Browse Prior Art Database

Dynamic Storage and Aging of Statistical Information

IP.com Disclosure Number: IPCOM000123470D
Original Publication Date: 1998-Dec-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 2 page(s) / 89K

Publishing Venue

IBM

Related People

Smith, GC: AUTHOR [+2]

Abstract

Problem When dealing with a highly distributed, networked system of applications which are working together, it is sometimes desirable to keep statistics on what and how much work each individual application has done over a given period of time. For instance, a user may want to view the amount of work a given process on a particular machine has done in the last week.

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

Dynamic Storage and Aging of Statistical Information

   Problem

   When dealing with a highly distributed, networked system
of applications which are working together, it is sometimes
desirable to keep statistics on what and how much work each
individual application has done over a given period of time.  For
instance, a user may want to view the amount of work a given process
on a particular machine has done in the last week.

   It is also desirable to isolate the applications that work
in a distributed system and the developers that create them from the
type of and way that statistics should be kept.  For example, some
users may need statistics grouped by the week and may only want the
last 4 weeks worth of statistics kept available.  Other users may
need statistics for each day, and may need to keep a full years worth
of data.  Some users want to know the size of the data in an
application processed while others may want to know the success of
the processing.  A developer should be able to create an application
that can participate in the network without knowing the specific
statistics a particular user needs to keep.

   Statistics throughout the system must be in a consistent
format, however, because a user should be able to use a single
interface for viewing all statistical information for the entire
system.

   Also, there must not be a requirement of a constantly
running process or thread that removes or cleans up old statistics
after when they are no longer useful.

   Solution

   Instead of forcing the each application to understand what
and how to store information about the work it has done, a system
can be created by which centrally located statistics objects handle
the work of collecting and aging themselves.

   Whenever an application completes a task, it retrieves one
of the statistics objects.  This retrieval could be from an object
request broker, or it could be as simple as restoring the object
from a shared file system, as long as the application has exclusive
use of the object.  The applica...