Defining a `lights-out' environment requires making choices. We help you determine just what those choices are

February  1997

What exactly is a lights-out environment? We compare what you might want in regards to scheduling, security, and software distribution and in the realms of management -- change, configuration, event, network, problem, and storage -- to what is possible now. (700 words)

Mail this article to a friend

You need to automate...we need a lights-out environment. That's easier said than done. How many times have you heard this before from management? "I want you to implement a lights-out production environment, we need to reduce the cost of our operations environment." It roles off their tongues so easily, no big deal right?

Get a clue in reality. We've been in IT for a combined 50 years now and we've probably heard it a few dozen times from the management we've reported to over the years. Even with all our travels around the world (several times), we have yet to see such an environment.

So let's try our own definition of what a "lights-out" environment should look like. In the following chart, we visit each discipline and summarize two objectives:

1. Nice to have: Summarizes the ultimate in automation.
2. Reality: Just what it says -- this is what's available now.

What "Lights Out" Really Means
You need	Nice to have	Reality
Change Management for online applications	Receives requests for change Determines potential impacts Tracks down and verifies approver Publishes change control notification Verifies status of changes and publishes	Routes request to approver Receives manual input for approval notification Publishes change control notification Prompts personnel responsible for making the change for status— once manual input is received, tool publishes changes.
Configuration Management/Release Distribution	Automated system detection on network to: Detect configuration of new system and/or device Log and report changes Global change to accept system from all monitoring processes on the network. Granting appropriate authority to new system Back out from a change if deemed unsuccessful	Track the configuration and "package" of objects to be changed Back out from a release or change
Event Management	Actively monitor the system (network, database, application, etc.) Automatically react and resolve problem using an information database.	Monitor the system (network, database, application, etc.) Automatically react to problem by notifying the proper support person without human intervention Human intervention to resolve problems
Network Management	Capture resource utilization (system, disk, network, bandwidth CPU, I/O, database transaction) real time Locate the bottleneck or problems Automatically resolve or tune with no human intervention	Capture resource utilization Notify support when threshold exceeds pre- defined limit, by way of the event monitoring system Provide information or hints for further analysis Procedure to address global performance trouble shooting
Problem management	Automatic problem detection anywhere on the system React and resolve with no human intervention	Log, track, escalate (auto paging, etc.) and report problems and their resolution Information support database
Scheduling	Detect new application batch processes and schedule Schedule jobs to run at the "best" time Check for dependencies Restart if failed processes	Schedule jobs Check for dependencies Restart if processes fail
Security	An integrated online application to: Handle request, approval, technical administration, and auditing for user access processing to the network database application and transaction levels. Detect, configure and authorize access privileges for new Applications introduced to the network, system, and database Automatic detection and reporting for security breach of Network, system, database, and transaction levels	An online application to process request, and approval routing, providing auditability, for user access to the system, database, and transaction levels Tools and streamlined process to administer access Individual tool to audit and secure network, system and database Interface event monitoring process when there is intrusion
Software Distribution	Automatically distribute software products from the server to the desktop, automating license requirements, software releases, and client architecture	Standardize on 3-tiered architecture to assure: Centralized administration for bug fixes, patches, upgrades Reduced requirements for system administration Consistency of software Reliability and totality of back-up processes
Storage Management	Backups start at scheduled time on system labeled tapes Robotic tape loading and archiving with built in cataloging Physically organizes tapes for third party vendor pick up and delivery Restore data files for users automatically For user requested restore, system automatically contacts third-party vendor for return of tapes	Reliable back ups start at scheduled time Robotic tape loading and auto cataloging systems Automatic completion status sent to control and display layer

Now it's time to put "lights-out" out for good. From now on, we'll refer to it as a "world-class" production environment.

If you have technical problems with this magazine, contact webmaster@sunworld.com

URL: http://www.sunworld.com/swol-02-1997/swol-02-unix.html
Last modified: