To hook up any peripheral to your computer, you use a bus that defines what happens on each end of the connection. Here's a look at the latest contenders.
Russell Kay
A bewildering array of bus technologies confronts the computer user today, and more are on the way. Expansion buses (e.g., the VL-Bus and EISA) are the route by which components can communicate directly with the CPU. A mezzanine bus, such as PCI (Peripheral Component Interconnect), Multibus, and Futurebus+, is a type of expansion bus that puts a bridge controller between processors and peripherals--in other words, it attaches another bus to the local bus--to add flexibility or extra processing capabilities. Other buses are designed to solve specific problems, such as the need for frequently removing or replacing small devices in laptops; PCMCIA address
es this with its ``hot-swapping'' capability, and more solutions are on the way.
But this month's state-of-the-art section isn't about any of those buses. Instead, it focuses on some of the newer peripheral buses designed primarily for hooking up external devices. While not so glamorous as the mezzanine buses, peripheral buses are generally more accessible and visible to the end user, especially nowadays when you try to connect more and more devices--tape backup units, CD-ROM drives, scanners, hard drives, RAID systems, video cameras, MIDI devices, sound systems, floppy drives, MO (magneto-optical) drives, network adapters, and more--to your computer.
In the past, the primary options for connecting these devices were the traditional parallel, serial, and SCSI ports. While those are still around, they are being souped up, redefined, and engineered to handle significantly increased bandwidth.
An important standard bus for hooking up peripherals is SCSI, which comes in a wide and sometimes b
ewildering array of flavors--narrow, fast, wide, fast and wide, SCSI-2, SCSI-3, and now serial. In ``SCSI and Beyond,'' Dinah McNutt sorts these out, showing where SCSI came from and, more important, where it's going.
Mark Clarkson's introduction ``Seriously Serial'' takes an in-depth look at the latest in serial interfaces, including the IEEE P1394 (Apple's version is known as FireWire) and the DEC/Signetics Access.bus. (Another high-speed serial bus that deserves future examination is Serial-Storage Architecture, or SSA, which IBM and Micropolis are developing.)
Clarkson also looks at what's being done with the enhanced parallel port to permit faster speeds and expanded capabilities for peripherals designed to hook up to or piggyback on existing PCs. After researching the field, Clarkson concludes that, while the new parallel-port standard is a useful interim measure, in the end, it stretches the port so far beyond its original design that it is unlikely to become a primary peripheral bus; the
serial buses are likely to carry the day.
Finally, another bus that shows significant promise for the future is Fibre Channel. In his examination of the standard, John Bryan describes how it can host a wide variety of interfaces and is usable for optical fiber and copper wire transmissions.
What's a Bus?
One of the problems in discussing peripheral buses starts with the term bus. For example, you can argue that SCSI is not really a bus but, instead, an interface--after all, that's what the I in the acronym stands for. So what are the differences, if any, between a bus, an interface, and a protocol? We used to make clear-cut distinctions among these terms: A bus was the hardware standard that governed how add-in boards would connect to the CPU and described the physical connections or connectors involved. An interface was a low-level description of the electrical signals that each side of a connection expected to see and how the hardware would interpret them. A protocol was a higher-level des
cription of how software would deal with the signals coming from the interface. Those differences are often quite blurred today. We talk almost interchangeably about buses and interfaces, and when you look at the nitty-gritty details that define these various standards, it's not necessarily clear where hardware stops, what needs to be done in firmware, and how much of the job is left to system and application code.
Nor is that the end of the confusion. Some of the peripheral buses that BYTE looks at here do not, for the most part, represent an either/or type of hardware standard as we used to understand it. Frankly, it's confusing when you discover that one thing you're calling a bus can be attached to or made to operate on another thing that you're also calling a bus. What do you call a SCSI device hooked up through a Fibre Channel bus that's been implemented on a card containing a SCSI host adapter that plugs into an EISA slot that's part of a PCI bus? The figure ``Where the Buses Are'' illustrates s
ome of the hierarchical complexity that can occur when more than one bus is attached to a computer system.
The articles in this section should help you make sense of some of the technologies, capabilities, and options that are available in today's peripheral buses, as well as those that are likely to be incorporated into tomorrow's computers.
Figure: Where the Buses Are
Not only do buses attach to computer systems, they can attach to one another, creating a hierarchical series of slots where different peripherals can be attached. This figure shows one possible configuration.
Russell Kay, a BYTE technical editor who writes sitting in front of his computer's bus, has been reporting on the computer industry since 1981. He can be reached on the Internet or BIX at
russellk@bix.com
.
Back Of The Bus
