Multics (Multiplexed Information and Computing Service) was a time-sharing operating system developed by MIT, Bell Telephone Laboratories, and General Electric Company since 1965 (Saltzer, 1974, pp. 8-24). Multics last known installation was shut in 2000 in Canada.
Features
Multics had some unique features. Multics had a program that controlled all hardware elements by relying on “symmetric multiprocessing, paging, and multiprogramming” features (Multics, 2013, para. 1; Daley and Neumann, 1965, pp. 84-100). The system had virtual memory. The hardware had to change virtual addresses from the CPU into real addresses (Multics, 2013, para. 1).
Users could gain access to virtual memory through hierarchical levels by depending on disks, storage and page components. The OS had a segmented memory, which the user could interact with directly. In some cases, different processes shared a segment. Every segment had addresses “consisting of up to 1MB or 0 to 256K words” (Multics, 2013, para. 1).
The architect linked “the file system with the memory access system” (Multics, 2013, para. 1). This allowed others programs to access files by referring to the memory. Multics also had high standards of language usages. It used PL/I language based on the IBM proposal. The use of high-level language for OS in 1960s was an ambitious approach at the time of Multics development.
In addition, the system could support several languages. The system also used relational database. This was the first commercial relational database in the market. Developers designed a secure system from the start of the project. Its security features earned B2 rating awarded by NCSC of the US government.
The OS had online reconfiguration (Schell, 1971, p. 86). The system could run without stopping throughout its lifespan. Users could add or remove some components of the system while it was still running. Such components included CPUs, disk drives, and memory among others.
Paul Karger and Roger Schell studied Multics vulnerability and noted that the OS could not meet security requirements in its original form (Karger and Schell, n.d, pp. 1-24; Downey, 1977, p. 45). Thus, it was not suitable for an open multi-level system. On the other hand, these researchers noted that Multics had superior security features than other current systems.
They observed that Multics could still work today as a basic Secret or Top Secret system. Multics could also provide a platform in which new systems could rely on to develop advanced secure multi-level systems. During software engineering processes, the team spent resources in order to create an effective OS (Gedda, 2009, pp. 1-4).
They developed and wrote the manual prior to the project implementation. The manual contained details about the system. Multics operated on a high-level language together with different codes, designs, programming, and layering technologies. These aimed at controlling the system complexity.
Goals
Developers of Multics aimed at creating a convenient system that could offer remote terminal abilities. They wanted to create a system that could operate in analogue manner like electricity and telephone services. The system had different approaches to configuration. These were alterable without the system or user involvement in the process. In addition, developers of Multics wanted to build a system with highly reliable internal files (Multics, 2013, para. 1).
The system could support selective processes of sharing information (Saltzer, 1974, pp. 388-402). Developers of Multics aimed to create “a hierarchical structure system, which could support information sharing, administrative roles, and decentralized user activities” (Multics, 2013, para. 2; Organick, 1972, p. 74).
At the same time, developers strived at building a system that would run on different types of applications. This would allow Multics to support several programming approaches and other external interfaces. Finally, the team wanted to create a system that could change based on unique needs of users and changes in technologies.
Impacts and influences on Modern day trends
Programmers of Unix operating system had extensive knowledge in the Multics operating system. Influences of Multics are evident in naming of features such as files and segments. However, these two systems had different internal design approaches. Unix aimed at improving some weaknesses in Multics. For instance, Unix wanted to develop a small and simple system in order to reduce the need for many resources.
In other words, Unix would operate on limited hardware compared to Multics. Some scholars claim that the name Unix, which was initially Unics, originated from Multics. They also argued that the ‘U’ stood for uniplexed rather than the complex Multiplexed of Multics. On this note, critics asserted that Unix developers rejected the complex approach to programming.
Instead, its developers concentrated on a single workable method for small computers. The name reflected a low or a ‘castrated’ form of Multics i.e., the pun was UNICS (Garfinkel and Abelson, 1999, pp. 21-41). Peter Seibel noted that Ken Thompson referred to Multics as an overdesigned and overbuilt system (Seibel, 2007, p. 449). These qualities almost made the system unusable.
MIT and Fernando J. Corbato claimed that the system was a success. However, Ken Thompson believed that Multics failed to achieve its goals (Gedda, 2009, pp. 1-4; Seibel, 2007, p. 449). Today, the hierarchical file system and the use of the shell in Multics still have strong influences on programmers because these methods provided simple approaches of replacing other processes while programming.
Fernando Corbato believed that the name Multics was serious and suggested the value and goal of the system. In this context, developers did not associate the name with any of the involved organizations. Corbato agreed that Multics and Unix shared a pun of the name Multics (Gedda, 2009, pp. 1-4).
Other operating systems with Multics influences
Apart from Unix, Multics has also influenced other operating systems. For instance, the Prime Computer operating system (PRIMOS) relied on Multics designs. In addition, the Domain/OS have also used Multics ideologies. The Stratus VOS operating system had relied on “Multics to design its internal and external features” (Seibel, 2007, p. 449).
For instance, Stratus VOS OS borrowed reliability, security, and availability features from Multics. These two projects had some resembling features in reliability, support, and security (Schroeder and Saltzer, 1972, pp. 157-170). Stratus VOS has a direct connection with the Multics, and it remains one of the surviving OS developed from Multics ideologies.
The ICL’s VME operating system relied on Multics security features. Such features had the ability to control access to resources at one side of the system. There are also other operating systems, which have borrowed other features from Multics. Overall, Corbato observes that modern computing relies on Multics approaches like “hierarchical file systems, file access controls, and dynamic linking on demand” (Gedda, 2009, p. 2) in order to develop their operating systems.
Corbato notes that the approach of high-level language was rather ambitious at the time of developing Multics. However, he asserts that it was necessary to improve the effectiveness of every input in the system. Today, he believes that developers would have picked a simple programming language than the PL/I.
References
Daley, R and Neumann, P. (1965). A General Purpose File System for Secondary Storage. Montvale, N.J: AFIPS Press.
Downey, P. (1977). Multics Security Evaluation: Password and File Encryption Techniques. Hanscom AFB, MA: HQ Electronic Systems Division.
Garfinkel, S., and Abelson, H. (1999). Architects of the Information Society: Thirty-Five Years of the Laboratory for Computer Science at MIT. Massachusett : MIT Press.
Gedda, R. (2009). CIO Blast from the Past: 40 years of Multics, 1969-2009. Web.
Karger, P., and Schell, R. (n.d). Multics Security Evaluation: Vulnerability Analysis. Web.
Multics. (2013). Web.
Organick, I. (1972). The Multics System: An Examination of Its Structure. Cambridge, MA: The MIT Press.
Saltzer, H. (1974). Protection and the Control of Information Sharing in Multics. Comm. ACM, 17(7), 388-402.
Saltzer, J. (1974). Introduction to Multics. Cambridge: MIT Press.
Schell, R. (1971). Dynamic Reconfiguration in a Modular Computer System. Cambridge: MIT Project MAC.
Schroeder, D. and Saltzer, J. (1972). A Hardware Architecture for Implementing Protection Rings. Comm. ACM, 15(3), 157-170.
Seibel, P. (2007). Coders at Work: Reflections on the Craft of Programming. New York: APress Publications.