Joseph A. Carchidi,
Group Engineering Manager,
PC Integration

In the 1990s, a major shift is occurring in personal computing, from isolated, individual work on desktops to work in groups whose members are located throughout an enterprise. To support this important change, Digital has developed a family of products, called PATHWORKS, that enables personal computer users to make the shift from the standalone machine to the network environment and the resources of larger computer systems.

The roots of PATHWORKS were in place as early as 1980. Digital's engineering management recognized that a significant part of the growth in the computer industry would be redirected from minicomputer to microcomputer products. As the 80s progressed, we learned from our experience in personal computer hardware development and from the direction taken by the growing and highly competitive microcomputer market that industry standard-based products were more important than unique technologies; that is, open systems, comprising standard devices and interconnects, were what customers wanted, not more proprietary systems.

Digital's VAXmate personal computer, introduced in 1987, was built on the industry standard model. Moreover, it offered something no other PC offered at that time: the VAXmate had the network built in. With foresight, engineering management determined that our microcomputer business would tie to our long-standing strength in building networks. Our strategy thus changed from a focus on hardware development to the development of microcomputer software.

The critical question then asked-and the one that lead to PATHWORKS development within Engineering-was whether to provide customers with an upgrade path similar to those of competitors in the PC LAN business at that time, i.e., file and print services, or a network environment that embraced the primary technologies used by customers, i.e., a complete set of networking applications that included file and print services, mail, X servers, and terminal emulators. The strategy that took hold was the latter; we would develop a broad set of products that recognized customers' investments in a range of personal computer and network software. Unlike other single-product PC LAN offerings, this set of products would be engineered to couple large server systems based on CISC and RISC technologies with the primary microcomputer systems and would support operation over a local or wide area network. Furthermore, the mapping between the disparate systems would have to be transparent to users, and without concessions on performance.

This chosen strategy, of course, was not the easier of the two to implement. One of our initial tasks was to select which operating systems to support among the many microcomputer operating systems available in the market. We decided to define the scope of our early development work by supporting the most widely popular personal computers, which are those based on the DOS, OS/2, and Macintosh operating systems. Another important decision was the choice of a network transport that would serve as the basis for the interconnection of the systems selected. We selected Microsoft's LAN Manager software as this transport. MS-NET, the predecessor to LAN Manager, had the advantage of being network transport independent, thus allowing us to utilize the DECnet network to extend the PC LAN software to a wide area network.

In the papers in this issue, you will read about some of the extensive work that has been accomplished since we first embarked upon this software effort. Engineers have designed and implemented file servers and network transports that allow PCs to access files, applications, storage, and print services on the larger VMS and ULTRIX server systems. Further, a PATHWORKS application, called eXcursion, brings together the X Window System, the Windows environment, and the DECnet network. The effect is to link X-so important to users of UNIX systems- with the PC DOS system environment. These combined efforts represent a hallmark in Digital's progress toward open, heterogeneous computing.

Our achievement in the Personal Computing Systems Group has been our steady progress toward providing customers the open computing environment they need. The breadth of our product offering has taken on clear definition within the last year, and we will now begin the work of adding depth to the PATHWORKS product set. The possibilities for future developments are truly astounding. Looking ahead five years from now, client workstations will have the power of supercomputers, and the dramatic progress in parallel computing will bring additional opportunities for data sharing and application developments which are in embryonic stages today. Our challenge in software engineering will be to make all these systems work together in a well-integrated, easy-to-use, well-deployed computing environment.