Multics

Multics implements a single-level store for data access, discarding the clear distinction between files (called segments in Multics) and process memory. The memory of a process consists solely of segments that were mapped into its address space. To read or write to them, the process simply uses normal central processing unit (CPU) instructions, and the operating system takes care of making sure that all the modifications were saved to disk. In POSIX terminology, it is as if every file were mmap()ed; however, in Multics there is no concept of process memory, separate from the memory used to hold mapped-in files, as Unix has. All memory in the system is part of some segment, which appears in the file system; this includes the temporary scratch memory of the process, its kernel stack, etc.

Segments are limited to 256 kilowords, just over 1 MB, because Multics hardware had 18-bit word addresses for the content of a segment. Larger files are "multisegment files" and are handled differently. The 256 kiloword limit was rarely encountered in practice, because at the time, one megabyte of memory was prohibitively expensive.

Another major new idea of Multics was dynamic linking, in which a running process can make external routines available by adding the segments containing them to its address space. This allows applications to always use the latest version of any external routine, since those routines are kept in other segments, which are dynamically linked only when a process first attempts to begin execution in them. Since different processes can use different search rules, different users can end up using different versions of external routines. Equally importantly, with the appropriate settings in the Multics security facilities, the code in the other segment can gain access to data structures maintained in a different process. Dynamic linking in Multics does not require special dynamic-link libraries (DLLs); a program can dynamically link to any executable segment to which it has access rights.

Thus, to interact with an application running in part as a daemon (in another process), a user's process simply performs a normal procedure-call instruction to a code segment to which it had dynamically linked (a code segment that implemented some operation associated with the daemon). The code in that segment can then modify data maintained and used in the daemon. When the action necessary to commence the request is completed, a simple procedure return instruction returns control of the user's process to the user's code.

Multics also supports extremely aggressive on-line reconfiguration: central processing units, memory banks, disk drives, etc. can be added and removed while the system continues operating. At the MIT system, where most early software development was done, it was common practice to split the multiprocessor system into two separate systems during off-hours by incrementally removing enough components to form a second working system, leaving the rest still running for the original logged-in users. System software development testing could be done on the second system, then the components of the second system were added back to the main user system, without ever having shut it down. Multics supports multiple CPUs; it is one of the earliest multiprocessor systems.

Multics is the first major operating system to be designed as a secure system from the outset.^[8] Despite this, early versions of Multics were compromised repeatedly.^[9] This led to further work that made the system more secure, and prefigured modern security engineering techniques. Break-ins became very rare once the second-generation hardware base was adopted; it has hardware support for ring-oriented security, a multilevel refinement of the concept of master mode.^{[citation needed]} A US Air Force tiger team project tested Multics security in 1973 under the codeword ZARF. On 28 May 1997, the American National Security Agency declassified this use of the codeword ZARF.

Multics is the first operating system to provide a hierarchical file system,^[10][11] and file names can be of almost arbitrary length and syntax. A given file or directory can have multiple names (typically a long and short form), and symbolic links between directories are also supported. Multics is the first to use the now-standard concept of per-process stacks^{[disputed – discuss]} in the kernel, with a separate stack for each security ring. It is also the first to have a command processor implemented as ordinary user code – an idea later used in the Unix shell. It is also one of the first written in a high-level language (Multics PL/I),^[1][12] after the Burroughs MCP system written in ESPOL, an expanded version of ALGOL.

The deployment of Multics into secure computing environments also spurred the development of innovative supporting applications. In 1975, Morrie Gasser of MITRE Corporation developed a pronounceable random word generator to address password requirements of installations such as the Air Force Data Services Center (AFDSC) processing classified information. To avoid guessable passwords, the AFDSC decided to assign passwords but concluded the manual assignment required too much administrative overhead. Thus, a random word generator was researched and then developed in PL/I. Instead of being based on phonemes, the system employed phonemic segments (second order approximations of English) and other rules to enhance pronounceability and randomness, which was statistically modeled against other approaches.^[13] A descendant of this generator was added to Multics during Project Guardian.^[14]

In 1964, Multics was developed initially for the GE-645 mainframe, a 36-bit system. GE's computer business, including Multics, was taken over by Honeywell in 1970; around 1973, Multics was supported on the Honeywell 6180 machines, which included security improvements including hardware support for protection rings.

Bell Labs pulled out of the project in 1969; some of the people who had worked on it there went on to create the Unix system. Multics development continued at MIT and General Electric. At MIT in 1975, use of Multics was declining and did not recover by 1976 to prior levels.^[15][16] Finally by slashing prices, MIT managed to lure users back to Multics in 1978.^[17]

In 1974 Honeywell entered into a development contract with the Air Force (with MIT as a sub-contractor) to develop a security kernel for Multics. This would involve reducing the size of the Multics hardcore by moving specific components of the supervisor out of Ring 0. One of the initial steps after carrying out a security evaluation was the implementation of a Multilevel security framework within Multics called AIM (Access Isolation Mechanism). This provided Mandatory access control which could be enabled to supplement the already existing Discretionary access control that Multics already possessed.^[18] The resulting Project Guardian ran until termination in 1976; whilst most of its changes were not added to Multics, some parts of the project such as the proposed Secure Front End Processor was productized by Honeywell as SCOMP (Secure Communications Processor).^[19] The SCOMP and it's STOP operating system would eventually evolved via XTS-200 and XTS-300 into current XTS-400 offering of secure operating systems.^[20]

Honeywell continued system development until 1985. About 80 multimillion-dollar sites were installed, at universities, industry, and government sites. The French university system had several installations in the early 1980s. After Honeywell stopped supporting Multics, users migrated to other systems, such as Unix.

In 1985, Multics was issued certification as a B2 level secure operating system using the Trusted Computer System Evaluation Criteria from the National Computer Security Center (NCSC), a division of the NSA; it was the first operating system evaluated to this level.

Multics was distributed from 1975 to 2000 by Groupe Bull in Europe, and by Bull HN Information Systems Inc. in the United States. In 2006, Bull SAS released the source code of Multics versions MR10.2, MR11.0, MR12.0, MR12.1, MR12.2, MR12.3, MR12.4 & MR12.5 under a free software license.^[21]

The last known Multics installation running natively on Honeywell hardware was shut down on October 30, 2000, at the Canadian Department of National Defence in Halifax, Nova Scotia, Canada.^[22]

In 2006 Bull HN released the source code for MR12.5, the final 1992 Multics release, to MIT.^[23] Most of the system is now available as free software with the exception of some optional pieces such as TCP/IP.^[24]

In 2014, Multics was successfully run on current hardware using an emulator created by Multicians Harry Reed and Charles Anthony.^[25] The 1.0 release of the emulator is available as of 2017^[update].^[26] Release 12.6f of Multics accompanies the 1.0 release of the emulator, and adds a few new features, including command line recall and editing using the video system.^[27]

The following is a list of programs and commands^[28] for common computing tasks that are supported by the Multics command-line interface.^[29][30]

• change_wdir (cwd) - change the working directory
• create_dir (cd) - create a directory
• copy (cp) - copy files
• list (ls) - list files and directories
• print (pr) - print the contents of a text file
• print_wdir (pwd) - print the working directory
• move (mv) - move a file or directory to a different directory
• rename (rn) - rename a file without moving it

• edm - a simple line editor
• emacs - Multics Emacs
• teco

• runoff (rf)

• echo
• if

• gcos (gc) - simulator to run GCOS programs
• sort, merge - sorting and merging of text and binary files

• logout
• who

The Multics shell language supports "active functions", which are similar to commands, but which return a string value. An active function is called by putting the active function name and the arguments to the active function in square brackets [ and ]. The string returned by the active function is substituted into the command in place of the call to the active function. For example, when the command echo [working_dir] is processed, the active function working_dir is run; it returns the full path of the working directory, which is substituted into the command, so that the echo command prints the working directory.

Some programs can act either as commands or as active functions; when run as a command, its result is printed, and when run as an active function, its result is returned as a string.^[30]

Some common active functions are:

• ceil - returns the smallest integer greater than or equal to the argument
• floor - returns the largest integer less than or equal to the argument
• home_dir (hd) - returns the home directory
• ltrim
• rtrim
• trunc - returns the integer part of the argument
• working_dir (wd) - returns the working directory

Peter H. Salus, author of a book covering Unix's early years,^[32] stated one position: "With Multics they tried to have a much more versatile and flexible operating system, and it failed miserably".^[33] This position, however, is said to have been discredited in the computing community because many of Multics' technical innovations are used in modern commercial computing systems.^{[7][better source needed]}

The permanently resident kernel of Multics, a system derided in its day as being too large and complex, was 135 KB of code.^{[citation needed]} The first MIT GE-645 had 512 kilowords of memory (2 MiB), a truly enormous amount at the time, and the kernel used a moderate portion of Multics main memory.

The entire system, including the operating system and the complex PL/I compiler, user commands, and subroutine libraries, consists of about 1500 source modules. These average roughly 200 lines of source code each, and compile to a total of roughly 4.5 MiB of procedure code, which was fairly large by the standards of the day.

Multics compilers generally optimise more for code density than CPU performance, for example using small sub-routines called operators for short standard code sequences, which makes comparison of object code size with modern systems less useful. High code density is a good optimisation choice for Multics as a multi-user system with expensive main memory.

During its commercial product history, it was often commented internally that the Honeywell Information Systems (HIS) (later Honeywell-Bull) sales and marketing staff were more familiar with and comfortable making the business case for Honeywell's other computer line, the DPS 6 running GCOS. The DPS-6 and GCOS was a well-regarded and reliable platform for inventory, accounting, word processing, and vertical market applications, such as banking, where it had a sizeable customer base. In contrast, the full potential of Multics’ flexibility for even mundane tasks was not easy to comprehend in that era and its features were generally outside the skill set of contemporary business analysts.^{[citation needed]} The scope of this disconnect was concretized by an anecdote conveyed by Paul Stachour, CNO/CSC:

When American Telephone and Telegraph was changing its name to just AT&T in 1983, a staffer from Honeywell’s legal department showed up and asked a Multician if he could arrange to have the name changed in all of their computerized documents. When asked when the process could be completed, the Multician replied, "It's done." The staffer repeated that he needed hundreds perhaps thousands of documents updated. The Multician explained that he had executed a global search and replace as the staffer was speaking, and the task was in fact completed.^{[citation needed]}

The design and features of Multics influenced the Unix operating system, which was originally written by two Multics programmers, Ken Thompson and Dennis Ritchie.^[34] Influence of Multics on Unix is evident in many areas, including the hierarchical file system, redirection, the shell, and the naming of some commands. But the internal design philosophy is quite different, focusing on keeping the system small and simple, and so correcting some perceived deficiencies of Multics because of its high resource demands on the limited computer hardware of the time.

The name Unix (originally Unics) is itself a pun on Multics. The U in Unix is rumored to stand for uniplexed as opposed to the multiplexed of Multics, further underscoring the designers' rejections of Multics' complexity in favor of a more straightforward and workable approach for smaller computers. (Garfinkel and Abelson^[35] cite an alternative origin: Peter Neumann at Bell Labs, watching a demonstration of the prototype, suggested the pun name UNICS – pronounced "eunuchs" – as a "castrated Multics", although Dennis Ritchie is said to have denied this.^[36])

Ken Thompson, in a transcribed 2007 interview with Peter Seibel^[34] refers to Multics as "overdesigned and overbuilt and over everything. It was close to unusable. They [Massachusetts Institute of Technology] still claim it's a monstrous success, but it just clearly wasn't". On the influence of Multics on Unix, Thompson stated that "the things that I liked enough (about Multics) to actually take were the hierarchical file system and the shell — a separate process that you can replace with some other process".

Dennis Ritchie wrote that the design of UNIX was influenced by CTSS.^[37]

The Prime Computer operating system, PRIMOS, was referred to as "Multics in a shoebox" by William Poduska, a founder of the company. Poduska later moved on to found Apollo Computer, whose AEGIS and later Domain/OS operating systems, sometimes called "Multics in a matchbox", extends the Multics design to a networked graphics workstation environment.

The Stratus VOS operating system of Stratus Computer (now Stratus Technologies) is very strongly influenced by Multics, and both its external user interface and internal structure bear many close resemblances to the older project. The high-reliability, availability, and security features of Multics are extended in Stratus VOS to support a new line of fault tolerant computer systems supporting secure, reliable transaction processing. Stratus VOS is the most directly related descendant of Multics still in active development and production usage today.

General Motors' Multiple Console Time Sharing System (MCTS) for the Control Data Corporation STAR-100 computer was based on Multics.

The protection architecture of Multics, restricting the ability of code at one level of the system to access resources at another, was adopted as the basis for the security features of ICL's VME operating system.

The Edinburgh Multiple Access System (EMAS) draws particularly on the one-level store concept used by Multics, providing access to files only by mapping them into memory. All memory space is associated with a segment.

• Time-sharing system evolution
• Peter J. Denning
• Jack B. Dennis
• Robert Fano – director of Project MAC at MIT (1963–1968)
• Robert M. Graham (computer scientist)
• J. C. R. Licklider – director of Project MAC at MIT (1968–1971)
• Peter G. Neumann
• Elliott Organick
• Louis Pouzin – introduced the term shell for the command language used in Multics
• Jerome H. Saltzer
• Roger R. Schell
• Glenda Schroeder – implemented the first command line user interface shell and proposed the first email system with Pouzin and Crisman
• Victor A. Vyssotsky

The literature contains a large number of papers about Multics, and various components of it; a fairly complete list is available at the Multics Bibliography page and on a second, briefer 1994 Multics bibliography (text format). The most important and/or informative ones are listed below.

• F. J. Corbató, V. A. Vyssotsky, Introduction and Overview of the Multics System (AFIPS 1965) is a good introduction to the system.
• F. J. Corbató, C. T. Clingen, J. H. Saltzer, Multics – The First Seven Years (AFIPS, 1972) is an excellent review, written after a considerable period of use and improvement over the initial efforts.
• J. J. Donovan, S. Madnick, Operating Systems, is a fundamental read on operating systems.
• J. J. Donovan, Systems Programming, is a good introduction into systems programming and operating systems.

• Jerome H. Saltzer, Introduction to Multics (MIT Project MAC, 1974) is a considerably longer introduction to the system, geared towards actual users.
• Elliott I. Organick, The Multics System: An Examination of Its Structure (MIT Press, 1972) is the standard work on the system, although it documents an early version, and some features described therein never appeared in the actual system.
• V. A. Vyssotsky, F. J. Corbató, R. M. Graham, Structure of the Multics Supervisor (AFIPS 1965) describes the basic internal structure of the Multics kernel.
• Jerome H. Saltzer, Traffic Control in a Multiplexed Computer System (MIT Project MAC, June 1966) is the original description of the idea of switching kernel stacks; one of the classic papers of computer science.
• R. C. Daley, P. G. Neumann, A General Purpose File System for Secondary Storage (AFIPS, 1965) describes the file system, including the access control and backup mechanisms.
• R. J. Feiertag, E. I. Organick, The Multics Input/Output System. Describes the lower levels of the I/O implementation.
• A. Bensoussan, C. T. Clingen, R. C. Daley, The Multics Virtual Memory: Concepts and Design, (ACM SOSP, 1969) describes the Multics memory system in some detail.
• Paul Green, Multics Virtual Memory – Tutorial and Reflections is a good in-depth look at the Multics storage system.
• Roger R. Schell, Dynamic Reconfiguration in a Modular Computer System (MIT Project MAC, 1971) describes the reconfiguration mechanisms.

• Paul A. Karger, Roger R. Schell, Multics Security Evaluation: Vulnerability Analysis (Air Force Electronic Systems Division, 1974) describes the classic attacks on Multics security by a "tiger team".
• Jerome H. Saltzer, Michael D. Schroeder, The Protection of Information in Computer Systems (Proceedings of the IEEE, September 1975) describes the fundamentals behind the first round of security upgrades; another classic paper.
• M. D. Schroeder, D. D. Clark, J. H. Saltzer, D. H. Wells. Final Report of the Multics Kernel Design Project (MIT LCS, 1978) describes the security upgrades added to produce an even more improved version.
• Paul A. Karger, Roger R. Schell, Thirty Years Later: Lessons from the Multics Security Evaluation (IBM, 2002) is an interesting retrospective which compares actual deployed security in today's hostile environment with what was demonstrated to be possible decades ago. It concludes that Multics offered considerably stronger security than most systems commercially available in 2002.

• multicians.org is a comprehensive site with a lot of material
  • Multics papers online
  • Multics glossary
  • Myths discusses numerous myths about Multics in some detail, including the myths that it failed, that it was big and slow, as well as a few understandable misapprehensions
  • Multics security
  • Unix and Multics
  • Multics general info and FAQ Includes extensive overview of other software systems influenced by Multics
• Open source emulator for the GE Large Systems / Honeywell / Bull 600/6000‑series mainframe computers
• Honeywell, Inc., MULTICS records, 1965–1982. Charles Babbage Institute, University of Minnesota. Multics development records include the second MULTICS System Programmers Manual; MULTICS Technical Bulletins that describe procedures, applications, and problems, especially concerning security; and returned "Request for Comments Forms" that include technical papers and thesis proposals.
• Current Multics development home page
• Official historical source code archive at MIT
• Link page to various Multics information at the Library of Congress Web Archives (archived 2001-11-30)
• Multics repository at Stratus Computer at the Wayback Machine (archived 2002-02-19)
• Multics at Universitaet Mainz
• Various scanned Multics manuals
• Multicians.org and the History of Operating Systems at the Wayback Machine (archived 2021-01-14), a critical review of Multicians.org, plus a capsule history of Multics.