Jane C. Blake
Managing Editor

In answer to a recent Journal survey question, half of those who responded said they would like to see a variety of topics in each issue; half prefer the single-topic format. This issue of the Digital Technical Journal will please those who like variety. The papers herein address networking, database performance, software processes, PCs, and chips. In the future, we will try to please readers on both sides of the question by presenting single-topic issues, such as Workflow Software upcoming, and issues that encompass a range of topics.

Our opening paper is about the GIGAswitch packet-switching system, used to increase data transfer among interconnected LANs and in workstation farms. At 6.25 million connections per second, GIGAswitch is among the industry's fastest multiport packet-switching systems. The paper by Bob Souza, P.G. Krishnakumar, Cüneyt Özveren, Bob Simcoe, Barry Spinney, Bob Thomas, and Bob Walsh is a substantive overview of GIGAswitch, the first implementation of which includes an FDDI bridge. The authors discuss the major design issues, including the data link independent crossbar, the arbitration algorithm (called take-a-ticket), and the techniques used to ensure the robustness, reliability, and cost-effectiveness of the switch.

Just as critical as network performance in high-end system environments, such as stock exchanges and banking, is the performance of server-based database management systems. In their paper on the DEC Rdb version 6.0 database, Lou Dimino, Rabah Mediouni, T.K. Rengarajan, Mike Rubino, and Peter Spiro examine earlier steps taken to establish good database performance on AXP systems and compare these and traditional approaches with the new enhancements that shorten the code paths, minimize I/O operations, and reduce stall times. Notable among performance results was a world record for transactions per second in a test of the Rdb version 6.0 database running on a DEC 10000 AXP system.

Data availability is another important characteristic of high-end systems. Volume shadowing is a feature of the OpenVMS operating system that increases availability by replicating data on disk storage devices. The port of Volume Shadowing Phase II to OpenVMS AXP is a case study of how a small engineering team can use software development processes to tackle complex system design tasks and deliver improved quality products on accelerated first-release schedules. Bill Goleman, Robert Thomson, and Paul Houlihan describe key process innovations, including early code inspections and profile testing, which simulates complex scenarios in order to reveal errors.

In the first of two papers on Alpha AXP PC products, Dave Conroy, Tom Kopec, and Joe Falcone trace the design choices, alternatives, and methodology used in the evolutionary development of the first AXP PCs. The authors explore the lessons learned from designing two experimental systems; the first demonstrated the feasibility of building a PC utilizing the DECchip 21064 microprocessor and industry-standard components; the subsequent system incorporated the EISA bus. They then review the design of the DECpc AXP 150 product, which built on the successes of the experimental systems and was completed in just 12 weeks.

Future Alpha AXP PCs and desktop systems will use new Alpha AXP microprocessors that have higher levels of system integration for performance yet employ packaging and clocking techniques that reduce system cost. Dina McKinney, Masooma Bhaiwala, Kwong Chui, Chris Houghton, Jim Mullens, Dan Leibholz, Sanjay Patel, Del Ramey, and Mark Rosenbluth explain the trade-offs and results of the 21066 design, which is the first microprocessor to integrate a PCI bus controller along with many other system-level functions.

At the close of this issue, we have listed the referees who offered their expert advice on the content and readability of papers submitted to the Journal between January 1993 and February 1994. We are grateful for the specialized knowledge they have shared with authors and editors alike.