GB2146812A - Optical storage system - Google Patents

Abstract

In a data storage system comprising a host CPU 10, coupled to a mass storage device 11, 20 by way of a storage director 14, 19 and a channel 12, has code normally adopted for communicating with a non-optical storage device, such as magnetic disc unit 20. The CPU 10 is provided with host resident software 15 to enable the CPU 10 to communicate with an optical storage device 11 as though it were a non-optical storage device. <IMAGE>

Description

SPECIFICATION Optical storage system This invention relates to an optical storage system, wherein data may be stored on, and read from, an optical media such as an optical disk, by means of a "computer".

For the past twenty years or so, there have been two major driving forces at work in the data processing industry which together have revolutionized the way business gathers, stores and interprets information. The first driving force has been the expansion of technological sophistication, as typified by the microcomputer chip. Computing power, which once required rooms full of equipment and kilowatts of electric power to run, can now be found in tiny silicon chips.

The second driving force has been simple economics; the steady and rapid decrease in the purchase cost of such technology. Particularly in the area of memory, as costs drop and capacities increase, there is an inevitable rush to take advantage of the new found memory space and fill it with information. Demand, in this respect, seems to perpetually outstrip supply.

Unfortunately, for users with exceptionally large data storage needs, magnetic based storage technology is beginning to slow down and level off. The theoretically design limits of head gap measurization and immediate recording densities is fast approaching. As a consequence, there is a real requirement for fast improvements in data storage technology. Accordingly, the present invention is directed to an optical storage system, compatible with current magnetic based stored systems, and which greatly increases the capacity of data storage systems.

Conceptually, the optical storage system of the invention combines many of the best features of magnetic disk and magnetic tape. Like magnetic tape, the optical system can store vast amounts of data on relatively inexpensive media. Like magnetic disk, the information once stored can be retrieved very rapidly and in either a sequential or direct access mode. Because of its write-once media, optical storage is uniquely well-suited to applications which require long term data stability and reliability, coupled with a permanent built-in audit trail.

Externally, the optical drive looks very much like a standard disk drive. Mechanically too, there are certain familial similarities: the optical platter is spun at high speed by a drive motor, and the read/write mechanism scans the surface of the platter, driven by a powerful voice-coil positioner, much like a standard magnetic disk unit.

But there the similarities end. Since writing to the platter and reading from the platter is accomplished by coherent beams of light, physical contact between the read/write mechanism and the media is eliminated.

When information is to be written to the optical storage media, the input data stream is used to modulated the output of a high-intensity infrared laser diode. The radiant energy from the laser is focussed to a fine point on the surface of the rotating media, which in turn creates very tiny marks on the active layers of the platter, in synchronization with the modulating pattern applied to the laser diode.

A much lower-power HeNe laser operating in the visible portion of the spectrum is used to read the optically-recorded data. The light from the HeNe laser is directed onto the platter surface, and the relected light is measured by a series of photodiodes. Variations in the intensity of the reflected light correspond to the presence or absence of marks on the media.

Data is read back from the platter as soon as it is written. Any errors detected at this time cause an immediate rewrite of the data. Circuitry and software combine to ensure that data is written essentially error-free.

The optical system is accessed and controlled via commands from the operator console, job control language (JCL) statements attached to application programs, in-programs calls from high level user languages using the host system access method via a complete set of channel commands.

Optical storage systems provide the advantages that significantly more data can be stored on line per unit of dry space, thereby resulting in proportional cooling and power requirement savings in the computer room. In addition, the data storage capacity of off line shelf space is much greater than that obtainable by the use of magnetic tape or disk systems. Based upon accelerated life test, optical media has an anticipated useful life of ten years, and an optical storage subsystem provided with a powerful built-in error detection and correction algorithm.

The optical storage media is not affected by magnetic fields, and data retrieval can occur at speeds comparable with non removable magnetic disks. Further, compared to the quantity of media required to store elect amount of data, optical storage devices have a substantially lower cost per megabyte than either magnetic tapes or magnetic disks.

While the direct adaptation of "computers in the control of the optical storage systems is technologically feasible, it is not economically practical in view of the fact the large expenditure in money and equipment currently invested in magnetic storage systems. The cost of conversion between systems dedicated solely to magnetic disk storage, and systems dedicated thoroughly to optical storage, is prohibited. Accordingly, the invention is directed to the provision of a system wherein present day post computer-magnetic storage systems may be compatible with optical storage systems.

In a conventional magnetic storage system, one or more magnetic disk systems is controlled by means of one or more storage directors of a controlled unit, the control unit communicating with a host CPU by way of channels, such as conventional IBM (RTM) channels. The present day said control unit, such as STC8880 control unit manufactured by storage technology corporation of Louisville, Colorado, or IBM (RTM) type 3880, while programmable, cannot adapt the overall system directly to enable use of optical storage systems with the host computer, much less enable use of both optical disk units and magnetic disk units with common control unit.

Accordingly, the present invention is directed to the provision of means wherein a minimum of modification is required, in current day storage systems for enabling the use of optical disk units, either alone or in combination with magnetic disk units.

Briefly stated, in accordance with the invention, the necessary adaptation of the storage system in general is effected by providing the micro code in storage directors to be employed with optical disk units, and, in addition, providing dedicated programming in the host computer, with respect to data to be stored or read from the optical disk unit. The micro code modification is required in the storage directors adapted for use with optical storage unit disk, in view of the fact that the primitive control systems required for control of optical disk unit are not fully ompatible with the primitive control signals required for control of magnetic disk units. It has been found, however, that mere adaptation of the storage director, by the use of such micro code, cannot solve the problem.Therefore, in accordance with the invention, the computer includes a program associated with data to be directed to, or received from, storage directors, associated with optical disk units, thereby to permit use of the host computer for the control of the optical disk unit without hardware modification. The host computer, then, associates blocks of data with type of storage media employed for such block, i.e., magnetic or optical, and, if the block of data is associated with opticzl storage units, controls the data n accordance with the additional program provided in the host computer. As a consequence, insofar as the user is concerned, it is not material either the data is optical or magnetically stored, conventional high level language being employed in the usual matter for programming.The system provides the advantage of the weekly increase storage capacity of optical disk systems, without the requirement of any hardware modification.

It is of course apparent that different types of control are required for magnetic and optical disk space, since the read/write processes employed in optical disk systems is quite different from that of the magnetic disk systems. In magnetic disk systems, the incremental areas of storage space may be rewritten, as desired. In optical disk systems, however, each incremental area or block of the optical disk may be written only once, the writting of new or changed data requiring the use of previously unused disk space, as well as the blocking of previously written disk space with respect to the given altered data.

In order that the invention will be made more clearly understood, it will now be disclosed in greater detail with reference to the accompanying drawings, wherein: Figure 1 is a simplified block diagram of an optical storage system in accordance with the invention, optionally incorporating magnetic storage; Figure 2 is a block diagram of a stored system in accordance with Fig. 1, illustrating the incorporation of both multiple optical disk units, as well as the optional incorporation of multiple magnetic disk storage devices; Figure 3 is a simplified block diagram of a modified optical storage system in accordance with the invention, employing a plurality of host computers for controlling a common plurality of optical disk units; Figure 4 is an internal view of an optical storage unit; Figure 5 is an illustration of the format of a platter;; Figure 6 illustrates the block format of a physical data block, and Figure 7 is an illustration of the operating panel.

Referring now to the drawings, and more in particular to Fig. 1, therein is illustrated a simplified block diagram of a data storage system including the arrangement of the present invention. The system includes a host CPU 10. In the preferred embodiment of the invention the CPU10 is an IBM370 or 308x mainframe computer, or a computer compatiable therewith. The computer 10 has an MVS (multiple virtual storage) disk operating system therein. MVS is a propietary operating system of IBM (RTM) corporation. The system further includes one or more optical disk units ODU 11. The host CPU is connected to the optical disk unit by way of channels 12, and control unit 13 including a storage director 14. The channels 12 are conventional IBM channels, such as a block multiplexer or selector channel, i.e., communication links, of the computer. The control unit 13 is comprised of a type STC8880 storage control unit manufactured by Storage Technology Corp. of Louisville, Colorado, or IBM3880. In this system, the host CPU has host-resident software 15, hereinafter referred to as OPSAM, is dedicated to control of the ODU. The storage director 14 of the control unit 13 has control-resident software, designated by the reference numeral 16, and the optical disk unit has storage unit resident software indicated by the reference numeral 17.

The control unit 13 may include one or more additional storage directors 19 coupled to the channels 12, for example, for control of one or more magnetic disk units 20.

The software which supports the optical disk unit hardware, and which together with the hardware comprises the total optical storage subsystem, thus consists of three distinct yet interrelated layers.

The host processor issues input/output instructions and provides temporary storage for data, status information, and the channel program to control the I/O channel. The channel fetches channel address words and channel command words and controls the transfer of data between the host computer and the storage director. The storage director interprets and executes commands the channel and controls the interfaces between the channel and the optical storage unit 11. The storage director also furnishes status signals to the host system and performs diagnostic evaluation of the storage director and optical storage unit.

In a typical system, such as shown in Fig. 2 the control unit 13 can contain two or four storage directors, and each storage director can communicate up to eight optical disk or storage units 11.

The optical disk unit is a storage device which offers large capacity and a high data transfer rate to provide low-cost reliable, on the line data storage. The optical storage unit is unlike magnetic storage units in that it has an optical head system which directs laser light to both read on and write from an optical platter. The laser technology increases storage capacity to 4.0 G bytes of user data per platter. Additionally, the platter service may be enclosed in a removable optical media unit.

A A preferred embodiment of an optical disk unit for use in the present invention as disclosed in our copending U.S. Patent Application Serial No. Case JPT-83054, for "optically storing Digital Data," filed September 19, 1983, the disclosure of which is incorporated by reference herein.

The optical storage unit responds to commands from the control unit 13, positions the access mechanism, reads or writes data, serializes or deserializes data, performs index control, and completes error detection and correction and corresponding functions.

The optical disk unit is comprised of an optical storage unit having onboard dedicated control electronics (optical device control). Unlike magnetic disk devices, an optical storage unit does not have a separate control module with device interface requirements. Since each optical storage unit has its own control electronics, an electronic failure only affects one unit and not a string of units.

The optical platter of the optical storage unit is spun by a drive motor. The read/write mechanism scans the surface of the platter, driven by a voice-coil positioner. Laser light is used to both read and write through the platter. Consequently, there is no physical contact between the read/write mechanism and the media.

When information is to be written to the platter, the data stream modulates the output of a high intensity write laser dyout. Laser light is collimated and focused to a fine point on the surface of the rotating media. Tiny marks are formed on the active surface layer of the platter to optically record the data stream pattern.

A A helim neon laser is provided to optically record data. The light from the laser is directed onto the platter surface, is reflected from the platter, and is measured by a group of photodiodes. Variations in the intensity of the reflected light corresponds to the presence or absence of marks on the media.

Data is read back from the platter as soon as it is written. Errors detected at this time cause a defect error to be skipped and an immediate rewrite of the data in a new area. Circuitry and microcode combine to insure that data is written in an error free manner.

The optical device control of the optical disk unit contains the logic necessary for two control interface functions and enables communications with two different storage directors. Thus, as illustrated in Fig. 3, each optical disc unit 11 can communicate with either of the storage directors 14 or 1 4A, enabling control by the host computer 10, 1 OA by way of the channels 13, 1 3A respectively. Each control interface provides the necessary interface paths between the optical device control and its assigned storage director.

The optical media unit is comprised of an optical platter enclosed in a protective housing. The cartridge protects the optical platter from damage during handling, storage, and shipping.

Additionally, the cartridge provides two external surfaces to display the platter serial number.

The platter serial number is a unique alpha numeric sequence assigned by the manufacturer prior to shipment. It is recorded on a platter but appears in both readable form and as a scannable bar code on the cartridge.

The media is based upon an aluminum substrate of the type employed for magnetic disks.

The optical platter provides a non-erasable, recording surface which does not require pre grooving in the data areas. The platter is divided into 71 6 concentric bands. Each band contains 49 tracks consisting of a single factory pre-recorded band positioning track, and 48 user data tracks.

Since the media is organized into bands and tracks, data blocks to be written are sequentially appended to the space remaining in any of the bands of the platter. This permits a number of separate data sets in a single optical platter. Once written, any data block may be addressed and retrieved sequentially or in random order regardless of its position within a band or track.

The host-resident software, in a preferred embodiment of the invention, the IBM proprietary executes in MVS/SP 1.3 operating system environment on an IBM or IBM compatible mainframe, and controls the way the end user interacts with the optical storage subsystem.

The host-resident software provides for interpretation of the job command language statements of the system (DD statements) to allocate and unallocate optical storage resources. It interprets and executes on line operater console commands and system programmer diagnostic commands; provides for a number of optical subsystem-specific utilities; provides the access method required to read/write to the optical storage sub-system.

The host interface, i.e., the host-resident software, thus acts as the interface between the host operating system and the optical storage subsystem. The main purpose of the host-resident software is to provide the vehicle through which the scheduling and the support of the optical storage subsystem can be accomplished without disruption to the host system software environment.

The control unit software is responsible for interpreting the channel command words which originate in the host CPU, and translating them into control interface commands which the control units 13 can execute.

There are five basic types of commands: Control, Write, Read, Sense, and Diagnostic.

Control commands transfer control information rather than data records between main storage and the Storage Director. This information may include an order code specifying a further action to be taken by the Storage Director or storage unit. It may also contain parameters defining the types of operations that are allowed or data areas which may be accessed. The data address field of the CCW designates the location containing the required additional information.

Read commands transfer data from optical storage to main storage. Data is placed in main storage in an ascending order of addresses, beginning with the address specified in the data address field of the Channel command word.

Write commands transfer data from main storage to optical storage. Data is fetched from main storage in an ascending order of addresses, beginning with the address specified in the data address field of the command word.

A sense command transfers 24 bytes of information from the storage director to the Channel.

Among the types of information provided by the sense bytes are error conditions detected in the previous operation and the current status of the Storage Director and storage unit.

Diagnostic commands are used for maintenance purposes only.

The optical storage unit contains a number of microprocessors, each responsible for different functions such as dynamic defects skipping processing, course seek and fine seek control, and overall read/write processing. The storage unit resident software receives the control commands as generated in the control unit 13 and converts them into assembly language commands appropriate to the storage unit internal microprocessors so as to ellicit the proper actions and resonses in the storage unit.

The physical data structure preferably employed on the storage media is described in greater detail in the following paragraph.

The general interface hardware, of the channels 12, provides communication between the storage director and up to eight channels. The interface is provided to two areas of the storage directory, i.e., a microcontroller and a data path. There are 18 inbound signals received from a channel and 19 out bound signals.

The control interface between the storage controller 13 and the optical storage unit passes 36 signals, as illustrated in Appendix D, in a daisy chain configuration which attaches each storage directory to a string of optical storage units. (From 1 to 8 units). Each optical storage unit has a dual path capability which permits the control of each unit to be shared between two storage directories by way of two separate control interfaces.

Referring again to Fig. 1, a further storage director 19 of the control unit 13 may be employed to control a magnetic disk unit 20. As illustrated in Fig. 2, in a more complete system, a plurality of such magnetic disk units 20 may be controlled by the common control unit 13.

The functional characteristics of both the Storage Control Unit and the Optical Storage Unit will be described.

STORAGE CONTROL UNIT The Storage Control Unit of the Optical Storage Subsystem is a Storage Technology 8880 with modified microcode. The Storage Control Unit interfaces with IBM S/370 OEMI Channels and the storage unit control logic, via the IBM Control Interface, and supports either 1.5 or 3.0 Mbyte/sec transfer rates. The 8880 provides inline and offline diagnostics and local and remote maintenance capabilities. It directs the Control Interface protocol, interprets the Channel command sequences and provides storage unit control logic sequencing. Each 8880 can contain two or four Storage Directors which interface with Optical Storage Units. Reference may be made to the 8880 Disk Storage Control Subsystem Product Description Manual for further information.

OPTICAL STORAGE UNIT The Optical Storage Unit (Fig. 4) of the Optical Storage Subsystem is a device with onboard dedicated control electronics (Optical Device Control (ODC)). Unlike magnetic disk devices an Optical Storage Unit does not have a separate Control Module with Device Interface requirements. Because each Optical Storage Unit has its own control electronics, an electronic failure only affects one unit and not a string of units.

The optical platter is spun by a drive motor; the read/write mechanism scans the surface of the platter, driven by a voice coil positioner. Laser light is used to both read from and write to the platter. Consequently, there is never any physical contact between the read/write mechanism and the media.

When information is to be written to the platter, the data stream modulates the output of a high-intensity write laser diode. Laser light is collimated and focused to a fine point on the surface of the rotating media. Tiny marks are formed on the active surface layer of the platter to optically record the data stream pattern.

A Helium Neon (HeNe) laser reads the optically recorded data. The light from the HeNe laser is directed onto the platter surface, is reflected from the platter, and is measured by a group of photodiodes. Variations in the intensity of the reflected light correspond to the presence or absence of marks on the media.

Data is read back from the platter as soon as it is written. Errors detected at this time cause a defect area to be skipped and an immediate rewrite of the data in a new area. Circuitry and microcode combine to ensure that data is written error free.

CONTROL INTERFACE The Optical Device Control contains the logic necessary for two Control Interface functions and enables communications with two different Storage Directors. Each Control Interface provides the necessary interface paths between the Optical Device Control and its assigned Storage Director.

SELF CHECKING CAPABILITY There are two error detection methods to ensure data integrity throughout the data path: parity checking and Error Detection Code (EDC). The EDC detects errors in the Data Buffer, on the media, and in the Read/Write Channel.

Data Path Parity Checking The data path is 1 6-bits wide throughout the machine and carries an odd-parity bit. Parity is generated at the point at which data is injected into the path, and checked at various key locations throughout the path.

Error Detection Code During a write operation all source data entered into the Data Buffer passes through an Error Detection Code (EDC) generator. The generator calculates the number of bytes of error-checking information which is entered into the Data Buffer as part of the logical block. The EDC information is written on the media in the EDC Resynchronizable Data Section and is protected by the ECC.

During a write operation the Readback Check circuitry passes the data through an EDC checker on the output of the Read Dynamic Defect Skipping Buffer. After the Data and EDC bytes are passed through the checker, the result in the checker should be zeroes. This check guarantees data integrity from the input of the Data Buffer through the entire write path and back through the readback circuits.

During a Read operation, the data which has been recovered from the media, placed in the Data Buffer and corrected by the functional microprocessor, is checked as it is shifted out of the Data Buffer. This check protects against a large error which may have exceeded the correction span of the Error Correction Code (ECC) and resulted in a false correction of the data (Chapter 6).

OPTICAL MEDIA UNIT The Optical Media Unit consists of an optical platter enclosed in a protective housing. The cartridge protects the optical platter from damage during handling, storage, and shipping.

Additionally, the cartridge provides two external surfaces to display the platter serial number.

The platter serial number is a unique alphanumeric sequence assigned by the manufacturer prior to shipment. It is recorded on the platter but appears in both readable form and as a scannable bar code on the cartridge.

The media, which is designed for a 10 year useful life, is based upon the existing 1 4-inch diameter precision aluminum substrate technology developed for magnetic disk. The optical platter provides a nonerasable recording surface which does not require pregroovig in the data areas. The platter is divided into 71 6 concentric bands. Each band contains 49 tracks consisting of a single factory prerecorded band positioning track (Home Address Track) and 48 user data tracks.

Because the media is organized into bands and tracks, data blocks to be written are sequentially appended to the space remaining in any of the bands of the platter. This permits a number of separate data sets in a single optical platter. Once written, any data block may be addressed and retrieved sequentially or in a random order regardless of its position within a band or track.

SOFTWARE SOFTWARE LAYERS The software (programming) which supports the Optical Storage Unit hardware, and which together with the hardware comprises the total Optical Storage Subsystem, consists of three distinct yet interrelated layers. These layers are referred to as host-resident, control unit-resident, and storage unit-resident software, respectively. For complete information on software refer to the Optical Storage Subsystem Application Programmer's Guide, System Reference Manual, System Programmer's Reference and related documentation.

HOST-RESIDENT SOFTWARE The outermost layer of the Optical Storage Subsystem programming is the host-resident software, It executes in an MVS/SP 1.3 operating system environment on IBM or IBMcompatible mainframes, and controls the way the end user interacts with the Optical Storage Subsystem.

The host-resident software provides for interpretation of JCL statements (DD statements) to allocate and unallocate optical storage resources. It interprets and executes online operator console commands and system programmer diagnostic commands; provides for a number of optical subsystem-specific utilities; and provides the access method required to read/write to the Optical Storage Subsystem.

In shrt, the host interface acts as the interface between the host operating system and the Optical Storage Subsystem. The main purpose of the host interface is to provide the vehicle through which the scheduling and the support of the Optical Storage Subsystem can be accomplished without disruption to the host system software environment.

CONTROL UNIT-RESIDENT SOFTWARE The middle layer of Optical Storage Subsystem software resides in the 8880 Storage Control unit. This software is responsible for interpreting the Channel Command Words (CCWs) which originate in the host, and translating them into Control Interface (CTL-I) commands which the 8880 can execute. The following section lists and describes the CCWs which the control unit recognizes. Chapter 9 describes the Control Interface.

Channel Commands There are five basic types of commands: Control, Write, Read, Sense, and Diagnostic.

Control Commands Control commands transfer control information rather than data records between main storage and the Storage Director. This information may include an order code specifying a further action to be taken by the Storage Director or storage unit. It may also contain parameters defining the types of operations that are allowed or data areas which may be accessed. The data address field of the CCW designates the location containing the required additional information.

Read Commands Read commands transfer data from optical storage to main storage. Data is placed in main storage in an ascending order of addresses, beginning with the address specified in the data address field of the Channel command word.

Write Commands Write commands transfer data from main storage to optical storage. Data is fetched from main storage in an ascending order of addresses, beginning with the address specified in the data address field of the command word.

Sense Commands A sense command transfer 24 bytes of information from the storage director to the Channel.

Among the types of information provided by the sense bytes are error conditions detected in the previous operation and the current status of the Storage Director and storage unit.

Diagnostic Commands Diagnostic commands are used for maintenance purposes only.

STORAGE UNIT-RESIDENT SOFTWARE The innermost layer of Optical Storage subsystem programming code is the storage unitresident software. The Optical Storage Unit contains a number of microprocessors, each responsible for different functions such as Dynamic Defect Skipping processing, coarse seek and fine seek control, and overall read/write processing.

This layer of software takes the CTL-I commands generated in the 8880 and converts them into assembly language commands appropriate to the storage unit internal microprocessors so as to elicit the proper actions and responses in the storage unit.

DATA FORMAT PHYSICAL DATA STRUCTURES The Optical Storage Subsystem places data on the optical platter according to a well-defined physical hierarchy (Fig. 5). In order of decreasing complexity, these structures are: Bands, Tracks, Sectors, and Blocks. They are described in the following paragraphs.

BANDS The platter is divided into 71 6 bands, of which 713 are available for user data. These data bands are in turn separated by 717 coarse servo tracks, which serve as band delimiters and servo positioning aids. The bands are addressed conecutively in ascending sequence, band 0 to band 715, beginning at the outer diameter. There are three bands which perform specific functions: Index, Platter Table of Contents (PTOC), and Field Engineering (FE).

FE Band The Field Engineering Band (FE Band) is the outermost band (0). It contains a number of prerecorded alignment tracks which are used to check reliability of the read Channel (Read/Write Channel).

Index Band The Index Band (Band 1) is the only band which uses short blocks consisting of 128 bytes.

Data in the Index Band is used to load the Index Memory. The storage unit reads the Index Band when the host CPU issues the 'Load Index Memory' CCW and builds the index memory tables in the local RAM. The tables are organized so that locating data within a particular band is done as quickly as possible.

Platter Table of Contents The Platter Table of Contents (PTOC) (Band 2) uses long blocks which consist of 7904 bytes.

It is used by the host access method to describe partitioning of data on the platter and defines every file contained on the platter.

TRACKS Each band contains 49 concentric tracks. The first track is a home address track, recorded at the time the platter is manufactured, while the remaining 48 tracks are used to record data.

Tracks within a band are numbered consecutively in ascending sequence from the outside to the inside of the band.

Single track capacity is 137,018 user bytes, which is used to record data, framing information, and error correction data. The framing and error correction methods, plus space for defect skips, require about eleven percent of the total space. User data capacity is 1 18, 560 bytes per track.

SECTORS Each track is physically subdivided into the same fixed number of equal length segments called sectors. The sector is the smallest unit of encoded information. Sector boundaries are precisely predefined by a sector pulse which is derived independently of the data track information. There are four types of sectors: 1. Resynchronizable Data Sector (RDSI The RDS sector format is a 2-byte sync word (resynchronization pattern), followed by 32 bytes of encoded user data, for a total of 34 bytes.

Each RDS starts at a known radial position and each track has a fixed number of RDS cells. For rotational orientation the RDS cells are sequentially numbered, starting at the track index point.

2. Preambic The preamble is a sector whose function is to provide Variable Frequency Oscillator (VFO) synchronization at the beginning of a data block and during defect recovery.

3. Block Separator-The block separator is a sector containing a unique pattern which identifies the beginning and ending of a physical data block.

4. Exception Mark-The exception mark is a sector whose unique pattern is written to identify media defects detected by the Read Back Check while writing a data block. It is also used to identify an incompletely written block at the end of a track.

PHYSICAL DATA BLOCKS A physical data block (Fig. 6) is a sequence of sectors written together as a continuous unit. A physical data block consists of a logical data block plus subsystem overhead. When a physical data block is read it is transformed into a fixed-length logical data block. However, the number of sectors which comprise a physical data block is variable because of media defects detected while writing, and the dynamic defect skipping which occurs as a result of such discoveries.

In the absence of detected errors, the physical data blocks recorded on a track have the following format: 1. A block separator sector 2. Two preamble sectors 3. One preamble/resynchronizable data sector 4. One sector containing the PID and one's complement of PID (each repeated once) 5. 247 Resynchronizable Data sectors containing 7904 bytes of user data (long bock) or 4 Resynchronizable Data Sectors containing 128 bytes of user data (short block). User data consists of LID, KEY, and Data.

6. An Error Detection Code Resynchronizable Data Sector (34 bytes) 7. Six Resynchronizable Data Sectors containing the Error Correction Code for the data (204 bytes) 8. A preamble sector 9. A Block Separator Sector Subsequent blocks are written beginning in the sector immediately following the last block separator of the previously written data block.

LOGICAL DATA BLOCKS User data is written to, and read from, a data track in fixed length logical units called blocks.

A block is a byte sequence whose data length is either 128 bytes (short block which is only used for the Index Band) or 7904 bytes (long block). The maximum number of short blocks per track is 222. The maximum number of long bocks per track is 15. A long bock logical structure consists of a physical ID (PID), Logical ID (LID), Key, and Data. A short block consists of PID and Data. To ensure compatibility users must adhere to the structure of a logical data block.

Physical ID The PID is the 8-byte Physical ID supplied by the Optical Device Control as data is written to the platter. The 8-byte entity is appended to the front of every physical block and indicates the band number, track number and relative record number within a band.

The PID is generated by the ODC when data is written to the platter. It can be returned as a part of the record to the channel when the Read Block and PID command or the Read Index Band and PID command are chained.

The four lower order bytes of the PID are called the Physical Search Argument (PSA). The host uses the PSA to address data on the platter. The PSA consists of a 2-byte band number followed by a 2-byte relative block number.

Logical ID The LID is an 8-byte Logical ID supplied by the host access method. The LID is a part of the user record; it can be used as an address and is a user option.

Key The KEY is a string of data of up to 64 bytes appended to user data by the host access method. The KEY is a part of the user record; it can be used for addressing purposes and is user optional.

Data Data consists of a variable number of bytes to comprise either short or long bocks. Short blocks can only be written in the Index Band. All other bands contain only long blocks.

RECORDS A record is the basic unit of transfer between a user and the host software access method.

The minimum length for a record is four bytes. The maximum length of a logical record depends on whether or not keys are recorded with each block as indicated in the following: with keys 2,01 7,28#2,001 ,92O bytes without keys 2,018,304 bytes FILES A file is a user-defined collection of logical records and is the smallest collection of records to which the user can refer. A file is created sequentially according to the entry sequence. There is no limit to the number of records nor to the total number of bytes of data which a file can contain. Once created, a file can be extended, even across platters, if necessary.

A user, through the Job Control Language of the operating system, can specify the maximum logical record length of a file and define a file as keyed or nonkeyed. If a file is defined as a keyed file, the length of the KEY field and the offset to the KEY from the beginning of each logical record must be constant and must be specified. Records can be added to a keyed file only in ascending KEY sequence order.

ACCESS MODES There are three access modes which the host-resident software provides to a user: 1. Sequential write 2. Sequential read 3. Direct read using keys, or logical identifiers, or relative record number SEQUENTIAL WRITE The host software access method in write mode allows the user to sequentially write data to a file. The user presents logical records to the write mode which performs the following functions: 1. Assigns the proper logical identifier to each physical block.

2. Checks the record to ensure that ascending KEY sequence is maintained if the file being written to is a keyed file.

3. Combines the logical records into physical blocks.

4. Constructs and issues the Channel programs to cause the physical blocks to be written to the proper Optical Storage Unit.

5. Constructs the index entry for a band when data is first written to a band. Constructs and issues the Channel program necessary to cause that entry to be written to the index band and to update the index memory.

6. Causes the allocation of more space when the space already allocated for a file is exhausted.

SEQUENTIAL READ The host software access method in sequential read mode allows the user to read sequentially through a file. Upon request the next logical record is returned to the user. The access method in sequential read mode performs the following functions: 1. The deblocking process necesary to retrieve logical records from the physical blocks read.

2. Constructs and issues Channel programs to retrieve physical blocks in the proper order.

3. Causes the mounting and demounting of platters in the proper order if the entity being read spans multiple platters.

DIRECT READ Direct read access is provided either by KEYs or LIDs. Any file can be processed directly by specification of logical identifiers. However, only keyed files can be processed directly using keys. In all cases of direct access the user specifies the search argument to be used for the retrieval of a logical record. The access method in direct read mode performs the following functions: 1. Determines to which volume the input operation is to be directed.

2. Constructs and issues, the Channel program which causes the search for the physical block or blocks containing the logical record requested.

3. Deblocks the physical block to retrieve the logical record.

SEARCH FUNCTIONS Every 7904-byte block begins with a variable length block header (BH). THe BH is from 8 to 72 bytes long depending upon whether the logical records contained in the block have usersupplied KEYs and the length of those KEYs.

If the logical records are keyed, the BH for each block contains a value equal to that of the highest valued KEY in the block. When performing a keyed search the host-resident software supplies the storage unit logic with the desired KEY as part of the Locate CCW. During a nonkeyed search the Locate CCW contains either the PID or LID.

During a search by KEY or LID the hardware retrieves a fixed-length field from the beginning of each block. On-board logic separates the appropriate subfields according to the type of search conducted. The resulting data pattern is compared against the host software-supplied search argument. The search proceeds in the following manner: 1. Hardware scans the index memory for the band which contains a KEY or LID value less than or equal to the target value.

2. The corresponding band is located.

3. The low search bound is set to 1. The high search bound is set to 48.

4. The median of the high and low bounds is calculated and a seek locates the corresponding track.

5. The first BH which can be read is read.

6. If the read value is greater than the target value, set the high bound equal to this value. If the read value is less than the target value, set the low bound value equal to this value.

7. If the high bound/low bound difference is greater than 1, go to step 4. If the difference is equal to 1, go to step 8.

8. The desired record (if it exists) must reside on either the low bound track or the high bound track.

9. Read sequentially through tracks L and H until the desired record is located or until a KEY or LID greater than the target value is located.

PERFORMANCE AND OPERATION INTRODUCTION The performance characteristics of the Optical Storage Unit and details the Operator Panel will be described.

AVAILABILITY AND SERVICEABILITY The following sections describe the major Availability and Serviceability advantages of the Optical Storage Subsystem.

AVAILABILITY Availability is defined in terms of both machine availability and data availability.

Machine Availability Machine availability is enhanced because each Optical Storage Unit has its own control electronics. An Optical Device Control electronic failure impacts only one OSU, rather than a string.

Machine availability can be further enhanced using the dual path feature. If an 8880 Storage Control Unit director fails, another path from the CPU to the OSU is available through a different director.

Data Availability Online fast-responsive-time systems demand very high data availability. In the event of an OSU failure, data availability is greatly enhanced by the customer-removeable 7440 Optical Media Unit. In addition to quick and convenient load/unload time the light (4.3 Ibs) and easily handled OMU reduces operator time.

SERVICEABILITY The Optical Storage Subsystem Serviceability program provides: 1. Ease of AccessAccess to all storage units is through the top or front of the unit.

2. Modular Design-All assemblies are built in modular form for ease of maintenance and are Field Replaceable Units.

3. Diagnostice The subsystem has the capability to interface with both local and remote diagnostic tools.

OPERATOR PANEL The Operator Panel (Table I) and Fig. 7 provides the controls and indicators necessary for system operator use in performing the following functions: 1. Power UP/DOWN sequence.

2. Load/unload optical media unit.

3. Monitoring storage unit activity.

Table OPERATOR PANEL

CONTROL/INDICATOR FUNCTION VOLUME/ID DISPLAY Alphanumeric display LOAD/ATTN SW Executes load sequence if necessary and sets 'READY' if load completed successfully READY IND Illuminated for successful LOAD Indicates OSU ready after Power On UNLOAD SW Executes unload sequence if necessary WRITE ENABLE IND/ LED indicates a write operation FILE PROTECT SW can be performed on the platter;; Switch inhibits writing.

DISPLAY LASER USAGE Displays in Volume ID the amount of time write laser has been in use ~IMPEL Causes an Initial Microprogram Load from the floppy disk *ERROR IND Indicates any abnormal condition SIDLE LOOP IND Indicates no operation in process DATAP TH (A,B) Allows cormiunication with 8880 ENABLE/DISABLE when in Enable position.

SW/IND Two switches/indicators on Dual Path models.

*POWER ENABLE SW/ Software controlled for power off; DELAYED POWER OFF override via Main Circuit Breaker.

NOTE: The Operator Panel is in two sections: a main panel and a subpanel. Items indicated with an * are located on the subpanel; all others are on the main panel.

ERROR RECOVERY INTRODUCTION Error recovery normally involves 8880/Optical Storage Unit and/or system invoked actions.

Recovery actions vary depending upon system configuration.

ERRORS A data error is an error in the bit pattern read from the platter. Errors are caused by defects on the platter surface, transient electronic noise, media noise, or a slight focusing or tracking error.

The major source of errors are media defects.

DYNAMIC DEFECT SKIPPING The Optical Storage Unit uses Dynamic Defect Skipping (DDS) to eliminate the major source of data errors. When a data error occurs while writing to the optical platter, it is immediately detected by using a readback check (RBC) beam, and the data in error is rewritten.

Consequently, all data written to the platter is correct.

CORRECTION OF DATA ERRORS Error detection and correction information is added by the Optical Device Control (ODC) to each block of data when it is written to the platter. When the record is read the error correction information is used to detect the errors that may be present and to correct data where possible.

The error information is generated with a set of predefined rules and is coded to check the validity for all data written in the recorded block.

The error correction code (ECC) generated by the ODC can be used to correct one of the following error conditions: 1. A single error burst contained within 192 bytes.

2. A double error burst, provided that the total length of bytes that contain the error bursts does not exceed 128.

3. Any combination of up to 96 independent error bursts, each burst spanning from 1 byte up to 95 bytes, provided that the total length of bytes that contain the error bursts does not exceed 96 and the distance between any three bursts is not a multiple of 32.

4. Only three error bursts, each burst scanning from 1 to 32 bytes, if distance between them is a multiple of 32 bytes.

Errors that can be corrected with the error correction code are called ECC correctable errors.

ODC hardware corrects those errors on-the-fly, with a two-data block latency required to start the need sequence. Correctable ECC errors are never seen by the host. A count of correctable errors and a count of the number of bytes read is maintained by the Optical Storage Unit and can be read by the host if error rate monitoring is desired.

If a data error is detected and the data is not correctable with the ECC, the 8880 uses channel command retry in repeated attempts to read the data correctly. In many cases, even though the first attempt to read the data is not successful, subsequent attempts made while the access mechanism and focus are offset from the normal position, result in the ECC correctable error. 8880 reports an uncorrectable ECC error only after all entry attempts have failed.

6.5 ERROR HANDLING Table II. Error Condition Table, identifies sense bit configurations set by the Storage Director in Sense Bytes 0, 1, and 2. Each configuration requires that a specific error recovery action (Table Ill be invoked by the host operating system.

Table III ERROR CONDITION TABLE

Byte Bit Name General Description Action Log 0 0 Command reject Programm1ng Error 2 No 0 0 Command reject Programming error wlread 2 No 2 0 Data Xfer info avail info available o O Command reject Programming error during 2 No 2 1 Write op spec write operation O O Command reject Programming error write 2 No 2 0 Data xfer info avail info available 2 1 Write op specified 0 0 Comnand reject Programming error during 2 No 2 1 Yri te op spec write op w/imrecise end 2 5 Inprecise end 0 0 Comnand reject Programming error during 2 No 2 0 Data xfer info write op w/valid write info avail and imprecise end 2 7 Write op specked 2 5 Imprecise end O O O Command reject A Write command received with 1 No 1 6 Write inhibited the Write Inhibit tab In the Read-Only position O 1 Intervention req'd OSU offline or not plugged 3 No for the address O 2 Bus out parity Bus out parity error 3 Yes O 2 Bus out parity Bus out parity error w/ 5 Yes 2 O Data xfer info avail read info available O 2 Bus out parity Bus out parity error w/ 1 Yes 2 1 Write op spec write operation O 2 Bus out parity Bus out parity error w/ 1 Yes 2 1 Write op spec rite operation and 2 5 Imprecise end imprecise end Table II (Cont.d) ERROR CONDITION TABLE

Byte Bit Name | General Description Action Log O 2 Bus out parity Bus out parity error w/ 5 Yes 2 0 Data xfer info write info available available 2 1 Write op specified O 2 Bus out parity Bus out parity error w/ 5 Yes 2 0 Data xfer info write info available available and imprecise end 2 1 Write op specified 2 5 Imprecise end O 3 Equipment check Equipment malfunction 4A Yes O 3 Equipment check Equipment malfunction w/ 5 Yes 2 0 Data xfer info avail read info available O 3 Equipment check Equipment malfunction w/ I Yes 2 1 Write op spec write operation O 3 Equipment check Equipment malfunction w/ 1 Yes 2 1 Write op spec write operation and 2 5 Imprecise end imprecise end O 3 Equipment check Equipment malfunction w/ 5 Yes 2 0 Data xfer info avail write info available 2 1 Write op specified O 3 Equipment check Equipment malfunction w/ 5 Yes 2 0 Data xfer info write info available available and imprecise end 2 1 Write op specified 2 5 Imprecise end O 3 Equipment check Equipment malfunction 1 Yes 1 O Permanent error Storage director retry exhausted or undesirable O 3 Equipment check Equipment malfunction w/ 1 Yes 1 O Permanent error Storage director retry 2 0 Data xfer info exhausted or impossible and read info available Table II (Cont.d) ERROR CONDITION TABLE

Byte Bit Name General Description Action Log O 3 Equipment check Equipment malfunction wt 1 Yes 1 O Permanent error Storage director retry 2 0 Data xfer info exhausted or 1mpossible and 2 1 Write op specified write info available O 3 Equipment check Permanent equipment 4 Yes 1 3 Operator message malfunction of the alternate storage director O 4 Data check Uncorrectable data check, I Yes 1 O Permanent error stor. control retry exhausted O 4 Data check Uncorrectable data check. 1 Yes 1 O Permanent error stor. control retry exhausted 2 0 Data xfer info avail w/read Info available -O 5 Overrun Command retry exhausted I Yes 1 O Permanent error on an overrun condition 0 s Overrun Command retry exhausted 1 Yes 1 O Permanent error on an overrun condition 2 0 Data xfer info avail wl read data available O 5 Overrun Command retry exhausted 1 Yes 1 O Permanent error on an overrun condition 2 0 Data xfer info w/wr1te data available 2 1 Write op specified available 1 1 End of data No block found, track 2 No searched was last track written in band, band was not completely written. on a read operation. 1 1 End of data No block found, track 2 No 2 0 Data xfer info searched was last track available written in band. band was not completely written, wired Info available.

Table II (Cont.d) ERROR CONDITION TABLE

Byte Bit Name General Description Action Log 1 2 End of band No block found, track 2 No searched was last track written in band, band is completely written1 no read/write data available.

.1 2 End of band No block found, track 2 No 2 0 Data xfer info searched was last track available written in band, band Is completely written. wiread info available.

1 2 End of band No block found, track 2 No 2 0 Data xfer info searched was last track available written in band, band is 2 1- rite op specified completely written1 w/ write info available.

1 2 End of band No block found, track 2 No 2 0 Data xfer info searched was last track available written in band, band is 2 1 Write op specified completely written. write 2 5 Imprecise end info available. Number of blocks transferred by channel does not equal number of blocks written to platter.

1 2 End of band No block found, track 2 No 2 1 Write op specified searched was last track 2 5 ImprecIse end written in band, band is completely written. number of blocks transferred by channel does not equal number of blocks written to platter.

1 4 No block found Search failed because 2 0 No track was empty, or band was empty or index search failed or block was not found on HA track for Read HA CCW.

Table II (Cont.d) ERROR CONDITION TABLE

Byte Bit Name General Description Action Log Storage Dlrector or ODC.

1 5 Operation error Invalid operation in 5 Yes 2 0 Data xfer info Storage Director or ODC available wired info available.

1 5 Operation error Invalid operation in I Yes 2 1 Write op spec Storage Director or ODC during write op 1 5 Operation error Invalid operation in 2 0 Data xfer info Storage Director or ODC available w/wrlte info available.

2 1 Write op spec 1 5 Operation error Invalid operation in 5 Yes 2 0 Data xfer info Storage Director or ODC available w/write info and imprecise 2 1 Write op spec end 2 5 Imprecise end 1 7 Track Format Error Unservoable track or 2 Yes unrecoverable mtd-block error on subsequent attempt to read/write.

2 0 Data xfcr info mid-block error w/read available data available.

1 7 Trach format error Unservoable trach or 2 Yes 2 0 Data xfer info mid-bloch error w/write available info available.

2 1 Write op specified 2 3 Environmental Statistical usage/error log 3 Yes data present information is present 2 7 Track empty No data present for 2 No diagnostic read.

Table III ERROR RECOVSRY ACTION TABLE

Action Explanation 1 Print console error message 2 Exit with programming error or unusual condition indication.

3 a. Repeat the operation (from the beginning of the CCW chain) once.

b. If the error condition persists, perform action 1.

4 a. Print console error message for the operator and/or customer engineer.

b. Perform action 4A.

4A a. Repeat the operation (from the beginning of the CCW chain?.

b. If the error condition persists after 10 retries, perform action 1.

5 a. Construct Restart LOCATE.

b. Determine the failed CCW, c. Repeat the failed operation and continue the CCW chain by executing: Restart LOCATE CCW TIC pointer to failed CCW determined by action Sb) d. If the error condition persists, perform action 1.

INTERFACE REQUIREMENTS INTRODUCTION This chapter describes the interface requirements between the Channel and 8880 and between the 8880 and Optical Storage Unit. It also supplies a list of signal lines and cabling requirements.

CHANNEL INTERFACE The Channel Interface hardware provides communication between an 8880 Storage Director and up to eight Channels. The interface is provided to two areas of the Storage Director: Microcontroller and data path.

There are 18 inbound signals received from the Channel and 19 outbound signals generated with 360/370 type line drivers. Table IV provides a list and description of Channel Interface lines.

CONTROL INTERFACE The Control Interface (CTL-I) provides the common logical and physical connection between the 8880 Storage Controller and the Optical Storage Unit. The CTL-I passes 36 signals (Table V in a daisy chain configuration which attaches each 8880 Storage Director to a string of optical storage units (from one up to eight units). Each Optical Storage Unit has a dual path capability which permits the control of each unit to be shared between two 8880 Storage Directors via two separate Control Interfaces.

CONMMUNICATION The Storage Director and each switch interface are attached in parallel to the signal lines of the CTL-I, allowing the simultaneous addressing or polling of Optical Storage Units by the 8880. Output signals from all switch interfaces attached to the same CTL-I are ORed together for input to the Storage Director on common signal lines. The protocol for this communication is based upon polling and selection and must be initiated by the Storage Director. The Optical Storage Unit can request communication by an Unselected Alert Line.

DUAL PATH OPERATION A storage unit communicates with two separate control interfaces through a dual path switch.

A dual path switch may be selected by one inerface while busy on the other or it may be in a neutral position and available to either interface for interrupt polling or selection. The Storage Director can select one of the eight storage units at a time to be connected to the CTL-I through the dual path switch. A selected storage unit remains connected to the CTL-I until it is instructed by a Storage Director to disconnect.

INTERFACE CABLING Table Vl lists channel Interface and Control Interface external cabling requirements.

Table IV CHANNEL INTERFACE SIGNALS

Line Group Line Name Description Channel Bus Parity Used to send data, I/O Out Lines 0 device address, commands, 1 control instruct ions 2 and similar informatIon 3 from the Channel to the 8880.

4 5 6 7 Channel Bus Parity Used to send data, I/O In Lines O IdentifIcation, sense data, 1 and status information 2 from the 8880 to the 3 Channel 4 5 S 7

Line Group Line Name Description Tag Lines Address In Used for special sequences Address Out and for inter locking and Command Out controlling information on Status In the buses.

Service Out Service In Data Out Data In Disconnect In Selection Operational Out Used for scanning or Control Operational In the selection of the Lines Hold Out attached I/O devices.

Select Out Select In Suppress Out Request In Mitering Meter In Used for the conditioning Control Clock Out of usage meters located Lines Mark O In in attached I/O devices and to condition enable/ disable switches.

Table v .CONTROL INTERFACE SIGNALS 8880 Storage Director Asserted Lines

Line Name Description ,Tag Bus Out Send control information to storage unit; 0 odd parity is always presented.

4 5 6 7 Parity Tag Gate Indicates the presence of a Tag command on Tag Bus Out.

Select Hold# Maintains corrinunication with storage unit until sequence of Tag and Unit commands is completed.

Sync Out Clocks Bus Out during data transfers from Storage Director to storage unit.

End Response Acknowledges an active Normal End or Check End initiated by storage unit.

Control Bus Out Transfers control or address information O to the storage unit; transfers data from the Director to selected storage unit during 2 a Transfer Out sequence; Odd parity is 3 always presented.

4 5 6 7 Parity Table V (Cont.d) CONTROL INTERFACE SIGNALS.

I Optical Storage Unit Asserted Lines

Sync In Clocks and controls timing during extended data transfers; controls data transfer rate between Storage Director and storage unit during execution of an extended Tag command Tag Valid Indicates that storage unit has validated and accepted a Tag command and modifier; Bus In data is valid Select Active Indicates that communlcatlon between Storage Director and storage unit is being maintained.

Normal End Indicates that execution of a Tag . command has been completed with no detected errors.

Check End Indicates that execution of an extended Tag command has been terminated abnormally Error Alert Reports an abnormal condition occurring while transferring data across the CTL-I Index Alert Indicates addressed storage unit has a Short Busy status with respect to switch interface FE Alert Active whenever a storage unit has FE Service Request set Control Bus In Transfers data from Optical Storage Unit O to 8880 2 ~3 4 5 6 7 Parity Table VI EXTERNAL CABLING REQUIREMENTS

Interconnected Cable Part Quantity Notes Units Description Number Channel to Channel 33007XXX 2 per Channel 1,3 8880 Tag/Bus 50031XXX 1 per Channel 1,3 EPO Storage Director CTL-I 33007XXX 2 per Director 1,2 to Optical EPO 50031XXX 1 per 1,2 Storage Unit storage unit Optical Storage Dual Path 33007XXX 2 per 1 Unit to Optical subsystem Storage Unit Optical Storage An AC power cable of 15 feet (4.6 m) Unit to AC power is provided with the Storage Unit.

Notes: 1. An 'X' in cable part number indicates length and can be specified in 5-foot (1.5 m) increments.

2. Maximum length of cable between Storage Director and Storage Unit String is 150 feet (46 ml.

3. 2880 Channel maximum cable length is 280 feet (85.3m) Length is reduced 15 feet for each Storage Director.

303X, 3041, 3081, and 4341 System maximum cable length is 400 feet (120m). Length is reduced 15 feet for each 8880 Control Unit & Storage Director. A maximum of 400 foot cables are needed for Channels with 3.0 Mbyte/sec transfer rate/data streaming feature.

All other CHNL-I cable lengths remain at 250 feet (76m).

While the invention has been disclosed and described with reference to a limited number of embodiments, it is apparent that modifications may be made therein within the scope of the invention, and it is therefore intended in the following claims to cover each such variation and modification as falls within the true spirit and scope of the invention.

Claims (10)

1. A data storage system wherein a host CPU is coupled to a mass storage device by way of a channel and a storage director of a control unit, in that order, the CPU having a determined operating system, wherein said mass storage device comprises an optical disk unit, said storage device having microcode therein for providing control signals adapted to said optical disk unit in response to signals from said channel, said host CPU having host-resident software operating on said operating system for enabling communication with said optical storage device as though it were a non-optical storage device.

2. The data storage system of claim 1, wherein said control unit comprises a second storage director coupled to control a magnetic storage device, and controlled from said host CPU by way of a channel, said host CPU communicating with said magnetic storage device indepen entity of said host-resident software.

3. The data storage system of claim 1, wherein said optical disk unit is arranged to communicate with said host CPU, selectively by way of a pair of storage directors of said control unit.

4. A method for communicating between a host CPU having a given operating system and an optical disk unit by way of a channel coupled to the CPU and a storage director of a control unit; said method comprising identifying blocks of signals in said CPU as data to be written to or read from said disk unit, modifying said control signals in said CPU with host-resident software and directing it to said storage director by way of said channel, and in response thereto, deriving control signals in said director for controlling said optical disk unit to read or write data.

5. In a data storage system wherein a host CPU is coupled to a plurality of mass storage devices by way of a channel and separate storage directors of a control unit, in that order, the CPU having a determined operating system; the improvement wherein at least one of said mass storage devices comprises an optical disk unit, the storage device therein having microcode therein for providing control signals adapted to said optical disk unit in response to signals from said channel, said system further comprising software operating on said operating system for enabling communication between said host CPU and optical disk unit as though said optical disk unit were a non-optical storage device.

6. The data storage system of claim 5, wherein another of said mass storage devices comprises a magnetic disk unit, said software being host-resident in said CPU, said software being operative only on data associated with said optical unit.

7. A data storage system comprising a CPU connected to a mass storage device by way of a channel, said CPU being normally adapted for communication with a non-optical type storage device, ~wherein said mass storage device is an optical storage device and means are provided for enabling communication between said CPU and said optical storage device.

8. A system as claimed in claim 7, in which said means enabling communication comprises a storage director associated with said optical storage device.

9. A system as claimed in claim 7 or 8, in which said means enabling communication comprises means adapting the operting system of said CPU.

10. A data storage system substantially as hereinbefore described with reference to the accompanying drawings.