Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1999-10-21
2003-08-19
Beausoliel, Robert (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
Reexamination Certificate
active
06609215
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention relates to information processing technology. More particularly the present invention relates to a system and method for simplifying rebuilding a system following a failure.
2. Description of Related Art
The UNIX operating system is a multi-user operating system supporting serial or network connected terminals for more than one user. It supports multi-tasking and a hierarchical directory structure for the organization and maintenance of files. UNIX is portable, requiring only the kernel (<10%) written in assembler, and supports a wide range of support tools including development, debuggers, and compilers.
The UNIX operating system consists of the kernel, shell, and utilities. The kernel schedules tasks, manages data/file access and storage, enforces security mechanisms, and performs all hardware access. The shell presents each user with a prompt, interprets commands typed by a user, executes user commands, and supports a custom environment for each user. Finally, the utilities provide file management (rm, cat, ls, rmdir, mkdir), user management (passwd, chmod, chgrp), process management (kill, ps), and printing (lp, troff, pr).
A multi-user operating system allows more than one user to share the same computer system at the same time. It does this by time-slicing the computer processor at regular intervals between the various people using the system. Each user gets a set percentage of some amount of time for instruction execution during the time each user has the processor. After a user's allotted time has expired, the operations system intervenes, saving the program's state (program code and data), and then starts running the next user's program (for the user's set percentage of time). This process continues until, eventually, the first user has the processor again.
It takes time to save/restore the program's state and switch from one program to another (called dispatching). This action is performed by the kernel and must execute quickly, because it is important to spend the majority of time running user programs, not switching between them. The amount of time that is spent in the system state (i.e., running the kernel and performing tasks like switching between user programs) is called the system overhead and should typically be less than 10%.
Switching between user programs in main memory is done by part of the kernel. Main system memory is divided into portions for the operating system and user programs. Kernel space is kept separate from user programs. Where there is insufficient main memory to run a program, some other program residing in main memory must be written out to a disk unit to create some free memory space. A decision is made about which program is the best candidate to swap out to disk. This process is called swapping. When the system becomes overloaded (i.e., where there are more people than the system can handle), the operating system spends most of its time shuttling programs between main memory and the disk unit, and response time degrades.
In UNIX operating systems, each user is presented with a shell. This is a program that displays the user prompt, handles user input, and displays output on the terminal. The shell program provides a mechanism for customizing each user's setup requirements and storing this information for re-use (in a file called profile).
When the UNIX operating system starts up, it also starts a system process (getty), which monitors the state of each terminal input line. When getty detects that a user has turned on a terminal, it presents the logon prompt; and once the password is validated, the UNIX system associates the shell program (such as sh) with that terminal (typically there are a number of different shells including ksh and csh). Each user interacts with sh, which interprets each command typed. Internal commands are handled within the shell (set, unset); external commands are invoked as programs (ls, grep, sort, ps).
Multi-tasking operating systems permit more than one program to run at once. This is done in the same way as a multi-user system, by rapidly switching the processor between the various programs. OS/2, available from IBM Corporation, One New Orchard Road, Armonk, N.Y. 10504; and Windows 95, available from Microsoft Corporation, One Microsoft Way, Redmond, Wash. 98052, are examples of multi-tasking single-user operating systems. UNIX is an example of a multi-tasking multi-user operating system. A multi-user system is also a multi-tasking system. This means that a user can run more than one program at once using key selections to switch between them. Multi-tasking systems support foreground and background tasks. A foreground task is one the user interacts directly with using the keyboard and screen. A background task is one that runs in the background and does not have access to the screen or keyboard. Background tasks include operations like printing, which can be spooled for later execution.
The role of the operating system is to keep track of all the programs, allocating resources like disks, memory, and printer queues as required. To do this, it must lensure that one program does not get more than its fair share of the computer resources. The operating system does this by two methods: scheduling priority and system semaphores. Each program is assigned a priority level. Higher priority tasks (like reading and writing to the disk) are performed more regularly. User programs may have their priority adjusted dynamically, upwards, or downwards, depending upon their activity and available system resources. System semaphores are used by the operating system to control system resources. A program can be assigned a resource by getting a semaphore (via a system call to the operating system). When the resource is no longer needed, the semaphore is returned to the operating system, which can then allocate it to another program.
Disk drives and printers are serial in nature. This means that only one request can be performed at any one time. In order for more than one user to use these resources at once, the operating system manages them via queues. Each serial device is associated with a queue. When a user program wants access to the disk, i.e., it sends the request to the queue associated with the disk. The operating system runs background tasks (called daemons), which monitor these queues and service requests from them. A request is then performed by this daemon process, and the results are sent back to the user's program.
Multi-tasking systems provide a set of utilities for managing processes. In UNIX, these are ps (list processes), kill (kill a process), and (run a process in the background). In UNIX, all user programs and application software use the system call interface to access system resources like disks, printers, memory etc. The system call interface in UNIX provides a set of system calls (C functions). The purpose of the system call interface is to provide system integrity, as all low level hardware access is under control of the operating system. This prevents a program from corrupting the system.
Upon receiving a system call, the operating system validates its authenticity or permission, executes it on behalf of the program, then it returns the results. If the request is invalid or not authenticated, the operating system does not perform the request, it simply returns an error code to the program. The system call is accessible as a set of ‘C ’ functions, as the majority of UNIX is also written in ‘C ’. Typical system calls are: _read —for reading from the disk unit; _write —for writing to the disk unit; _getch —for reading a character from a terminal; _putch —for writing a character to the terminal; and_ioctl —for controlling and setting device parameters.
The fundamental structure that the UNIX operating system uses to store information is the file. A file is a sequence of bytes, typically 8 bits long, and is equivalent to a character. UNIX keeps track of files internally by assigning each one a unique
Hamilton, II Rick A.
Lipton Steven Jay
Beausoliel Robert
Bonzo Bryce P.
International Business Machines - Corporation
Mims Jr. David A.
Yee Duke W.
LandOfFree
Method and system for implementing network filesystem-based... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for implementing network filesystem-based..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for implementing network filesystem-based... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3128989