Sun-2
A workstation built not for glamour but for the grimy work of bootstrapping networked Unix—underpowered by later standards, yet stubbornly capable where it counted.
The Sun-2 was never meant to dazzle. It arrived as a transitional machine, soldered together from Motorola’s slightly enhanced MC68010 CPU and a Multibus backplane that owed more to industrial control systems than to desktop elegance1911. Yet within university labs and engineering offices, it carved out a reputation: not for speed, but for being just capable enough to run SunOS and talk to the network—often without a disk of its own1. The standard Sun-2/120 shipped diskless, relying on Ethernet to fetch its operating system from a server, a configuration that spoke volumes about Sun’s early vision of distributed computing1. This was not a machine for standalone heroics, but for quiet participation in a larger system.
Memory was a fiddly affair. The base configuration started with a single 1 MByte board, and expansion required physical DIP switch settings—eight tiny toggles at position U506 on each memory board—to assign a starting address in 1 MByte increments1. Up to three additional boards could be added to the Sun-2/120, for a maximum of 4 MBytes1. The Sun-2/170 allowed only two expansions, also capping at 4 MBytes, though the part numbering of memory boards hints at possible undocumented configurations: 2MB and 4MB CPU boards existed (part numbers 501-1163, 501-1164, 501-1208), suggesting factory-built variants that sidestepped the manual configuration burden4.
Its display, at 900 x 1152 pixels, was a full-page vertical format that matched the paper-centric workflows of technical drafting and code listing911. Graphics were handled by the Sun-I Video system or, optionally, the Sun Color Video Controller (part number 501-1014), a board that brought modest color capability to the Sun-2/120X and other variants14. Floating-point math, absent from the CPU, could be offloaded to the Sky Floating Point Processor (part number 370-1029), a dedicated coprocessor that plugged into the backplane14.
The bus architecture was a telling compromise. Most Sun-2 models used Multibus, an aging standard originally designed for 8086 systems911. The CPU board used the Multibus P1 connector only for power—+5 volts and ground—leaving data and address lines to be routed elsewhere, a design quirk that simplified power distribution but reflected the ad-hoc integration of disparate components1. Some Sun-2 models used VMEbus instead of Multibus911. Storage options included an SMD controller (cs 451) and SCSI controllers such as the 501-1138 and 501-10454. Notably, Sun-2 systems lacked support for SCSI-3, a limitation that would later complicate diskless booting as storage evolved2.
Software support was anchored in SunOS, with Release 3.4 current by January 1988 and 3.5 arriving weeks later3678. The operating system ran from on-board memory and I/O spaces, mapped into the CPU’s address space as Type 0 (23-bit, 0x0–0x7FFFFF) and Type 1 (14-bit, 0x0–0x3FFF)51012. The PROM monitor provided basic boot and diagnostic functions, though it lacked a diagnostic switch or EEPROM—troubleshooting meant interpreting boot messages and manual jumper settings911. Two PROM versions existed: an older one used up to SunOS 4.0, and a newer version with added features introduced afterward911.
Despite its limitations, the Sun-2 played a dual role in software development: it could act as a host for cross-compiling to Sun-3 and Sun-4 systems, and conversely, it could be the target of compilation from those more powerful platforms7. This made it a node in a development pipeline, not a dead end. Software availability was modest but focused: Sun offered Modula-2, FORTRAN, Pascal, and PHIGS (a 3D graphics API) specifically for the Sun-2, with PHIGS licenses priced at $3,600 on either 1/4" or 1/2" media1516. SunLink products, enabling IBM mainframe connectivity, were compatible with SunOS 3.x but not later releases, marking the Sun-2 as a machine tied to a specific software epoch156.
The Sun-2 was never sleek. It came in desktop, deskside, and fileserver configurations, its form dictated by function rather than aesthetics911. It was replaced not with fanfare but with inevitability, as the 68020-based Sun-3 and then SPARC-based Sun-4 moved the goalposts beyond reach. But for a moment, in the late 1980s, the Sun-2 was enough: a networked Unix terminal with just enough memory, just enough I/O, and just enough CPU to keep the lights on in a lab or engineering office. It was not elegant, but it was honest.
References
- 800-1110-01 - SunOS 1.1 System Manager's Manual for the Sun Workstation - Models 120 and 170
- 812-8910-01 SoftwareTechnicalBulletinOct89
- 812-8801-03 - Software Technical Bulletin March 1988 (1988)
- 813-2004-14 CardcageSlotAssignments&BackplaneConfigRevA 29Sep88 (2004)
- 800-3378-10 ChangePagesAddendaDocbox 24Apr89
- 812-8801-04 - Software Technical Bulletin April 1988 (1988)
- 812-8801-02 - Software Technical Bulletin February 1988 (1988)
- 812-8801-01 - Software Technical Bulletin January 1988 (1988)
- 800-1736-10 SunPROMUsersManualRevA 9May1988 (1988)
- 800-1780-10 WritingDeviceDrivers RevA 9May88
- 800-1736-10 PROMUsersManual RevA 9May88
- 800-1304-05 WritingDeviceDriversSunWorkstationRevA 19Sep86
- 812-8912-01 SoftwareTechnicalBulletinDec89
- Sun US Price List Dec88