both disk and marking parameters will be discussed.

This paper describes a new method for measuring optical disk mechanical characteristics significant for both drive design and disk interchangeability. This method enables the simultaneous and accurate detection of disk properties using a compact measuring head. This paper also discusses a new acceleration criterion for optical disks and an appropriate clamping force for correct measurement of their mechanical characteristics.

Sumitomo's mass production plant for write once optical memory disks has been operational since the summer of 1985. The product disk, which has the sandwich structure with an air gap and which has the sensitive layer based on tellurium, exhibits equal or better read / write characteristics and reliability compared with ones from the laboratory. It will be demonstrated that the disks are particularly suited for high speed and high density data recording at an adequetly low error rate and also that the result of accellerated aging tests assures an excellent archivability.

Birefringence is generally caused by the orientation of the optical anisotropic molecular chain introduced during the injection molding of plastic substrates. It was found that the degree of birefringence of molded disk depends on the value of the difference between the polarizabilities parallel and perpendicular to the molecular chain. This difference in polarizabilities is proportional to the optical elastic constant of the polymer in the elastic state. It was determined that the negative and positive segmental polarizabilities may be fully compensated by mixing two polymers with opposite molecular optical anisotropy in an optimal ratio. Using this method an injection molded substrate was developed which demonstrated minimal birefringence.

This paper describes the status of magneto-optic product development and the media scaleup activities. Examples are cited from the 3M magneto-optic disc production facility to show that M/0 has emerged as the dominant erasable technology. We present data indicating a complete resolution of issues such as substrates and product lifetime that were associated with M/0 technology in the past. In the end we reflect on what lies ahead for M/0 technology and the developments needed for second generation products.

Thin films of alloys in the Fe-Tb-Co-Gd quaternary have received considerable attention as candidates for optical read-write storage media. Although the magnetic properties of many different compositions of films have been examined, a systematic investigation of film composition has been tedious due the effort involved with fabrication of many different composite or alloy targets. Multi-target magnetron sputtering (MTS) offers a convenient method to easily prepare a wide range of thin film compositions. In this study, thin films in the Fe-Tb-X ternary (X = C or 0) have been prepared using MTS to examine the effects of two common sputtering system impurity elements, carbon and oxygen. The magnetic properties of these films, measured using a vibrating sample magnetometer, are correlated with the microstructural characteristics determined using optical and electron microscopy techniques. In addition, the effect of room temperature aging on the magnetic properties of thin films overcoated with carbon will be described.

The effect of the groove profile on the characteristic properties of a magneto-optical (MO) disk with a quadrilayer structure (A1N/GdTbFe/A1N/A1) prepared on a glass substrate was investigated, using a micro Kerr-effect measuring apparatus. Taking a diffraction effect into account, we have calculated the dependence of signal quality on the groove profile, being in agreement with experiments. The carrier to noise ratio reaches to 51 dB at high recording density where the bit length is 1.0 μm. In order to obtain a better understanding of the reliability of the MO disk, the MO disks were exposed to various environments with respect to temperatures and humidities. Though the readout signals of recorded bits do not change after the aging tests, the recording sensitivity is reduced slightly after exposing at high temperatures. Therefore, the lifetime of the MO disk depends upon the stability of recording sensitivity and was evaluated as about 18 years using the Arrhenious equation.

This paper discusses the various components of the noise in the readback signal from an optical disk and suggests methods for measuring them. An analysis is presented of the relationship between optical efficiency of an M-0 head, light intensity at the detector(s), choice between PIN and avalanche photodetectors and the resulting noise levels. A number of measured noise spectra from grooved and ungrooved substrates and disks are presented and mathematical models are used to explain the results. In particular a simple model of an M-0 system with differential detectors is included.

Te-Ge-Sn-Au thin films were studied for phase change type rewritable disk media, in order to obtain the fast crystallization speed and the thermal stability. Films were prepared by co-evaporation method. It was found that through static record/erase measurements the threshold crystallizing pulse duration of Te60Ge4SnllAu25 film was only 1 usec at 2 mW of laser power; that is less than one tenth compared with that of Te80Ge5Sn15 film, while the threshold amorphizing laser power of them were almost the same, 6 mW at 0.2 psec of pulse duration regardless of Au concentration. Its crystallizaton temperature of about 130°C was enough to maintain the good thermal stability. Through DSC, X-ray and electron diffraction studies, the first appeared crystalline state in crystallization process, corresponding to the drastic change in optical property, showed only one phase of metastable simple cubic(SC) structure. It was concluded that the appearance of the SC structure made the crystallization speed higher. The obtained thin film was a candidate for the simultaneously erasable and recordable material.

Hermetically-sealed media are needed to suppress thermal deformation whidh occurs during the writing and reading process in erasable optical disks. The photopolymerization process ( 2P process ) is therefore examined for application to the stamping and sealing process. The replicability of grooves in the stamping process is studied by Fourier analysis position signal for the focused beam. Distortions in grooves replicated from a pregrooved stamper are seen to depend on the shape of the plastic substrate used in the 2P process. Various photopolymerizable adhesives have been tried to create hermetically-sealed media in the sealing process. Examination of media read/write performance shows that the hardness of the adhesive plays a key role in preventing thermal deformation during the write process.

One of the problems to be solved for the development of magneto-optical disc drivers is how to detect signals with a good carrier to noise ratio. The birefringence observed in the disc substrate is one of the factors which give serious effect on the ratio.1 We have studied the optical anisotropy of injection molded polycarbonate substrate and found that the optic axis representing the birefringence is almost perpendicular to the disc surface though it has been widely believed that the axis is parallel to the surface. Based upon the newly found fact, we have calculated the effect of the birefringence upon the carrier to noise ratio.

Antimony-Selenium alloy film with Bismuth has good characteristics for erasable optical disk media. The mechanism of the reversible change is phase transition between amorphous and crystalline states. It is confirmed by small-angle X-ray diffraction analysis that Se-Bi component plays an important role in this reversible change. The minimum erase time is 1 μs. The film has high transition temperature and activation energy of crystallization. It has a long life span of written spots in amorphous state. Recording film of Sb-Se-Bi alloy is formed on 20 cm 0 PMMA substrate and is covered with protective films. The disk samples are written and erased at a speed of 1800 rpm, using an optical head with two diode laser beams. Carrier-to-Noise ratio is more than 55 dB at 5 MHz, 30 KHz bandwidth. The newly developed disk realized high performance and durability for the use of an erasable video file system.

In-Se based amorphous-crystalline phase change recording films are analysed in terms of structure, optical properties, crystallization speed, and crystallization activation energy. Crystallization by a laser beam is possible by a 0.2μs pulse. Crystallization temperature(heating rate 5°/C min.) measured by using a differential scanning calorimeter is 135°C and crystallization activation energy is calculated to be 2.6eV. The life of the amorphous state is estimated6 to be longer than 10 years at 60°C. The phase change cycles can continue even after 10 cycles in stationary state experiments for test samples. In addition, possibility of single beam overwrite(rewriting without prior erasing) is verified by using a 1.6 μm-diameter round laser beam spot on a 5-inch-diameter disk rotating at 2400rpm. The laser beam power is modulated at 1.77MHz and 0.88MHz between an amorphizing power level and a crystallizing power level. The amorphizing power level is 14mW and the crystallizing power level is 7mW. The experimental results show that the overwrite the same as magnetic disk is possible by using the In-Se based recording film.

Optical disc memory systems, using a laserdiode, are one of the most promising devices for large capacity computer data storage. Recently we developed a device, suitable for high volume production, based on a 5-1/4" form factor. The basic specifications, features and development considerations of this new optical memory device are described.

We present in this paper an overview of Optotech's 5984 Optical Disk Drive. Key features such as the modulation code, the disk format, defect mapping scheme and the optical head and servo subsystem will be singled out for discussion. Description of Optotech's 5984 disk drive The Optotech 5984 optical disk drive is a write-once-read-mostly (WORM) rotating optical memory with 200 Megabyte capacity on each side of the disk. It has a 5 1/4 inch form factor that will fit into any personal computer full-height slot. The drive specification highlights are given in Table 1. A perspective view of the drive mechanical assembly is shown in Figure 1. The spindle that rotates the disk has a runout of less than 10 um. The rotational speed at 1200 revolutions per minute (rpm) is held to an accuracy of 10-3. The total angular tolerance from perfect perpendicular alignment between the rotating disk and the incident optical beam axis is held to less than 17 milliradians. The coarse seek is accomplished through a stepping motor driving the optical head with 1.3 milliseconds per step or 32 tracks per step. The analog channels including read/write, the phase lock loop and the servo loops for focus and track control are contained on one surface mount pc board while the digital circuitry that interfaces with the drive and the controller is on a separate pc board. A microprocessor 8039 is used to control the handshake and the sequence of R/W commands. A separate power board is used to provide power to the spindle and the stepping motors. In the following we will discuss some of the salient features in the drive and leave the details to three accompanying Optotech papers. These salient features are derived from a design that is driven by three major considerations. One is precise control of the one micron diameter laser spot to any desired location on the disk. The second consideration is effective management of media defects. Given the state of the art of the Te-based disk technology with an average raw defect density of approximately 10-5(compared to 10-draw error rate in high density magnetic hard disks), elaborate defect management tools are required to assure data integrity. The last consideration is, needless to say, low cost and high reliability.

This paper describes a tracking servo system developed for a small size optical disk system. In this tracking servo system, new techniques, such as two stage servo system and tracking error integration memory, are applied for improving track access time and data transfer rate. The two stage tracking servo system has a unique position detector, which outputs a lens position error signal relative to the optical head. Using this detecting signal, the objective lens and the optical head are controlled co-operatively. Therefore direct track access from the current track to the target track is achieved without lens vibration. Also, average track access time is reduced. The tracking error integration memory integrates track error signal synchronized with disk rotation. This memory output signal is equivalent to track eccentricity after several disk rotations. Therefore, a tracking servo system with this tracking error integration memory can achive highly accurate track following even under highspeed disk rotation and large track eccentricity. As a result, disk rotation speed can be upped and data transfer rate can be improved.

Certain seek techniques were developed for the Optotech 5984 drive to accommodate the use of a stepper motor coarse actuator and a hardware environment designed to minimize complexity and to achieve a low-cost end-unit goal. Of these techniques, the following will be discussed: fine seek actuation offset compensation, and velocity control in fine seek actuation. The fine seek actuation offset compensation provides necessary correction to fine seek drive values due to the effects of the limited positioning resolution of the stepper motor coarse actuator. The velocity control technique developed for fine seek actuation provides sufficient velocity control at track crossing frequencies up to 15Khz using a relatively slow, low-cost microprocessor.*

The well known push-pull method has a problem in tracking offset caused by deviation of the beam axis and disk inclination. The tracking offset degrades the tracking servo performance and the track following range is conventionally restricted to about ±50pm. To solve this problem and enlarge the track following range, we have developed a new tracking servo technology which uses two actuators interlinked by servos. This method completely eliminates the deviation of the beam axis and enables a track following range of ±400 μm.

A new servo method with eccentricity correction circuit for an optical disc equipment is developed. When eccentricity exists, the relative position between tracks and rotation center of a disc varies sinusoid according to the disc rotation. Eccentricity correction is performed by moving actuater according to the synchronized sinusoidal correction signal. Amplitude and phase of the correction signal are equalized to those of track movement by microcomputer. This eccentricity correction can reduce a 200 μm eccentricity down to 10 μm, improve tracking ability, and shorten track access time.

Design of a high performance linear head positioner fitted for an optical disk drive is reported. First, a flat and small positioner structure with a linear motor consisting of one coil and two magnetic circuits is invented. Next, a new design method to make drive force large, motor size small, and resonant frequency high is discussed by combining motor design with vibration analysis. Finally, the flat and small head positioner with 4.8 N at 1.6 A and about 6 kHz resonant frequency is developed by this design method.

Plan: 1) Actual Positioning Method 2) Sample Format Versus "Continuous" Format 3) Sample Access Methods Actual Positioning Method The small track pitch (equal to or less than 1.7 um) of digital optical disk devices makes a conventional mechanical access to a track very difficult to achieve. In fact, the eccentricity of the central hole (about 40 um) does not allow an absolute mechanical access to find a track.

The new tracking method named composite method is analyzed both theoretically and experimentally. It is found that this type of disk realizes high optical stability and accurate track following, which is not influenced by optical head characteristics. The composite method, a conbination of the push-pull and the wobbled tracking method, has the number of wobbled sectors in the order of 16-32. In this method, the wobbled tracking technique is used to suppress the low frequency offset component in the push-pull tracking error signal due to aging or soiling of the optical system. The relation between the number of sectors, track offset suppression performance and influence of missampling were simulated by servo simulator; the results were found to be quite consistent with the experimental results.

It seems almost paradoxical that beams of light can be moved and steered at very high speeds using a variety of scanning methods, yet the optical disk drives now being designed and marketed for data storage applications have comparatively long access times. Knowing that optical data storage has unrealized potential is of interest, but of more immediate concern is the recognition that poor access performance is a serious design issue. Magnetic disk drives offer average seek times in the 15-25ms range, compared to about 80-500ms (or more, for CD ROMS) for current optical drives. This performance disparity exists, in part, because the relatively massive "optical heads" in use today cannot be transported across the radius of a disk as quickly as a stack of much lighter magnetic heads. Any potential distance advantage that the optical drive might have, due to its substantially higher track density, is offset by the magnetic drive's use of a multi-disk stack. As a result, the drive must achieve similar radial accelerations during seeks if it is to have similar average access times. The inability of current optical drives to approach the access speeds of comparable magnetic drives significantly reduces the competitiveness of optical products in major segments of the very large data storage market. This shortcoming is especially disturbing when we know that opto-mechanical scanners typically operate in the 1-10ms range and that non-mechanical scanning techniques can be substantially faster than that.

Real time error checking is required in optical disk systems to improve data reliability and to increase data throughput to/from the computer. This paper describes a two-beam optical head which meets the requirement. The head is composed of two laser diodes (830 nm for write and servo operations, 780 nm for read operation), dichroic mirrors, and an achromatic objective lens. New optical servo technologies have been developed and incorporated in this optical head. One is a new focusing technology (called the Wax-wane method) which needs only two optical components. Another is a new tracking technology (called the ILT method) which enables a tracking range of ±400 μm.

A compact-size optical head for magneto-optical disk drives has been developed. The head is designed to be built into a disk drive which is of the same size as a full-height, 5.25" floppy disk drive. By using a newly developed semiconductor laser that can radiate a beam with a small divergent angle and of small eccentricity, we have attained simplification of a collimating system and ease of assembly. To obtain a high readout carrier-to-noise ratio, we have adopted the differential readout method and the high frequency superposition method. Moreover, to stabilize a focus servo system when a focused spot traverses guide grooves, we have chosen the pupil obscuration method for focus error detection. With these technologies, continuous irradiation of 8mW onto a disk surface and a carrier-to-noise ratio in excess of 50dB within 30kHz bandwidth have been achieved.

An axial gradient index singlet collimator lens for use in the compact disk system has been designed and successfully fabricated. Cost and manufacturing constraints have led to a plano-convex design with the axial gradient index profile symmetric about the plane, perpendicular to the optical axis, passing through the center of the lens. Diffraction limited performance has been obtained in the collimator singlet.

An optical pickup for a CD player, which has a simplified optical system, has been developed. It contains only two molded-plastic lenses. One is a plastic aspheric objective for focusing a spot on the disc. The other is a single lens for servo sensor optics. The objective is a finite-conjugate type, and it enables the pickup to be almost the same height(27mm) as our previous model, though the new one has a straight optical path. In this paper, we describe (1) a method to decide the objective specifications considering the relation between the aberration and the readout characteristics, (2) aberration problems using a plastic objective and the compensation method of spherical aberration, and (3) characteristics of the focusing sensor which employs the single plastic lens.

The performance of optical disk drives is dependent on the quality of the optical components used in the read/write head. In this paper we discuss the effects of wavefront aberrations on the tracking and focusing servo signals. A heterodyne grating shearing interferometer which can measure the performance of the optical components in the drive environment is then described. Experimental results related to this interferometer is also presented.

Frequency-domain optical recording based on the phenomenon of persistent spectral hole burning permits very high storage densities in excess of 109 bits/cm2 and very fast data access times through the combined use of spatial and spectral data storage. This paper reviews recent advances in the development of storage materials and describes exploratory investigations of technological issues of frequency-domain optical recording such as simultaneous high-density spatial and spectral recording and fast data readout.

In optical disk systems using push-pull tracking method, there have been problems of tracking offset errors caused by a disk skew and low signal-to-noise ratio due to a surface noise of the pregrooved area. In order to achieve higher track density recording, these problems have to be solved. We have developed a new type of on-land composite pregroove method for higher track density recording. In this method, the header data and the user data are recorded on the land area between two pregrooves on the disk. Each pregroove has 10-20 of tracking offset detection flags once per sector at each header area.

A new recording format for second generation optical disks has been developed that promises to yield capacities and data rates significantly greater than those offered by first generation systems. This format is a combination of a high linear density modulation scheme, made possible by the use of prepit servo techniques, and an offset track geometry. This geometry yields a reduction of track spacing, as compared to non-offset structures, and allows the track spacings to be further reduced as a function of increasing radius. These reductions permit drives using this CAV format to achieve the capacity improvements ascribed to CLV techniques, without CLV's attendant liabilities. Capacities in excess of 550 MB per 5 1/4" surface are projected with this format.

This discussion concerns the requirements for a format used in digital optical recording, specifically the 5-1/4" form factor. An implementation, based upon a fully sampled approach, is described in detail. Format requirements The function of a format is to provide a framework for the operation of a media-drive combination. In essence it should guide the drive hardware and firmware through the necessary steps to store and retrieve data. In the case of optical recording the following functions are required: . focussing (on the information carrier) . tracking (along the data tracks) . accessing (across the data tracks) . synchronization (to the pre- and self-written data) . modulation (of the data, i.e., encoding/decoding) . sectorization (for logical retrieval) The last function includes the features of block size, error correction and block (sector) header structure, and is determined by (external) requirements such as data transfer, allowed data error rate and expected media error rate. Sectorization will not be addressed here, because the emphasis will be focused on the hardware related servo (first four items) and modulation-functions.

Marketing and performance requirements dictated the need for a direct read during write (DRDW) capability for the 5984 optical disk drive product line. This report outlines the the mechanism and features of the DRDW technique developed at Optotech to allow authenticity verification of the bits written in real time.

The quality of substrate that exists between the optical head and the media sensitive layer is known to critically influence the drive performance. First, we report here the degradation of the servo and read signals by the parallelism of the two surfaces of the plastic substrate. It is conclusive that the peak to peak variation of plastic parallelism in the radial direction over one revolution can cause tracking instability and data cross talk. Consequently, the seek accuracy and bit timing accuracy in the optical drive are degraded. Second, we report that the variation of birefringence over one revolution results in the variation of light feedback into the laser cavity. This increases the laser noise, and thus degrades the read performance, especially when the laser output power is in the lower range.

Due to the nonlinear nature of the thermal optical recording process and the inherent difficulty in tracking and maintaining focus of a submicron-sized light spot, peak-shift values of 40 to 50% are likely to occur during the use of an optical data storage system. Wideband SNR of 30 dB or more will, therefore, be required to attain acceptable phase margin (i.e., reliable operation). Similarly, magnetic recorders working at high linear density will experience peak-shift values of 20 to 30%; wideband SNR from 23 to 29 dB will be required in those machines, depending upon whether or not error correction coding (ECC) is employed.

The effects of birefringence due to substrate and optical components on the sum and difference signals in optical memory systems are studied. A method is presented for calculation in an arbitrary optical system. Computed results are given to elucidate trends for a typical magneto--optic system employing two polrizing beamsplitters and differential detection, and comparison is made with experimental results. The effects of birefringence on the signal--to--noise ratio are discussed.

The performance of a magneto-optic readout system is easily degraded by optical retardations, or polarization-dependent phase shifts. Optical interference structures meant to enhance the magneto-optic effect can produce retardation which reduces the readout signal. Polarization-dependent optical elements (prisms, mirrors, etc.) can either exacerbate or compensate for this problem. Retardation due to substrate birefringence is an even more serious problem because it can cause relatively large intensity fluctuations at the photodetectors. Injection-molding techniques which minimize normal-incidence birefringence in polycarbonate substrates have been demonstrated; however vertical birefringence, which causes large retardation for marginal light rays, remains an unsolved problem.

The types of errors likely to occur with optical disks are randomly distributed short error bursts combined with less frequent longer error bursts. This error environment can be treated successfully with some form of Reed-Solomon EDAC code, since R-S codes provide good correction capability for long error bursts. Two forms of Reed-Solomon coding have been given serious consideration by the optical disk community: product codes and long-distance codes. Product codes have the advantage that they use short codewords, and thus operate faster and require less complexity in a VLSI implementation. Long-distance codes have the advantage that they use a smaller amount of redundancy to achieve the same error correction power as product codes, and thus leave a greater percent of the gross disk capacity for user data. The trade-offs associated with each of these methods are examined. An advanced form of Reed-Solomon encoding is discussed. This specific R-S technique reduces operating time and silicon complexity, and permits certain operations which are either impossible or very difficult with older R-S methods.

This paper describes some of the advantages of optical disk technology over magnetic disk and tape technologies for use in the harsh and rugged environments of military appli-cations. Lockheed's Model 300 Write-Once-Read-Mostly (WORM) optical disk drive environ-mental specifications are discussed as they relate to the requirements of Military Speci-fication MIL-E-5400, the General Specification for Electronic Equipment, Airborne.

An image retrieval system has been developed using CD-ROM ( Compact Disc - Read Only Memory ). The images such as documents, drawings, etc., are distributed in the CD-ROM disk and user can retrieve, see on CRT, and make hardcopies of them. The main features of the system are as follows: 1) A high data reliability which guarantee an error rate of less than 10 -12 is achieved with newly developed ECC LSI for CIRC and decoding software for Layered ECC adopted in the CD-ROM data format. 2) Using an image compression technique based on the G4 facsimile format, more than 5000 letter sized documents can be stored in a disk. 3) The system can display images on a high resolutional (200 TPI) CRT and print hardcopy by a 400 TPI laser beam printer. 4) The system can be operated by common 16 bit personal computers. The several applications of this system are also discussed.

This paper describes the development of the basics of an optical disk testing system used to test 12 inch, write once, Alcatel Thomson Gigadisk (ATG) media that are used at the Library of Congress in a pilot document storage and retrieval system. Since very little is known regarding the longevity of optical disk media and the fact that disk manufacturers are still refining processing techniques, any conclusions regarding error patterns, failure modes, or longevity may be superceded by a new "batch" of disks. Therefore, this paper focuses on the development of procedures for testing disks that can be used as the write once optical disk technology continues to advance.

This presentation is intended to provide the listener with a general overview of the optical media market. It deals with the basic questions and concerns expressed by those who are about to become involved in optical storage. Areas touched upon include the various types of optical media available, their storage capacities, how they're made, how they are used, life expectancy of media, states of various standards efforts, current and projected pricing and availability, market trends, and growth projecting for the next five years.

Many optical recording media companies characterize media properties using carrier to noise measurement methods. While this method is easily accomplished with off-the-shelf measurement equipment, it has proven to be a very unreliable method of characterizing the performance of optical recording materials. Optical recording data channel is known to have several unique features relating to nonlinearities and multiplicative noise sources which makes the use of carrier to noise ratio measurements ambigious or inaccurate. Alternative measures of making performance measurements using signal to noise ratio parameters provides substantially greater accuracy, eliminate the ambiguities and provide better insight into the operating properties of the recording material. Signal to noise ratio measurements (SNR) are made using a digitizer and software processing of the data channel output. SNR measurements are useful to both media and drive companies unlike CNR measurements which are useful only as a qualitative measure presented by media companies. The purpose of this paper is to evaluate the differences between CNR and SNR measurement and to show the substantial improvement in characterization accuracy which can be achieved by SNR measurements.

System application and integration issues drive the operational requirements of an optical disk memory. This paper describes a typical system integration problem and four possible solutions. The operational requirement is to WRITE a continuous input data stream into the optical disk memory for several days, and READ random files from the memory at the same time. This can be done at continuous WRITE rates of 24 Mbps (3 MBytes/sec). The four solutions involve different ways to use the RCA Optical Disk Jukebox (ODJ) and different configuration options that may be specified when ordering the ODJ. These options include multiple heads per turntable, multiple turntables per ODJ and I/O staging buffers. Data is given on the number of file READ accesses per minute vs. simultaneous WRITE rate. Diagrams of the ODJ configuration options are included.

Optical drives have unique characteristics that lend themselves to applications that previously were difficult or impossible to satisfy. Obvious examples are: a. Removable storage of huge volumes of data - Transportability - Jukebox use - Security b. Write once - Audit trails that cannot be erased c. Long lasting archives - Do not need to be rewritten every few years - Not vulnerable to magnetic fields - Randomly accessible This paper identifies some features and algorithms that provide high efficiency in achieving the benefits listed above. These features are used in some of the subsystems in which the OSI LD1200 is integrated. Gaining the benefits of optical disks by integrating them into an existing computer system was hampered because existing software drivers were designed for magnetic disk drives. Existing drivers assume the ability to update a file--including the directory--by rewriting in the same sector. The requirement for new and/or modified host adapters and software drivers has provided opportunities for existing adapter houses as well as for new entrepreneurs. This has led to subsystems in which different characteristics of optical drives are emphasized and has resulted in a wide variety of capabilities available for use in applications.

In the late 1990's supercomputers will have computational performance approaching one trillion floating point operations per second. These high performance systems will contain tens to hundreds of giga bytes of internal high speed memory either distributed, centralized or combined to provide the required bandwidth for efficient computation. The impact on the on-line storage and mass storage archive subsystems is estimated from actual user experience with current supercomputers. The on-line portion of the storage subsystem must satisfy the dual constraint of high capacity and very high transfer rates, while the mass storage portion may have greater capacity with a lower transfer rate requirement. The actual 1990's storage subsystem will be composed of both magnetic and optical disk and tape units. Some of the factors critical to design of the total storage subsystem are examined.

The National Space Science Data Center (NSSDC) has been investigating the use of optical disk technology in order to accommodate the growing need for an inexpensive, convenient method of receiving, storing, retrieving and disseminating data. The main focus to date has been on Write Once Read Many technology employing a 12" drive, with a data storage capacity of 1 Gigabyte per optical disk. A pilot project developed under a research contract with the University of Texas at Dallas provided the integration of an Optimem 1000 with a PDP 11 host computer. The NSSDC has developed a software and hardware system for integrating 12" optical disk drives with VAX VMS/Files-11 so that the optical drive appears as if it is a native DEC peripheral. The NSSDC is running performance tests on the drives. The NSSDC is submitting a mass buy of optical drive systems that interface with VAX/VMS to be distributed to scientific installations around the country. This will facilitate the exchange of data between scientific investigators and the NSSDC in the near future, by providing a temporary standard for VAX and MicroVAX science users. The long-term needs of the NSSDC are similar to that of potential optical disk users who will be exchanging or distributing optical disk platters. These needs are for standards at all levels, including interchangeability of blank media and different optical drives and interchangeability of media written on different host systems. At present, the only standard in WORM technology is the size of the footprint of the drive.

Mass memory systems based on erasable optical disk media are expected to play an important role in the Space Station era of the 1990's and beyond. Substantial work is already underway in both media and disk memory development. NASA has initiated investigation of spaceborne applications, and has begun definition of optical disk mass memory system controller requirements. This work is closely coupled with a multi-agency Government program to develop an erasable optical disk buffer memory1; program goals are one terabit (10 to the 12th bits) capacity, at up to 1.6 gigabits per second transfer rate. Current status of the NASA investigations is reported, including discussions of potential applications and the implications of environmental considerations, self test and reconfiguration philosophy, multi-user network interfaces, and variable (changeable) input/output data rates, all associated with the potentially unique usage in space flight. A baseline optical disk memory system architecture is included as a starting point to indicate interfaces. The role and applicability of interface standards such as SCSI and IPI are addressed. Finally, issues and technologies considered critical to space-flight optical disk usage, and requiring continued research and development, are presented.

The efficient storage and retrieval of technical information in wholly digital form has begun. The Computer Aided Logistics Support (CALS) initiative of the Office of the Secretary of Defense (OSD) has established a goal to transition from today's paper-based system to a digital system for logistics and technical information. This action will meet the projected information distribution requirements of the DoD and provide both an example and stimulus for similar action throughout our industrial society. A number of major aerospace contractors are already focusing on the efficient storage and retrieval of technical information from very large digital databases. However, for this information to be useable in the field, it must be made available in logical unit sizes and distributed in a convenient medium. This paper suggests that the optical memory card could become that medium. Using technology similar to today's writeable optical disks, optical memory cards will make it possible to conveniently distribute and use digital information. For example, the Drexon® LaserCardTM brand of optical memory card, while only the size of a standard credit card, can hold up to 2 megabytes of updateable digital information (about 800 pages of text or 200 pages of integrated text and graphics) and can be issued in on-demand quantities. Thus, optical memory cards have the potential to significantly contribute to the overall needs of both government and industry.

The RCA Optical Disk Jukeboxes (ODJ) delivered in 1984 and 1985 use unformatted, write-once 14-inch Du Pont media. This media can store 10 GBytes on one side of the disk. Unformatted media allows maximum utilization of the disk surface and makes it easy to use alternate types of media on the ODJ. Media formulation changes are more easily accommodated, allowing more freedom in media advancements. Because the media is unformatted, the amount of data which can be stored on the disk will vary, depending on the way it is used. Record sessions of less than 10-100 MBytes reduce the amount of data stored on the disk. This paper gives data on the media capacity vs. record session length and discusses how this impacts system integration issues. File access times, access rates and transfer rates are given for the first ODJ as delivered and for the most recent ODJ in our laboratory when using the Hyper ChannelTM interface to a VAXTM computer.

We describe the implementation of an archive of scientific data, which will expand at a rate of 1 Terabyte per year, starting from the middle of 1987. The archive, now ready in a prototype fashion, is located at ESO in Garching near Munich, and will serve both the ESO (European Southern Observatory) and the ST-ECF (Space Telescope European Coordinating Facility) needs. Details on the three major system components, that is User Interface, Catalogue Handler and Archive Handler are given, with particular emphasis on the last one, which uses WORM optical disk storage, for which a specific disk format has been adopted.

Extensive data have been taken on the Drexler Technology Corporation's DrexonR recording media. The data have been analyzed to determine the media's response function. The data indicate that the media conforms to a form of the Generalized Mathematical Model of Noise Sources under some recording conditions. A model of the media under these conditions is developed and used to evaluate the effects of recording laser noise.

Part 1: Review of the CRAY-1A/StorageTek (STK) 0,0'3 interface strategy. Part one of the paper is based on a previous publication and will discuss the problems involved in constructing a high-speed data path between a CRAY-1A and the STK optical digital data disk (0D3) utilizing Network Systems Corporation (NSC) HYPERchannel* adapters. The test connection architecture, data structures, data rate experience, and resulting inter-face designes for a Local Device Network (LDN) will be discussed. Part 2: Examination of the interface strategy currently being implemented for a CRAY X-MP/ 4800 Part two of the paper will discuss an interface strategy being put in place for two large-scale computers, a CRAY X-MP/4800 and a CRAY-1A using a new NSC adapter. This strategy can be expanded to a number of large-scale host computers within a local network. At present, this interface does not deal with any optical device. It is being constructed to interface high-speed host computers to IBM-compatible storage devices which are also accessed by a Mass Storage Control Processor (MSCP) node in the NCAR Local Network (NLN). This portion of the paper emphasizes the interface that will be in place whenever an OD**3 device becomes available. Thus, we hope to show (ahead of device availability) what a typical large user of such OD 3 devices expects in the way of interface requirements and logical data formatting of the devices.