Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Copying Large Data Objects with Directory Attributes

IP.com Disclosure Number: IPCOM000104561D
Original Publication Date: 1993-May-01
Included in the Prior Art Database: 2005-Mar-19
Document File: 4 page(s) / 123K

Publishing Venue

IBM

Related People

Gladney, HM: AUTHOR

Abstract

Disclosed is a method of improving the performance of a distributed electronic library of digital images and other objects. The kind of library to which this applies is shown in the figure.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 42% of the total text.

Copying Large Data Objects with Directory Attributes

      Disclosed is a method of improving the performance of a
distributed electronic library of digital images and other objects.
The kind of library to which this applies is shown in the figure.

      Sketch of a Digital Library - The digital library reproducts
essential characteristics of systems based on paper, emulating
aspects of using or managing a library which has books, pictures, and
other material objects.  Like its material counterpart, the digital
library is intended for objects worth retaining for long times and
valuable to many people.  Since some expect documents to be stored
remotely, a library service network is planned (see figure).  Each
library client-typically a workstation-uses a cache.  Each library
consists of a catalog server and one or more image servers.  The
catalog store, a data base describing the contents of a library,
controls access to anything in that library.  Each image store maps
object identifiers to the bit-streams which constitute the library
collection.

      Figure.  Network of Library Clients, Servers, and Stores:
image simply means "data object representing something other than
itself"

      Summary of the Invention - For large data objects being copied
between backing stores and caches as required for such digital
libraries, certain network messages, database calls, and record
schema are required whether or not searches are done directly in the
cache catalog.  The idea is to extend these messages, calls, and
schema by the most useful search attributes.  The costs (in cycles
and latencies) are not appreciably increased by slightly longer
records, so normal service overhead is not visibly increased.
However, queries which would have otherwise required workstation to
backing store communications can be executed entirely in
workstations.

      To what has been disclosed elsewhere, adding the following
steps is now proposed:

1.  The step of including, as part of the network messages for any
    object copying or moving procedure, an additional authentication
    string which is hard to guess, with this string being chosen
    afresh more frequently than session authentication strings,
    possibly for every object move/copy.

2.  The step of including, as part of the network messages conveying
    the response to the client for any image moving or copying order,
    such information from the library server catalog as is needed to
    shadow, in the cache catalog, those portions of the library
    server catalog which describe each object which is held in the
    client cache.

3.  The additional step in which the aforementioned message and cache
    catalog contents contains object timestamps-last referenced and
    last changed-as recorded by the library servers' clock.

4.  The additional step in which the last changed timestamps are used
    to resolve conflicting updates in two...