dtj v01-08 feb1989

This issue of the Digital Technical Journal, published in February 1989 by Digital Equipment Corporation, is dedicated to Storage Technology. It showcases Digital's systems-level approach to designing integrated storage and information management products, emphasizing the Digital Storage Architecture (DSA) for interoperability and future-proof customer investments.

The document covers several key areas:

1. Hierarchical Storage Controller (HSC): This section details the HSC as a high-performance, tightly coupled multiprocessor designed as a storage server for VAXcluster systems. It elaborates on its hardware and software structure, emphasizing mechanisms for achieving high performance through techniques like overlapped and optimized seeks, multiple simultaneous transfers, and interprocessor communication through queues and counters. Performance aspects, including processor and bus contention, and latency optimizations, are discussed.

2. RA90 Disk Drive Technology: The RA90 is presented as a leading-edge magnetic disk drive with 1.216 gigabytes of formatted capacity and an average seek time of 18.5 milliseconds. The papers describe the design decisions and technological improvements that contributed to its high capacity, speed, and reliability. This includes advancements in thin-film heads, disk media, and the critical role of Control Systems Technology (e.g., servomechanisms for precise head positioning, adaptive runout correction, automatic bias force correction, and the use of digital signal processors) to ensure accurate and stable operation at high track densities. Kerr Microscopy is also introduced as a crucial tool for observing magnetic domains in thin-film heads, aiding in their design and manufacturing.

3. Fault Management (VAXsimPLUS): This service tool is a fault manager for DSA fixed media disk drives. It's designed to detect, diagnose, recover from (via "Autocopy" – dynamic disk substitution), and report system faults, thereby increasing system availability and preventing data loss.

4. Database Management Systems (VAX DBMS and Rdb/VMS):

   • High Availability Mechanisms of VAX DBMS Software: This paper details how VAX DBMS ensures high system availability in a VAXcluster environment. It explains how data is shared across nodes using cluster-wide global sections and the VMS lock manager, how transparent cluster failover and database recovery are achieved (including undoing uncommitted updates and database unfreeze), and the role of after-image journaling (AIJ) and on-line backup for media error recovery.
   • A Relational Database Management System for Production Applications: This article focuses on enhancements to VAX Rdb/VMS software, originally designed for ad-hoc queries, to support large-scale production applications. Key improvements include the introduction of hash indices, advanced structuring and partitioning capabilities to minimize I/O bottlenecks, record placement control, query optimizer enhancements, and a fast, multi-threaded on-line backup utility.

In essence, the journal provides a comprehensive overview of Digital's integrated approach to storage solutions, from the physical design and control of disk drives to the software layers that manage data integrity, availability, and performance in complex computing environments like VAXclusters.