Abstract

The SCSI parallel interface (SPI) is designed to provide an efficient peer-to-peer I/O bus with the maximum number of hosts and peripherals determined by the bus width (i.e., 8 or 16). Data may be transferred asynchronously or synchronously at rates that depend on implementation. The objectives of the SCSI parallel interface are: a) To provide host computers with device independence within a class of devices. Thus, different disk drives, tape drives, printers, optical media drives, and other SCSI devices may be added to the host computers without requiring modifications to generic system hardware. b) To provide compatibility such that conforming SPI-2, SPI-3 devices may interoperate with SPI-5 devices given that the systems engineering is correctly done. The interface protocol includes provision for the connection of multiple SCSI initiator ports (i.e., SCSI devices capable of initiating an I/O process) and multiple SCSI target ports (i.e., SCSI devices capable of responding to a request to perform an I/O process). This international standard defines mechanical, electrical, timing requirements, command, and the task management delivery protocol requirements to transfer commands and data between SCSI devices attached to an SCSI parallel interface, i.e. to allow conforming SCSI devices to interoperate. It is intended to be used in conjunction with the SCSI command sets. The resulting interface facilitates the interconnection of computers and intelligent peripherals and thus provides a common interface standard for both system integrators and suppliers of intelligent peripherals. This standard is a functional description. Conforming implementations may employ any design technique that does not violate interoperability. This standard has made obsolete single-ended and multimode signalling alternatives. Implementations that use single-ended or multimode signalling alternatives should reference the SCSI Parallel Interface-2 standard (ISO/IEC 14776-112).