A recent survey of 150 IT Managers in large organizations conducted by Harris Interactive, and annotated by Mike Vizard in an IT Business Edge article, calls attention to some of the issues facing organizations that have “too many” applications.
Over 82% of respondents noted that they encountered monetary losses due to “unused, unresponsive, or crashed applications”, but only 32% of respondents said they plan to introduce an application monitoring tool in the next 18 months. I think those two numbers ought to be reversed: Less than “20%’ should be encountering losses, and ”80%” should be implementing application monitoring.
What To Do
As noted in the annotations of the article – you can’t manage what you can’t see, and somebody needs to make some hard decisions about what gets installed and maintained in the enterprise. Also consider this practical reality -- if the applications are not being used, or are not reliable in the form in which they’re installed – then it’s quite likely that nobody is producing useful work from them anyway. Get rid of them!
Here’s a list of things to get the ball rolling:
- Perform a complete inventory of the software that is currently installed.
- Identify which applications are not being used and uninstall them – permanently.
- Identify which applications can be replaced.
- Identify which applications need to be updated and get them updated. If you cannot update an application to address security issues or reliability issues, it’s time to get rid of it – permanently.
- Where multiple versions of the same application are in use, standardize on a specific version – it doesn’t have to be the newest – but pick one!.
Here’s another collection of things to consider:
- Identify which applications are dependent on other outdated or insecure software. For example, dependencies on old versions of Java or MDAC are not good. If there aren’t updated versions of the software that can be used securely and reliably, find a replacement as soon as possible.
- Identify which applications are dependent upon elevated privileges (i.e. they require Administrator rights).
- Identify which applications are dependent upon specific (read: ancient) operating systems. If the vendor isn’t going to update the application to run on Windows 7, get rid of the application and the vendor!
- Identify which applications can be shared or virtualized. Maintaining one copy of an application is much more efficient than maintaining several; in addition, reducing the number of instances in use may save licensing and/or maintenance costs.
In many cases, these situations are going to require a bit more effort to determine whether the vendor has a solution or has committed to delivering a solution, or whether the application will need to be replaced. But now is the time to start developing a response plan to issues that can be identified today.
For the applications that are selected to be kept, they need to be monitored. Monitoring includes a number of aspects, including:
- actual usage
- required updates
How to Get There From Here
Getting from where those 150 surveyed organizations are today to a lean, streamlined proactively monitored, healthy application environment is not going to happen overnight, so don’t try. The solutions can be phased in over time, as appropriate for the needs of the organization, but monitoring applications will require a number of tools.
- Application usage can be monitored through software inventory or event log monitoring software.
- Required updates should be monitored through an effective patch management solution.
- Reliability and performance are handled with application performance monitoring software.
- For virtualized applications you should make use of virtualization management software.