Plan 9 replaced Unix at Bell Labs as the organization's primary platform for research and explores several changes to the original Unix model that improve the experience of using and programming the system, notably in distributed multi-user environments. Plan 9 was a Bell Labs internal project from its start during the mid 1980s. In 1992, the first public release was made available to universities. In 1995, a commercial second release version was made available to the general public. In the late 1990s, Lucent Technologies, who had inherited Bell Labs, dropped commercial interest in the project. In 2000, a non-commercial third release was made under an open source license. In 2002, a non-commercial fourth release was made under a new free software license.
A user and development community, including current and former Bell Labs and MIT members, continues to produce daily minor releases as ISO images. Bell Labs still hosts development. The development source tree is accessible over the 9P and HTTP protocols and is used to keep an installation up to date.
Unix attempted to remove these distinctions, and model all input/output as file operations. All device drivers were required to support meaningful read and write operations as a means of control. This lets programmers use utilities like mv and cp to send data from one device to another without being aware of the underlying implementation details. However, at the time, many key concepts (such as the control of process state) did not seem to map neatly onto files. As new features like Berkeley sockets and the X Window System were added, they were incorporated to exist outside the file system. New hardware features (such as the ability to eject a CD in software) also encouraged the use of hardware-specific control mechanisms like the ioctl system call.
The Plan 9 research project rejected these different approaches and returns to the file-centric view of the system. Each Plan 9 program views all available resources, including networking and the user-interface resources (like the window it is running in), as part of a hierarchical file system, rather than specialized interfaces.
/dev, but these addressed only devices attached by hardware, and did not address networked devices. Under Plan 9 printers are virtualized as files, and both can be accessed over the network from any workstation.
Another Plan 9 innovation was the ability for users to have different names for the same "real world" objects. Each user could create a personalized environment by collecting various objects into their namespace. Unix has a similar concept in which users gain privileges by being copied from another user, but Plan 9 extends this to all objects. Users can easily spawn "clones" of themselves, modify them, and then remove them without affecting the resources from which they were created.
Plan 9 introduced the idea of union directories, directories that combine resources across different media or across a network, binding transparently to other directories. For example, another computer's
/bin (applications) directory can be bound to one's own, and then this directory will hold both local and remote applications and the user can access both transparently. Using the same system, under Plan 9 external devices and resources can be bound to
/dev, making all devices network without additional code.
/procdirectory, in which all running processes are listed, illustrates how these features work together to produce a greater whole. This special Plan 9 "file system" has also been adopted by Linux and other later operating systems. Processes appear as named objects (sub-directories with info and control files) under
/proc, along with other kernel resources, giving the user a dynamic I/O channel to send commands to them and read data from them. The user does not have to use a limited set and form of system calls to interact with the kernel from compiled programs; rather, he or she can use tools such as
catto search, query and manipulate processes.
Users can also mount
/proc directories (and any other special file systems) from any other machines into their namespace as well, interacting with them as if they are local. The result is a distributed computing environment assembled from separate machines — terminals that sit on users' desks, file servers that store permanent data, and other servers that provide faster CPUs, user authentication, and network gateways, all using the existing hierarchical directory/name system familiar to most computer users. A user can "build" a system by collecting up directories on fileservers, applications running on servers, printers on the network and then bind them all together into their personal namespace running on a terminal.
/netis the API for all TCP/IP, and it can be used even with scripts or shell tools, writing data to control files to write and read connections. Relevant sub-directories like
/net/udpare used to interface to prospective protocols. You can implement a NAT by mounting a
/netfrom a perimeter machine with a public IP, while connecting to it from an internal network of private IP addresses, using the Plan 9 protocol 9P in the internal network. Or you can implement a VPN by mounting a
/netdirectory from a remote gateway, using secured 9P over the public Internet.
Here would be an example of using union (a stack) directories in
/net: just like inheritance in OOP, you can take one (possibly remote)
/special directory and bind another local special directory on top of that, adding some new control files and hiding others. The union directory now is like a child object instance of the original parent. The functionality of the original can be partially modified. Consider the
/net file system. If you modify or hide its
/net/udp sub-directory you may control or extend the UDP interface with local filter processes, still leaving the original
/net/tcp running intact, perhaps in a remote machine. Note that name space is per process: if you give an untrusted application a limited, modified
/net union directory, you restrict its access to the net.
All this makes it easy to combine "objects" or file systems written in different languages on different systems, while using standard naming, access control and security of the file system, largely transparently to the programmer.
This is similar to the facility offered by the
mount_portal command in BSD which by convention is mounted on
/p instead of
/net with only
Key to supporting the network transparency of Plan 9 was a new low-level networking protocol known as 9P. The 9P protocol and its implementation connected named network objects and presented a file-like system interface. 9P is a fast byte-oriented (rather than block-oriented) distributed file system that can virtualize any object, not only those presented by an NFS server on a remote machine. The protocol is used to refer to and communicate with processes, programs, and data, including both the user interface and the network. With the release of the 4th edition, it was modified and renamed 9P2000.
An installable runtime environment exists for x86, and Plan 9 has been ported to MIPS, DEC Alpha, SPARC, PowerPC, ARM and other architectures. The system is written in a dialect of ISO/ANSI C. Several applications were originally written in a language called Alef, but have since been rewritten in the same C dialect. Plan 9 can import POSIX applications and can emulate the Berkeley socket interface through ANSI/POSIX Environment APE. Recently, a new application called linuxemu was developed that can be used to run Linux binaries; it is, however, still a work in progress.
However, Plan 9 itself has never surpassed Unix in popularity, and remains primarily a research tool. Plan 9 has been criticized as "seem[ing] to function mainly as a device for generating interesting papers on operating-systems research." Eric S. Raymond in his book The Art of Unix Programming speculates on Plan 9's lack of acceptance:
"Plan 9 failed simply because it fell short of being a compelling enough improvement on Unix to displace its ancestor. Compared to Plan 9, Unix creaks and clanks and has obvious rust spots, but it gets the job done well enough to hold its position. There is a lesson here for ambitious system architects: the most dangerous enemy of a better solution is an existing codebase that is just good enough."
Other critics of Plan 9 include those critical of Unix in general, where Plan 9 is considered the epitome of the "Worse is better" school of operating system design. Common criticisms include the relative lack of "polish" and development in Plan 9's windowing system and Plan 9's relative lack of maturity as a commercial-grade body of software.
Plan 9 proponents and developers claim that the problems hindering its adoption have been solved, and its original goals as a distributed system, development environment, and research platform have been met, and that it enjoys moderate but growing popularity. Inferno, through its hosted capabilities, has been a vehicle to bring Plan 9 technologies to other systems as part of heterogeneous computing grids.
Inferno is a derivative product, originally started by the same group that created Plan 9, and now developed by a British company, Vita Nuova Holdings. Inferno shares many of the same design concepts as Plan 9, but uses a new application programming language, Limbo, and an accompanying virtual machine, Dis. Inferno is marketed as an open source embedded operating system.
"Plan 9 from User Space" (or plan9port or p9p) is a port of most of the notable Plan 9 libraries and applications to Unix-like operating systems.