Novell Home

Saving Time with App Launcher: One Guy's Story

Novell Cool Solutions: Feature
By Mark Poole

Digg This - Slashdot This

Posted: 28 Jul 1999
 

We recently rolled out a product called EMPAC, which is an enterprise class client-server application used by our facilities management department. Without App Launcher (aka NAL), this project would have taken much longer, and some of what we did would effectively have been impossible. This is a comparison of how things went on the major project tasks, and how they would have gone without App Launcher. We were using ZEN 1.1.

Task: Provide Oracle client
Task: Test app object
Task: Install first wave
Task: Install second wave
Task: Move to production database
Task: Provide ongoing maintenance
Total impact of App Launcher on project


TASK: Provide Oracle client on local machines

Without App Launcher:

The program requires the Oracle client files to be on the local machine. We have already built our systems around running the client files from our Novell servers. We were told that we would need to change that. Basically, this would have required us to completely revisit all of our machines, retest every database application that we use, and start maintaining Oracle locally. Obviously this was not an acceptable solution.

With App Launcher:

App Launcher allowed us to dynamically map drives. This, combined with a few registry entries, allowed us to keep using Oracle the way we were, and not affect any other applications.

Savings:

I don't even want to think about what it would have cost us in testing, installation, and future support. Let's just accept that this would have been huge.

TASK: Test to verify app object functionality

Without App Launcher:

We spent about a week testing the client to make sure that everything was working right and that we weren't going to see any problems.

With App Launcher:

This step actually took longer with App Launcher. After we built the application object, we had to spend about two hours testing it, and making sure everything worked ok. We also used macros to define which Oracle instance that application would connect to. This was done to make a couple of future steps easier.

Savings:

App Launcher cost us two hours at this step. This is where a lot of people are afraid of the extra overhead in using products like ZEN. It will pay off later however.

TASK: Install first wave of users on test database

Without App Launcher:

When we got ready to roll the product out to the first groups of people, we were using a test database so that people could make sure that everything was working ok before we let them loose in the production area. Our first wave included 71 clients. Considering that each would have had to have the correct Oracle client installed, followed by the application client, followed by having everything configured, an hour is the absolute least time it would have taken to do the install. That would have been after training the installers, and doesn't include the time spent tracking the clients down and scheduling the installs, or walking to the end users' offices. (These are major problems in the university setting.) For us that would add on at least 20 minutes per install. So, based on 1:20 per install, this step would have taken about 94 hours.

With App Launcher:

It took me about one hour to add each of the names to a group which gave them the App Launcher icon. This completely avoided training, contacting the end users, and travel time. Total time spent: 1 hour.

Savings:

  • No App Launcher: 94 hours
  • App Launcher: 1 hour
  • Difference: 93 hours

TASK: Install second wave of users on test database

Without App Launcher:

At the beginning of the following week, we added an additional 95 users over a period of three days. Without App Launcher, that would have taken about 126 hours. If we wanted to get done in the three days, we would have had to put six people on it full time. This is assuming we could contact every client in those three days, and complete the installs.

With App Launcher:

This took about an hour and a half.

Savings:

  • No App Launcher: 126 hours
  • App Launcher: 1.5 hours
  • Difference: 124.5 hours

TASK: Switch everyone to the production database

Without App Launcher:

A few days later, we decided we were ready to switch everybody over to the production database. There are two ways we could do this. We could go to each machine and change where the application pointed (which is fairly simple), or we could go to each machine and change the TNSNAMES file to point to a different database (which is not as simple). I would estimate the time to make the change about five minutes. Since this is really easy, we will add on training at no time cost, but we still have to worry about the 20 minute overhead.

With App Launcher:

Since we had set up the database name as a macro, all we had to do was change that and increment the application object version number. This took less then ten minutes to do and test.

Savings:

  • No App Launcher: 69 hours
  • App Launcher: 10 minutes
  • Difference: 68 hours 50 minutes

TASK: Provide ongoing maintenance

Without App Launcher:

At times, we will have to do maintenance on the database. This will require that most people not be allowed to log in, while a few administrators do their work. In the past, we have tried just asking the end users not to log in during maintenance with little success. We couldn't think of a good way to do this without App Launcher.

With App Launcher:

We can create a schedule that removes the App Launcher icon when we need to be doing the maintenance. Since each of our administrators has a fully local install from the testing phase, this won't affect them and they will be able to do what they want. It takes about ten minutes to set this up.

Projected Savings:

  • No App Launcher: Unknown
  • App Launcher: 10 minutes
  • Difference: Unknown

Total impact of App Launcher on Project

App Launcher allowed us to handle the roll out of this project in under a week. It has also given us the flexibility to easily alter, disable, or reinstall the application as needed. These numbers don't include the amount of time that each of the end users would be without their computer. Based on my estimates, App Launcher cost us about eleven hours of end users time, about four minutes per person. Without App Launcher, it would have cost us about 180 hours of lost time, about 1:05 per person. So, our end users saved 169 hours. This gives the university a total savings of 453 hours. Add to this the fact that each additional App Launcher user we add to the system gives us another savings of two hours and thirteen minutes, and this program proves its worth. None of these times take into account the amount that we saved from not having to redo our Oracle infrastructure, or from having to figure out a simple way to keep the end users out of the database when they shouldn't be using it.

There were a couple of issues that came up during this project. First, although we have yet to have a single client install fail, we have had several machines that didn't meet the minimum requirements. Because the people didn't see the icon when they were supposed to, they contacted tech support and we had to figure out what the problem was. What would be great is the ability to display the icon, but have it provide information about which of the minimum requirements isn't being met. The other feature request that came out of this was the ability to schedule down time for an application object as opposed to when it is available. This is a small difference, but it would make configuring the application object for maintenance a little easier.

[Editor's Note: These two issues have been addressed in ZENworks 2. ZENworks 2 does offer the ability to show an application even when a system requirement is not met. The application is disabled but the user gets information telling them why it is not available. And as far as scheduling is concerned, you can control the availability of the application through the application schedule. Therefore you can schedule down time according to a pre-set schedule.]

Total Savings:

  • No App Launcher: 289 hours
  • App Launcher: 4 hours 50 minutes
  • Savings: 284 hours

About the Author

Mark Poole is a ZEN power user from James Madison University in Harrisonburg, Virginia. He is a frequent, and much appreciated, contributor to Cool Solutions, and can be found at odd hours of the day and night solving people's ZEN problems in the Support Forums.


Novell Cool Solutions (corporate web communities) are produced by WebWise Solutions. www.webwiseone.com

© 2014 Novell