Browse Prior Art Database

Clone Copy Service Daemon for Networked Client Code Distribution

IP.com Disclosure Number: IPCOM000013257D
Original Publication Date: 2002-Jan-01
Included in the Prior Art Database: 2003-Jun-18
Document File: 4 page(s) / 61K

Publishing Venue

IBM

Abstract

Problems and Invention Advantages

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 46% of the total text.

Page 1 of 4

Clone Copy Service Daemon for Networked Client Code Distribution

Problems and Invention Advantages

The network computer (NC) is a typical model of thin client devices in
network computing infrastructure. Despite its expected advantages, NC
failed to succeed in the recent IT market place.

With its design, NC requires its operating system kernel and application
programs ("boot code") to be provided from the server machine (such as
large PCs and workstations). Especially when a group of NCs are powered-on
at the same time, it is common that the server machine ("boot server")
receives multiple NCs' file download requests and generates heavy network
traffic. Because of this network traffic contention, users may need to wait
for more than 20 minutes to have all the NCs ready for operation. This is
one of the major reasons of NC's failure at this market introduction stage.

As for IBM's NC "Network Station" case, the operating system kernel is
about 2MB in size and more than 10MB of program code may need to be
downloaded to each NC depending on the application programs to use. For
DBCS environment, the operating system kernel grows more than 3MB.
Therefore, depending on the server machine capability and LAN bandwidth
availability, a large number of network computers may require more than
half an hour to make them operational, for example.

This impractical user situation can be improved by increasing the number of
server machines or by starting the NCs at different timing so that only 5
to 10 NCs are accessing the server machine in actuality at one time.
However, those resolutions are costly and/or less-flexible from actual
users' operation point of view.

This invention is to reduce this network contention by defining an
intermediate node NC ("node client") between the server machine and other
remaining NCs ("subordinate clients"). This node client re-distributes the
boot code to its subordinate clients after the node client received the
boot code from the server machine. The node client achieves this
distribution structure only by software implementation but does not require
any additional hardware investment.

For example, in 100 NCs, 10 NCs are selected as the node clients, and
configured to receive the boot code from the server machine. The
subordinate clients (remaining 90 NCs) are to wait until these node clients
are ready and re-distribute the boot code to the subordinate clients. As
each download to the node client is completed, the node clients are started
and the Clone Copy Service Daemon is activated to provide the boot code for
each 9 subordinate clients. The node client starts to provide the copy of
his boot code to his subordinate 9 NCs. By taking this file distribution
mechanism, the time required that the all NCs' are ready will be saved
logically around 70-80%. This problem resolution eliminates one of the
major portion of the current network computer disadvantages.

Methodology

The following figure shows the case that all NCs are requesting the
operating syst...