Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
2000-09-13
2004-02-03
Metjahic, Safet (Department: 2171)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000
Reexamination Certificate
active
06687716
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to the management of sharable files, and more particularly to a file consistency protocol for maintaining consistency among sharable files.
2. Description of the Related Art
In today's computing environment computers are increasingly used in a networked environment. As an example, a majority of workplace users are connected to an internal network or system where each user works from a station, such as computer, connected to a server or servers that are part of the network. Each user's computer contains various applications that operate on files which are stored on the server or on their own station, which may offer analogous services. Therefore, a first user may access the server or analogous device to use a file by way of an application on the first user's machine. In addition, a second user may also access the same file and use the contents of the file with an application on the second user's machine. Nonetheless, the situation arises in this shared environment where the first user changes data in the file which is also being used by the second user. As such, file consistency is necessary to ensure that all copies or instances of the file in use contain the same information as updates are performed to each instance of the file.
Currently, the methods available to maintain file consistency among multiple copies of a file include making an independent copy of the file and then manually or explicitly programmatically merging the changes back into the original file; a UNIX type file consistency scheme where an original file is locked after changes are made to a copy; and a distributed file system approach. Each of these prior art techniques will now be discussed in more detail.
As mentioned above, one method of ensuring file consistency requires that an independent copy of the file with a different file name be made. When the user writes changes to with the independent copy, the user must manually integrate the changes made to the independent copy back into the original file, which may be stored on a server. Of course, the changes cannot be merged back into the file while the original copy is in use by another user. As can be appreciated, integrating the changes between files requires significant effort on the part of the users accessing the file, decreases efficiency, and increases the amount of disc space necessary to store the various file copies. Furthermore, when the user makes an independent copy, the independent copy is required to be saved under a different file name (or in a different place, e.g., directory, disk, system, etc). Thus, a subsequent user may not know the new file name or location of the independent copy and may inadvertently access an outdated, invalid copy.
The UNIX operating system also has a method for maintaining file consistency among different copies of files accessible by many users. If multiple copies of a file are being accessed by different users, the UNIX operating system, or applications using the UNIX file system may automatically lock the file so that all users are prevented from saving changes to the file. In order to save changes, the UNIX operating system, file system, or application will also require the user to save the changes to an independent file having a different filename, thereby creating a copy. Thus, a user must wait until all others using the file have closed the file (i.e., making the file no longer shared), or the user must save the file under a different name, which is undesirable for the previously mentioned reasons.
In addition, UNIX allows a user to change the file attributes of the file such that the file is writable by everyone at all times and copies of the file containing different data exist on the UNIX system. As a result, when a file is writable, any modifications made by a first user are independent from other modifications made by a second user. Therefore, the modifications made by the first user do not appear on the copy in use by the second user. By the same token, any modifications made by the second user do not appear on the copy of the file being used by the first user. Thus, all the changes made by individual users are saved to individual copies and are not saved to the original file. Consequently, the original file, and copies of the original file, do not contain the modifications made by both the first and second, etc. users.
As also mentioned, another approach used to facilitate data consistency of a shared file is to use a distributed file system. Under this approach, the distributed file system distributes file data over multiple computers thereby allowing the data to reside at multiple locations simultaneously. With this approach, a user A may have a file A stored on a hard drive of user A's computer. If a user B wants to access the file on user A's computer, the distributed file system copies the file A to user B's memory or hard drive. In this situation, when user B is accessing the file A, a flag is set in the file A which informs the distributed file system that the user B is making changes to the file A. The flag enables the locking of the file A such that when user A attempts to use the file A in a manner that conflicts with user B's usage (i.e., changing the data in the file A), this setting prevents user A from changing the data. When the user B closes the file A, and the user A attempts to write to the file A, the distributed file system informs the user A that changes have been made and the distributed file system will ask the user A if the user A wants a copy of the file A containing the changes. However, this scheme is inefficient because the distributed file system must track who is modifying the file and the distributed file system must ask individual users, such as user A, when they are trying to access the file A, if the user A wants a fresh copy of the file A containing the changes. Additionally, the inefficiency is further compounded since the distributed file system must provide each user with a new copy in each instance when a separate user attempts access to the file. Furthermore, the distributed file system is not always capable of globally providing a new copy of the file containing the changes to all users which, in some situations, may number hundreds of users. In summary, under the distributed file system approach, the user A and the user B may not simultaneously write to the file A. Instead, if the user A is working with the file A in an exclusive manner, the user B must wait for user A to close the file A before user B may access file A. The file A can still be opened, however, the file will be read-only (i.e., no writing is allowed), and its contents are not guaranteed to be valid/current.
An example of a distributed file system is an Andrew file system. The Andrew file system allows a user to share files over a network and a subsystem. However, the Andrew file system requires an undesirable amount of overhead, such as CPU cycling time and increased network traffic. In addition, overhead is further increased due to the increased amount of user intervention involved with the Andrew file system. The Andrew file system requires user applications to communicate with one another before certain actions, such as data modifications, are commenced. Therefore, as the number of shared users on the network using the Andrew file system increases, the complexity of operating an Andrew file system increases.
Another problem with shared environment platforms where users access files on a shared storage medium relates to file locking and file sharing. An example of the current locking/sharing mechanisms available are a revision control system (RCS) and a source code control system (SCCS). With these approaches, a user C can open a file C in a shared mode and a user D may also access the file C. However, when the file C is opened in the shared mode, neither user may ever modify existing data in the file C or add data to the file C. Therefore, the file C (i.e., whil
Goddard Brian
Metjahic Safet
Radiant Data Corporation
LandOfFree
File consistency protocols and methods for carrying out the... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with File consistency protocols and methods for carrying out the..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and File consistency protocols and methods for carrying out the... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3282572