Browse Prior Art Database

Protocol for Maintenance of Shadow Directory Partitions

IP.com Disclosure Number: IPCOM000119701D
Original Publication Date: 1991-Feb-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 5 page(s) / 152K

Publishing Venue

IBM

Related People

Austin, JH: AUTHOR [+3]

Abstract

Disclosed is a protocol which uses sequence numbers to minimize network traffic while maintaining currency of information in shadow directory partitions at nodes which are interconnected using an asynchronous connectionless technology, such as System Network Architecture Distribution Services (SNA/DS). This protocol is applicable to the general distributed data environment if a single location can be identified (termed "master") for a given data subset (termed "partition") which always has the most up-to-date values of the subset and if this subset is replicated at other locations (termed "shadow") by moving the values in some asynchronous way from the master location to the shadow location.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Protocol for Maintenance of Shadow Directory Partitions

      Disclosed is a protocol which uses sequence numbers to
minimize network traffic while maintaining currency of information in
shadow directory partitions at nodes which are interconnected using
an asynchronous connectionless technology, such as System Network
Architecture Distribution Services (SNA/DS).  This protocol is
applicable to the general distributed data environment if a single
location can be identified (termed "master") for a given data subset
(termed "partition") which always has the most up-to-date values of
the subset and if this subset is replicated at other locations
(termed "shadow") by moving the values in some asynchronous way from
the master location to the shadow location.

      An electronic directory may be segmented into information
partitions in a fashion which is similar to the way the telephone
directory is segmented into telephone books (Fig. 1).  These
partitions are replicated in various network nodes to provide
high-speed access to the directory user in a fashion similar to
printing copies of a phone book and distributing these books to
convenient locations like a phone booth where a user may wish to
refer to the contained information.  Having once replicated the
partition (i.e., published a phone book), the shadows (printed
copies) fall behind in current information content when compared with
the master partition which is being updated with newer information.

      In the conventional directory scheme, a full copy of the master
partition would be sent to each shadow partition location totally
replacing the older version (Fig. 2) in much the same way that a new
phone book is printed and distributed to telephone customers
superseding the previous book.  Since the bulk of information in a
directory is stable over a long period of time, this scheme is quite
wasteful of network resources.  The obvious solution is to distribute
just the changes as they occur.  But when the interconnecting network
is connectionless and asynchronous, some changes may arrive out of
order or not at all yielding a corrupted information base.

      The term "folded integer" is used to indicate a positive value
integer with special meanings applied to the comparison and
arithmetic operators that refer to it.  When the highest value is
incremented, the result is one (1); when 1 is decremented, the result
is the highest value. When values "A" and "B" are compared, "0" is
less than any other value and "A" is greater than "B" if fewer
decrement operations could be performed on "A" to get to "B" than
increments on "A" to get to "B" (e.g., if the maximum positive value
is 255, then "5" is greater than "150" and "67" is less than "150").

      With the disclosed protocol, each master partition contains a
CUR_SEQ_NUM "folded" integer value which is set to zero when the
master is created and is incremented w...