Fact-checked by Grok 2 weeks ago

Workplace OS

Workplace OS was a microkernel-based operating system project initiated by IBM in 1991, designed to serve as a unified platform capable of hosting multiple operating systems through compatibility "personalities" such as OS/2, AIX, and POSIX, thereby improving software portability and reducing maintenance costs across diverse hardware environments. The project emerged from IBM's "Grand Unification Theory of Operating Systems" (GUTS), which sought to standardize subsystems and enable a single microkernel to support various OS environments, including servers and personal digital assistants, while leveraging the Mach microkernel developed at Carnegie Mellon University. In collaboration with Apple through the Taligent joint venture formed that same year, Workplace OS incorporated object-oriented user interfaces and programming models, with plans to integrate personalities for MS-DOS with Windows and future iterations of OS/2 and AIX as modular "adaptors." Development emphasized processor independence, symmetric multiprocessing support, and beta releases to hardware vendors, aiming to reassert IBM's influence in the PC market amid competition from Microsoft Windows. Despite an investment exceeding $2 billion over five years, Workplace OS faced significant challenges, including technical complexity, performance overhead from the microkernel design, and market confusion over its role relative to established systems. The sole commercial release occurred on January 5, 1996, as the Warp Connect PowerPC Edition (version 1.0), targeted at PowerPC hardware but undermined by low demand, cost overruns, and delays with the PowerPC 620 chip. The project was discontinued in March 1996, marking it as a notable in IBM's and highlighting the risks of ambitious architectures in the 1990s computing landscape.

Introduction

Objectives

Development of Workplace OS was initiated by in 1991 as part of its broader to consolidate its fragmented operating system amid growing in the market. The emerged from the "Grand Unification Theory of Operating Systems" (GUTS) initiative, aiming to develop a single, modular operating system that could serve as a common foundation across diverse hardware and software environments. This effort was closely tied to IBM's alliance with Apple and , formalized in October 1991, which focused on the PowerPC processor to challenge the Intel-Microsoft dominance and enable cross-platform compatibility. The primary goal was to create a -based OS that supported multiple "personalities," allowing it to emulate and run applications from various existing systems such as , AIX, , and even Windows, without requiring extensive rewrites. By factoring out common services into a portable , IBM sought to enhance across platforms including PowerPC, x86, and others, thereby reducing development and maintenance costs for both and third-party developers. This design emphasized , enabling personalities to provide application-specific interfaces while sharing underlying kernel resources for efficiency. Strategically, Workplace OS aimed to unify IBM's disparate OS offerings—spanning from desktops to servers—thereby minimizing internal fragmentation and improving developer productivity through reusable components. It positioned IBM to compete more effectively against Microsoft's and various Unix variants by offering a versatile, future-proof platform that promised binary compatibility and scalability across device classes. These objectives reflected IBM's response to the rapid evolution of computing in the early , prioritizing and cost savings in a increasingly driven by standardized ecosystems.

Core Architecture

Workplace OS employed a architecture designed to provide a minimal, portable foundation for operating system services, drawing from the 3.0 kernel developed at . The kernel itself handled only essential primitives, such as (IPC) via , thread management, abstraction, and basic I/O port access, while delegating all other functionality—including file systems, device drivers, and networking—to user-space servers. This approach ensured that the kernel remained small and verifiable, reducing the to enhance overall system security and reliability. Central to the architecture were the Basic OS Services (BOS), a set of core utilities implemented as modular user-space servers that provided foundational capabilities like process management, memory allocation, and device configuration, independent of specific operating system personalities. Device drivers operated similarly as isolated servers, allowing for hot-swapping and fault containment without risking kernel stability. These components formed part of the broader Personality Neutral Services (PNS) layer, which offered shared, hardware-agnostic utilities accessible to multiple personalities, such as OS/2 or Unix emulations. The design principles of Workplace OS emphasized reliability through strict of services, preventing a failure in one user-space component from propagating to the or other modules, thereby improving over traditional systems. Scalability was achieved by supporting environments and (), with the facilitating resource sharing across networked nodes or multi-processor setups. Extensibility was a core tenet, enabled by the personality model that allowed third-party developers to implement new OS environments atop the neutral and PNS without modifying the core. In one sentence, this architecture aligned with broader objectives by enabling multi-personality support for legacy applications. Compared to monolithic kernels, such as those in early Unix variants where all services run in privileged kernel mode, Workplace OS's offered superior modularity and fault isolation, as a buggy driver could be restarted independently without system reboot. It also diverged from OS/2's hybrid model, which integrated microkernel-like with substantial monolithic components for performance, by adopting a purer separation that prioritized long-term portability and extensibility over immediate optimization, though this introduced overhead in IPC-heavy operations.

Development

Initial Phases and GUTS

The GUTS project, or Grand Unification Theory of Operating Systems, was an internal effort launched in 1991 to develop a microkernel-based capable of supporting multiple operating system personalities on a single unified kernel. This initiative aimed to address the fragmentation in IBM's OS portfolio by creating a common foundation for systems like , AIX, and , emphasizing modularity and portability across hardware platforms. Building on GUTS, the Workplace OS project emerged in the early 1990s as IBM sought to evolve its operating system strategy amid shifting industry alliances, including influences from Apple's Pink project via the Taligent joint venture. Development involved key contributions from researchers at IBM's Thomas J. Watson Research Center, who adapted GUTS concepts into a more comprehensive microkernel design based on Carnegie Mellon University's Mach kernel. This transition focused on enhancing security and interoperability while maintaining compatibility with existing software ecosystems. Key milestones in the initial phases included the prototyping of the Workplace OS during 1991 and 1992, where efforts centered on core (IPC) mechanisms—such as for efficient kernel-user interactions—and the implementation of basic services like and thread scheduling. These prototypes were tested on early hardware targets to validate the 's ability to host OS personalities concurrently. Despite these advances, the early development faced significant challenges, particularly in integrating legacy code from IBM's operating systems, which required extensive to the environment without compromising stability. Additionally, proved difficult for applications, as the added abstraction layers in the design introduced overhead that hindered responsiveness in time-sensitive scenarios.

Demonstrations and Reorganization

In 1993, showcased a prototype of Workplace OS at the trade show, demonstrating its operation on PowerPC hardware with support for multiple operating system personalities, prominently featuring while initially planning integration with AIX. The demonstration highlighted the system's modular architecture, which aimed to enable seamless compatibility across environments, though it did not fully exhibit live switching between personalities in public view. During the event, Chairman announced that Workplace OS would not replace AIX, citing performance penalties associated with the design that drew resistance from AIX users unwilling to accept reduced efficiency. The demonstration elicited positive industry feedback on the potential of Workplace OS's modularity for improving and reducing maintenance costs, but it also underscored concerns about the prototype's immaturity, particularly in and , which influenced to adjust its development priorities and timelines. These demonstrations built on earlier internal prototyping efforts like GUTS, providing external visibility to the project's ambitious goals. In parallel, collaborated with Taligent, its with Apple, to incorporate object-oriented extensions into Workplace OS, aiming to embed applications directly into the operating system environment for enhanced integration and reusability. Taligent's contributions focused on creating a where individual applications could function as built-in OS components, though cultural and technical challenges limited the venture's impact. Following the 1993 demonstrations, IBM underwent significant business reorganization in 1994 under CEO , who shifted the company toward customer-focused strategies and away from overly ambitious OS replacement projects. This included consolidating operating system development, including Workplace OS efforts, under the Personal Software Products (PSP) division, which handled and related initiatives, to streamline resources and address the high costs of fragmented development—estimated at nearly $2 billion over five years for the project. The restructuring prioritized performance and market realities over pure architectural innovation, marking a pivot in 's software strategy.

Preview Release and Roadmap

The first developer preview, released in December 1994 as the OS/2 PowerPC SDK Beta 1, provided tools for building and testing applications on PowerPC hardware, including a beta C++ and MetaWare High C/C++ cross- running on x86-based systems. This preview marked an early implementation of the foundation, derived from CMU 3.0, with initial focus on the OS/2 personality while laying groundwork for concurrent support of other operating systems like AIX and Windows. The roadmap envisioned Workplace OS 2.0 by 1996, featuring comprehensive PowerPC architecture support, including 64-bit capabilities on the PowerPC 620 processor, full Windows compatibility through personality modules, and enhancements for to enable seamless application migration across environments. Developer tools emphasized the Workplace OS Application Environment, a for applications between personalities without major rewrites, targeting efficiency in multi-OS scenarios. IBM positioned Workplace OS as the cornerstone of its global computing strategy, serving enterprise servers, desktops, embedded systems, and personal digital assistants (PDAs) by unifying and reducing maintenance costs through a shared base.

Technical Components

Microkernel Foundation

The Workplace OS microkernel, officially known as the IBM Microkernel, was built upon an enhanced version of the Mach 3.0 microkernel developed at , providing a minimal set of core mechanisms to support modular operating system construction. This foundation emphasized a small , approximately 70,000 lines of , to minimize the privileged-mode footprint while enabling scalability across devices from handhelds to supercomputers. The design relocated non-essential services, such as file systems and schedulers, to user-mode servers, promoting reliability through isolation and . Central to the microkernel's operation was its interprocess communication (IPC) model, which relied on a message-passing system inherited from , utilizing ports as endpoints for communication between threads and tasks. Ports facilitated both synchronous and asynchronous modes: synchronous IPC blocked the sender until message delivery and processing, ensuring reliable coordination, while asynchronous modes allowed non-blocking notifications for efficiency in distributed scenarios. Threads served as the basic execution units, operating within tasks that acted as resource containers akin to UNIX processes, sharing and other attributes to enable efficient intra-task communication. This port-based threading model supported concurrent execution across multiple processors, with host and processor sets allowing dynamic resource allocation. Kernel services were deliberately limited to essential primitives, including thread creation, management, and termination; basic operations via objects for and ; and scheduling mechanisms for dispatching based on and . support included caching of virtual addresses to reduce overhead in address translation, complemented by a default for handling page faults in the absence of personality-specific pagers. and I/O handling were also kernel-provided, routing hardware events to appropriate user-mode handlers while maintaining minimal intervention. To address the inherent performance overhead of designs compared to monolithic s, optimizations focused on efficiency, such as to avoid unnecessary data duplication between user and spaces. Additional techniques included caching frequently accessed rights and message descriptors to minimize context switches and lookup costs, with Mach-derived costing around 2,300 CPU cycles per operation in baseline implementations. These measures aimed to mitigate in cross-server communications, though evaluations later highlighted persistent challenges in achieving parity with integrated designs. The included security enhancements for higher-assurance configurations, supporting environments via capability-like port rights and isolation between tasks. While many features were planned, only basic implementations were available in the 1996 OS/2 Warp PowerPC release before the project's discontinuation.

Personality Support

Workplace OS implemented personality support through user-space servers that emulated various operating system environments, such as the OS/2 Manager, AIX, and , allowing multiple personalities to run concurrently atop the . These servers functioned by intercepting and translating application system calls into (IPC) messages directed to the , enabling seamless compatibility with legacy software without requiring kernel modifications. Each operated as a replaceable , constructed from user-level servers and libraries that could be swapped or extended independently of the core . For instance, the personality included support for Win-OS/2, which allowed Windows applications to execute within the environment by routing calls through the personality layer. This facilitated the coexistence of diverse runtime environments, with a dominant personality typically managing the and application launching. Despite these advantages, personality support introduced challenges, including performance overhead from the translation of system calls into , which could degrade responsiveness compared to monolithic kernels. The system's extensibility allowed for the development of custom personalities, including object-oriented ones derived from Taligent's contributions, which integrated advanced frameworks for future-oriented applications while maintaining compatibility with the microkernel's mechanisms. This flexibility positioned Workplace OS to potentially host specialized environments beyond standard Unix or variants, though practical adoption was limited by the project's discontinuation.

Hardware and Platform Integration

Workplace OS was primarily designed for the PowerPC architecture as part of the between , Apple, and , aiming to establish a unified platform for high-performance computing. The system targeted IBM's RS/6000 server line, which utilized PowerPC processors for enterprise workloads, as well as the Power Series laptops, such as the models 850 and 860, intended for mobile professional use. Compatibility extended to embedded systems within IBM's ecosystem, leveraging the modular to support diverse hardware configurations without deep kernel modifications. Ports to the x86 architecture were developed to broaden adoption, with Workplace OS planned for release on platforms following initial PowerPC focus, though full binary compatibility across architectures proved challenging. Additional support for other RISC variants was outlined in the roadmap to enable multi-architecture deployment. These efforts emphasized abstraction layers to maintain hardware portability, allowing the OS to run on both big-endian PowerPC systems like RS/6000 and little-endian x86 environments through techniques, such as 486-compatible software for x86 binaries on PowerPC. The driver model in Workplace OS treated drivers as modular servers operating at level, rather than code, to facilitate sharing across multiple operating system personalities. This approach provided multi-architecture abstraction, enabling drivers to interface with independently of the underlying , and supported hot-plugging for dynamic addition or removal without . For instance, PCI-based peripherals on RS/6000 servers and laptops could be managed through these servers, reducing downtime in and scenarios. Porting Workplace OS across architectures highlighted significant challenges in achieving binary compatibility, particularly due to differences—PowerPC's big-endian format versus x86's little-endian—necessitating runtime translation or recompilation for drivers and low-level components. 's integration strategy addressed this by prioritizing in the , ensuring that platform-specific adaptations, such as for RS/6000's or ThinkPad's , did not compromise the OS's cross-personality goals. This hardware focus complemented the personality layers, allowing seamless of legacy software on varied platforms, though only PowerPC versions were commercially released in 1996.

Discontinuation and Aftermath

Reasons for Cancellation

In March 1996, IBM canceled the Workplace OS project after investing nearly $2 billion over five years of development. The announcement came amid mounting challenges, marking the end of an ambitious effort to create a unified microkernel-based operating system capable of hosting multiple "personalities" for OS/2, AIX, and other environments. This followed the limited commercial release of OS/2 Warp Connect PowerPC Edition on January 5, 1996. Technical delays were a primary factor, particularly in achieving stability for the personality layers that were intended to emulate legacy operating systems on the Mach-derived . The complexity of integrating these personalities led to protracted , with the personality—the only one nearing completion—suffering from incomplete features like networking support and unstable file systems in its PowerPC preview release. Performance gaps further exacerbated issues, as the 's design prioritized modularity and fault tolerance over raw speed, resulting in benchmarks that lagged behind competitors such as Microsoft's , which offered superior efficiency on x86 hardware. These shortcomings were particularly evident in server workloads, where customers favored high-performance Unix variants from and Sun over IBM's experimental approach. Market dynamics also played a decisive role, with the rapid dominance of x86 architecture eroding support for alternative platforms like PowerPC. IBM's heavy reliance on PowerPC—tied to the project's roadmap for cross-platform portability—proved misguided as the chip's declined sharply due to performance shortfalls in models like the PowerPC 620 and limited hardware adoption. In response, IBM pivoted strategically to bolster established products, redirecting resources toward Warp for desktop stability and AIX for enterprise Unix needs, effectively sidelining Workplace OS to avoid further dilution of focus. The dissolution of the -Apple Taligent joint venture in December 1995 compounded these pressures, as Taligent's overlapping goals for a portable object-oriented OS mirrored Workplace OS ambitions but failed to yield viable results. Apple's sale of its Taligent stake back to underscored the venture's unfeasibility, prompting IBM to absorb remaining assets while abandoning the broader vision of a personality-agnostic platform that had been previewed in earlier roadmaps.

Internal Assessments

Following the cancellation of the Workplace OS project in 1996, IBM conducted internal evaluations that underscored the challenges encountered during its development. These assessments, detailed in a comprehensive by of California-Riverside researchers, identified the microkernel's inherent complexity as a primary obstacle to achieving rapid iteration and optimal performance. The architecture's reliance on user-level servers for all system services, including extensive (IPC) mechanisms derived from , resulted in substantial overhead and intricate interdependencies that complicated and integration across diverse hardware platforms. Critiques of focused on the project's overambitious scope, which sought to consolidate multiple operating system personalities—such as , AIX, OS/400, , and even Windows compatibility—onto a single under the Grand Unified Technology Strategy (GUTS). This expansive vision led to fragmentation among development teams spread across divisions and external collaborations, including the Taligent with Apple, fostering coordination issues and diluting focus. Internal reviews estimated the effort consumed nearly $2 billion over five years, representing about 0.6% of IBM's annual revenue at the time, yet yielded no commercially viable product. Despite these setbacks, the assessments highlighted positive outcomes, particularly advancements in technology that enabled more robust, modular communication between system components. These innovations provided a blueprint for fault-tolerant designs and were licensed to nearly 20 corporations, universities, and research institutes, fostering external research in microkernel-based systems. Elements of the framework and modularity principles informed subsequent server technologies, including enhancements in the eServer iSeries platform, where improved and communication efficiency supported enterprise workloads. Insights from project contributors emphasized the enduring value of modularity, even amid failure.

Reception

Industry Perspectives

The development and subsequent cancellation of Workplace OS elicited mixed reactions from key partners and competitors in the technology industry during the early to mid-1990s. Apple's involvement through the Taligent joint venture, formed in 1992 to advance object-oriented operating system technology from Apple's Project Pink, initially held promise for a unified platform across Apple and IBM ecosystems. However, as Taligent struggled with integration challenges between Apple's object-oriented model and IBM's procedural Workplace OS components, Apple withdrew in 1995, licensing the technology and leaving Taligent as an IBM subsidiary. This outcome deepened Apple's frustrations with collaborative OS efforts, contributing to its strategic pivot toward acquiring NeXT Software in 1996 to salvage its operating system initiatives, viewing the Taligent experience as a missed opportunity for a cohesive, next-generation platform. Competitors like expressed skepticism toward Workplace OS, seeing it as a potential disruptor to their growing dominance in PC operating systems. In , amid escalating tensions in the IBM-Microsoft partnership over , Microsoft executive publicly critiqued IBM's shifting strategies, stating, "I can’t tell you what IBM’s strategy is, but I do know it’s not working together with us," highlighting concerns that Workplace OS's approach and multi-personality support could fragment the market rather than consolidate it under Windows. Despite these critiques, early 1994 previews of Workplace OS garnered some praise for its innovative modularity; analyst noted in that Taligent's role was to "create an environment… where all applications… are built directly into the operating system," positioning it as a forward-thinking alternative, though he warned it could become "either the definitive operating system of tomorrow or a massive flop." Analyst reports from the mid-1990s underscored Workplace OS's technical prescience in modularity but flagged its timing and platform choices as missteps. A 1993 analysis described the Taligent-Workplace OS collaboration as facing "technological challenge… to develop an … [that] may be even tougher" due to cultural and architectural clashes between partners, predicting delays that ultimately materialized. By 1995, as refocused amid low demand, reports highlighted the project's ahead-of-its-time design—capable of supporting diverse "personalities" like and AIX—but criticized its emphasis on x86 compatibility at a moment when RISC architectures like PowerPC were gaining traction, rendering it untimely for broader adoption. Hardware vendors, including , expressed concerns over the limited software ecosystem for PowerPC, with beta testers noting integration delays that hindered adoption. The project's trajectory influenced mid-1990s debates on PowerPC adoption, as Workplace OS was envisioned as a key enabler for the IBM-Apple-Motorola alliance's RISC processor. Intended to run natively on PowerPC hardware while emulating x86 environments, it aimed to accelerate enterprise shift away from dominance; however, its abandonment in 1996, following experimental ports like for PowerPC, underscored the platform's software ecosystem gaps. This fueled skepticism among hardware vendors and developers, contributing to PowerPC's stalled momentum in PCs despite initial hype, as competitors like Microsoft's brief support for PowerPC (in version 3.51) also waned due to similar OS maturity issues.

Academic Analysis

Academic scholars have examined Workplace OS as a significant in , particularly highlighting the tensions between and performance in large-scale system design. Analyses from the late noted that the project's use of Mach 3.0 as a foundation for multi-server personalities enabled high through user-level servers but introduced substantial overhead from synchronous message-passing mechanisms that mismatched contemporary hardware capabilities. This , intended to support scalable enterprise environments, resulted in servers of excessive size and complexity, complicating maintenance and integration. Similarly, studies described the microkernel's design goals for portability and personality-neutral services but emphasized how fine-grained object-oriented components exacerbated performance degradation, serving as a cautionary example for balancing layers in distributed systems. Workplace OS's implementation influenced subsequent microkernel research, particularly in optimizing (IPC) efficiency, a core challenge exposed by its heritage. Jochen Liedtke's seminal 1995 SOSP paper critiqued first-generation microkernels like —upon which Workplace OS was built—for their bloated IPC primitives, which imposed unnecessary synchronization and context-switching costs, leading to up to an slower performance compared to monolithic kernels. This analysis directly informed the design of second-generation microkernels such as L4, where Liedtke prioritized , asynchronous IPC to minimize overhead while preserving ; subsequent evaluations, including a 2009 study by Matthias Daum et al. in the Journal of Automated Reasoning, referenced Workplace OS alongside as exemplars of these early pitfalls, underscoring how their inefficiencies spurred innovations in kernel-minimalism for and applications. Although exokernel designs from the mid-1990s, such as those proposed by Engler et al., diverged by emphasizing application-specific over traditional IPC, the broader discourse on Workplace OS contributed to debates on shifting OS responsibilities to user space for enhanced customization. Critiques in academic literature from the late often framed Workplace OS as over-engineered for requirements, prioritizing ambitious multi-personality support at the expense of practicality. Assessments argued that the system's elaborate factoring of OS components into modular servers created unnecessary interdependencies, inflating development costs and hindering deployment in production environments. These debates, appearing in and ACM proceedings, influenced later research toward pragmatic variants, emphasizing verifiable over comprehensive feature sets.

Long-Term Legacy

Despite its abrupt cancellation in 1996, Workplace OS has endured as a pivotal in the challenges of developing general-purpose operating systems, informing subsequent research on kernel design trade-offs between , security, and performance. The project's architecture, built on a heavily modified version of the 3.0 , aimed to support multiple hardware platforms and OS personalities for seamless application compatibility, but encountered significant overhead in (IPC) and that undermined its viability for mainstream use. This experience contributed to a broader understanding in the field that pure , while theoretically advantageous for fault isolation, often imposed unacceptable performance penalties in desktop and workstation environments without extensive optimizations. The concept of OS personalities—user-space servers emulating environments like , AIX, and Unix to run unmodified applications—represented an ambitious precursor to multi-environment in operating systems, though its in Workplace OS failed to generalize effectively due to complexity and integration issues. This approach highlighted limitations in achieving transparent compatibility layers, influencing academic and industry critiques that shaped the evolution of subsystem designs in later systems. For instance, the project's documented shortcomings prompted reevaluations of similar ideas, reinforcing the preference for kernels in production environments while preserving principles for specialized applications. In the broader landscape of paradigms, Workplace OS's emphasis on kernel modularity and personality-based indirectly bolstered interest in lightweight OS techniques, though its direct technological inheritance within IBM's appears limited, with no verified adoption in evolutions of or AIX. Instead, the project's legacy manifested in the redirection of efforts toward embedded and real-time domains, where systems like demonstrated viability for stability-critical uses such as automotive controls. Retrospectives from the have drawn parallels to ongoing explorations, noting how Workplace OS's pitfalls prefigured performance optimizations in projects like the L4 family, which addressed Mach's inefficiencies through and fast messaging. Modern operating systems exhibit conceptual echoes of Workplace OS's modular ambitions, particularly in designs prioritizing scalability across devices. Google's OS, announced in and based on the microkernel, reflects renewed pursuit of a unified, personality-like framework to support diverse hardware without dependencies, though it draws more directly from L4 than from Workplace OS's heritage. Similarly, Android's layered architecture enables modular extensions, but remains rooted in a monolithic , underscoring the enduring tension between Workplace OS's vision and practical implementation hurdles. These parallels, highlighted in analyses, position the project as an early, if unsuccessful, milestone in the quest for versatile OS foundations. Archival materials from Workplace OS remain scarce, with no public release of full ; however, snippets and technical from the are preserved in repositories and IBM's historical , facilitating simulations in research settings to study behaviors. These resources have supported educational recreations of the mechanism, aiding analysis of its limits without requiring binaries.

References

  1. [1]
    Workplace OS - ArchiveOS
    Sep 24, 2025 · The entire Workplace OS platform was discontinued in March due to very low market demand, including that for enterprise PowerPC hardware. A ...
  2. [2]
    Workplace OS History: IBM's $2 Billion Microkernel of Failure - Tedium
    Feb 28, 2019 · IBM bet big on the microkernel being the next big thing in operating systems back in the '90s—and spent billions with little to show for it ...
  3. [3]
    IBM WORKPLACE OS KERNEL FROM TALIGENT IS THE KEY TO THE COMPANY's FUTURE PLANS - Tech Monitor
    ### Summary of IBM Workplace OS and Related Plans
  4. [4]
    New Alliance : IBM and Apple Officially Unveil Accord
    Oct 3, 1991 · Produce new computer operating system software that will work on both companies' high-powered desktop computers called workstations. An ...
  5. [5]
    [PDF] The Failure of Personalities to Generalize Brett D. Fleisch Computer ...
    IBM proposed Workplace O perating S ystems (OS) , the porta b le successor of OS/2 b ased on the IBM micro- kernel that supports multiple personalities. IBM en-.
  6. [6]
    [PDF] Microkernels in a Bit More Depth - CSE CGI Server
    IBM Workplace OS (1991–96). → Unify IBM's operating systems (and produce cost savings). •. DOS, OS/2, Posix, AIX, OS/400, WIndows (binary compatible). • all on ...
  7. [7]
    [PDF] Complexity revisited: learning from failures and successes
    IBM Workplace OS (1991-6). One microkernel operating system for all IBM products ... “Personalities” for OS/2, AIX, OS/400, Windows. Binary compatibility ...
  8. [8]
    [PDF] Alook inside the next generation of computing environments ...
    Jan 19, 1994 · IBM's Mach-based Workplace OS, meanwhile, will adhere to the pure microkernel doct1ine, relegating the pager, the scheduler, the secu1ity system ...<|control11|><|separator|>
  9. [9]
    [PDF] Microkernel-based Operating Systems - Introduction - TU Dresden
    Oct 10, 2017 · Case study: IBM Workplace OS. • Main goals: – multiple OS personalities. – run on multiple HW architectures. ARM. PPC x86. MIPS. Alpha. Mach ...
  10. [10]
    http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.46.7315
  11. [11]
    OS/2 Warp, PowerPC Edition
    So no, OS/2 for PowerPC is not the Workplace OS IBM talked about in ~1993. But yes, it's the closest thing to Workplace OS that ever existed. John Cole says ...Missing: Comdex | Show results with:Comdex
  12. [12]
    Ibm Rides Into Microsoft Country - Bloomberg
    ... Gerstner set up to benchmark IBM's ... operating systems, say the OS/400 of IBM's AS/400 minicomputers. ... IBM's Personal Software Products Div. The lack of ...
  13. [13]
    IBM PSP Developer Support News - 1994 - Issue 1 - OS2World.com
    This session gives attendees an opportunity to understand the direction of the Network Transport component of Workplace OS, as well as an early opportunity for ...
  14. [14]
    OS/2 Warp, PowerPC Edition | OS/2 Museum
    ### Summary of OS/2 Warp, PowerPC Edition Details
  15. [15]
    OS/2 PowerPC Beta 1 SDK : IBM - Internet Archive
    Jul 7, 2019 · Two CD-ROMs, IBM part number 31H5773. Includes the PowerPC operating system and cross-development tools that run on Intel-based OS/2 2.1x or Warp.Missing: Workplace | Show results with:Workplace
  16. [16]
    Workplace microkernel and OS: a case study
    ### Summary of Workplace OS Microkernel and OS: A Case Study
  17. [17]
    [PDF] Microkernels - CSE CGI Server - UNSW Sydney
    Fleisch and Mark Allan A. Co.: "Workplace microkernel and OS: a case study." Software: Practice and Experience 28.6 (1998): 569-591. Page 11. © Gernot Heiser ...
  18. [18]
    Workplace OSs: Doors to the Future - EDM2
    Oct 12, 2022 · The Microkernel's memory management provides the mechanisms of memory objects and functions to manipulate all this. Mach IPC (ports) manage ...
  19. [19]
    Workplace Microkernel and OS: A Case Study - OSnews
    Nov 25, 2013 · The basic premise of Workplace OS work was: 1) IBM would adopt and improve the CMU Mach 3.0 microkernel for use on PDAs, the desktop, ...Missing: 1991 | Show results with:1991
  20. [20]
    None
    Nothing is retrieved...<|control11|><|separator|>
  21. [21]
    Experience with the development of a microkernel-based, multiserver operating system
    **Summary of Content from https://ieeexplore.ieee.org/document/595173**
  22. [22]
    [PDF] PowerPC An Inside View - ibmfiles.com
    AIX and OS/2 for the PowerPC, and Motorola is responsible for porting. Windows ... In the Workplace OS architecture, device drivers are provided as common.Missing: demonstration | Show results with:demonstration
  23. [23]
    [PDF] Prep Standardizes PowerPC Systems: 12/27/93 - CECS
    Dec 27, 1993 · Workplace OS will support x86 Windows applica- tions via the Win-OS/2 subsystem, just as in today's. OS/2, using 486 emulation software ...
  24. [24]
    How IBM became the poster child of operating system failure during ...
    Mar 2, 2019 · All the developers back then knew is that the operating systems available in server rooms or on desktop computers simply weren't good enough, ...
  25. [25]
    TECHNOLOGY - Dec. 20, 1995 - Los Angeles Times
    Dec 20, 1995 · Apple, IBM End Taligent Partnership: Apple Computer Inc. and International Business Machines Corp. are ending their partnership in Taligent ...
  26. [26]
    [PDF] Microkernel-based Operating Systems —Introduction— - TU Dresden
    Oct 18, 2022 · “A microkernel does no real work.” – Kernel only provides inevitable mechanisms. – Kernel does not enforce policies. • But what is inevitable?<|separator|>
  27. [27]
    Half an operating system: The triumph and tragedy of OS/2
    Nov 29, 2019 · IBM hastily wrote an experimental port of OS/2 Warp for PowerPC, but abandoned it before it was finished. Workplace OS never got out of early ...
  28. [28]
    https://books.google.com/books?id=G0SCDRC-VPQC&pg=PA93
  29. [29]
    http://archive.fortune.com/magazines/fortune/fortune_archive/1993/04/05/77699/index.htm
  30. [30]
    https://link.springer.com/article/10.1007/s10817-009-9119-8
  31. [31]
    Experience with the Development of a Microkernel-Based, Multi ...
    The operating system personalities were constructed out of a set of user-level personality and personality-neutral servers and libraries. While we made a number ...
  32. [32]
    Proving Fairness and Implementation Correctness of a Microkernel ...
    May 5, 2009 · 21. Fleisch, B.D., Co, M.A.A.: Workplace microkernel and OS: a case study. Softw. Pract. Exp. 28(6), 569–591 (1998). Article Google Scholar.
  33. [33]
    Towards trustworthy computing systems: taking microkernels to the ...
    This paper argues that microkernels are the best approach for delivering truly trustworthy computer systems in the foreseeable future. It presents the NICTA ...Missing: influence | Show results with:influence
  34. [34]
    Microkernels rule! - Embedded
    Jun 25, 2008 · Microkernels, first invented around 1970, are based on the idea of moving the operating system's services out of the kernel (the code that runs ...