Transcript Remote Automated Administration of Public Computer

 2000, Yantar Analitik
Remote Automated Administration of
Public Computer Terminals (RAAPCT)
Dan Trubow
31 October 2000
CS689 Research Methods
 2000, Yantar Analitik
RAAPCT











 2000, Yantar Analitik
Definitions
Comparison of TV/Computer Admin
Research Problem and Objective
Importance
Background & Current Research
Research Design and Plan of Action
Deliverables
Delimitations
Timing
Feasibility
Conclusion
What is Computer Administration?





Installing & Updating Operating System
Installing & Updating Application Software
Authentication and Access
Security
Verification of Proper Operation
 2000, Yantar Analitik
What is RAAPCT ?



Remote – via the Internet
Automated Administration – NO human intervention
Public Computer Terminal
 Located in a business or organization
 Ready access around the clock
 Limited functionality, mainly for Internet access
 Not part of a LAN or WAN, but stand alone
 Could be a server/terminal setup
 2000, Yantar Analitik
Television Set-Up & Administration





Find a suitable location
Connect to power and antenna or
cable
Connect ancillary equipment
Perform station programming
Use for years with little further
administration
 2000, Yantar Analitik
Computer Set-Up & Administration
Find
a suitable location
Connect to power and modem
Connect ancillary equipment
Load operating systems and programs
Prepare for years of continual administration
Computer administration is a widespread and repetitive task that
should be fully automated!
 2000, Yantar Analitik
Research Problem



Why isn’t the administration of
computers already fully
automated?
What hardware and software
characteristics hinder/assist full
automation?
Can computer administration
become fully automated?
 2000, Yantar Analitik
Research Objective
 Develop
an automated software system
for administering public computer
terminals
 Duplicate
administrative tasks and remove ALL
human intervention from the remote computer
 Make
clear the reasons why computer
administration is not currently fully
automated
 2000, Yantar Analitik
Importance


The focus impact:
 Reduce routine administrative cost
 Ease deployment of computers to admin-incapable
businesses and organizations
Numerous other benefits
 Simplify update of software and systems
 Redirect administrators to more strategic issues
 Prepare for distributed application computing
 2000, Yantar Analitik
Background and Research Review
[1992]Depot
& [1994]Soft.
[1998]Comparison of Large-Scale UNIX/NT Install.
Microsoft docs.


[1999]NT FAQs, NT Automated Install Tools.
[2000]Windows 2000 Automated Deployment.
[2000]Automated
Generic Operating System Install.
Altiris, Mirimar, ON Technology, Norton Systems.
The proposed research differs in that the focus is not on replication and is
not on incremental improvement.
 2000, Yantar Analitik
Research Plan

Divide and Conquer
Learn how to profile and transmit a hardware configuration
from a clean computer
Enumerate steps to automatically select and transmit an
OS based on the received hardware configuration
Identify steps to remotely load an OS and user
applications, automatically, and the the configuration of
software residence on the computer
Profile a remote testing procedure for verifying proper
function
Highlight security issues
 2000, Yantar Analitik
Delimitations





PC platform with no peripheral equipment
Linux operating system
One user application (Netscape Communicator)
Network connection will be via modem
Prototype system considered operational when:
 Operating system is loaded on remote computer
 Application is loaded and one update performed
 Functionality can be remotely verified
 2000, Yantar Analitik
Deliverables

A ‘scaled-down’ prototype system for remote
automated administration of computer terminals

A paper detailing characteristics and weaknesses of
current hardware platforms, operating systems and
software applications that hinder full automation of
computer terminal administration
 2000, Yantar Analitik
Feasibility


This research effort will produce valuable results even
without complete success in one of the deliverables
 If the prototype contains lots of workarounds, then
a strong hardware/software limitations document
should result
 If there aren’t many limitations, then the prototype
should be a strong product
Other feasibility considerations are researcher specific
 2000, Yantar Analitik
Timing
Jan-01
Hardware Profile
Select OS
Install OS/Apps
Remote Test
Security
Prototype Test
Limitations Paper
 2000, Yantar Analitik
Mar-01 May-01
Jul-01 Sep-01 Nov-01
Jan-02
Resources & Special Needs




A server and clean client
Linux source code
Netscape communicator
Availability to systems
administrators
 2000, Yantar Analitik
Conclusion




Computer administration should
be easy and automated.
There are specific reasons why it's
not automated!
A prototype automated system for
remote administration will be
developed
Hardware/software characteristics
limiting full automation will be
documented
 2000, Yantar Analitik
 2000, Yantar Analitik