Here is an interesting disparity for you to think about. If you have invested a reasonable amount in network and systems management tools, your IT department should be able to:
- Keep your servers and networks up and running
- Monitor on an ongoing basis how utilized they are
- Use change management software and processes to manage keeping them up to date
- Use capacity planning tools to understand when to upgrade them
- Commit to service levels in terms of the availability and capacity of these resources
On the other hand, you are highly unlikely to have an analogous set of capabilities for your applications. For example, you probably:
- Cannot keep your applications up and running as smoothly as your servers and networks
- Cannot know on an ongoing basis how utilized an application is relative to its capacity to support users and work
- Do not know what the performance (response time) the application is delivering to each of your end users
- Cannot quickly and effectively identify problems in the applications, resolve them, and get the changes into production
- Cannot commit to application service levels that credibly promise availability and performance levels of the applications to your users.
There are three groups of vendors who collectively stand some chance of addressing this set of issues:
- Traditional infrastructure monitoring vendors (BMC, CA, IBM/Tivoli, NetIQ, HP, etc.). While these vendors products are very widely installed in enterprise IT shops they have neither a robust ability to measure applications availability and performance, nor any ability to assist in the application problem resolution process. These vendors are therefore the least likely of the three groups to be the source of a meaningful solution to the problem.
- APM focused monitoring vendors (Wily, Veritas, Mercury, Compuware, Quest, VIEO, ProactiveNet, eG Innovations, Premitech, etc.). These vendors all focus heavily upon measuring applications performance (response time), and offer in some cases (J2EE) a reasonable ability to pinpoint problems in middle tier applications. However, none of these vendors (as of yet) combines a comprehensive ability to monitor a variety of different applications architectures (J2EE, COM+, .Net, ASP, Citrix, etc.) for end user performance with an ability to help the problem resolution process once an application problem is identified.
- Problem resolution vendors (Identify Software, Mercury, Compuware, AviCode, XtremeSoft, etc.) offer a variety of capabilities to drill into specific source code environments and either pinpoint the problem or at least get the developer close to the source of the problem. However, these products are for the most part difficult to use in a "monitor all of the time mode" since they usually consume to much resource and generate too much data to be left on all of the time on every production server.
In summary the market opportunity for management of applications while they are in production is wide open. In order to seize this opportunity, someone will meet the requirements below and seize the brass ring:
- Monitor every aspect of the application no matter how it is architected (J2EE, .Net, etc.) and no matter where it runs.
- Collect meaningful performance data (response time) about every user of every application
- Operate in a manner consistent with the requirements of products that are running all of the time on all production servers (low resource utilization, and smart management of volumes of data produced)
- Automatically map the application components across the application system for each user of the application.
- Synthesize 1-4 above into true root cause reports that tell administrators and developers where the problems are and what to fix.
Someone will grab the brass ring and run off with a huge market opportunity. Who will it be?
Bernd Harzog
CEO
APM Experts
Bernd.harzog@APMExperts.com
