The Rise of Continuous Application Performance Management

Michael Vizard

Everywhere we turn these days, the application architecture is getting more complex. Not only do applications now run embedded in browsers using code based on AJAX, those applications are communicating with multiple tiers of applications that extend from servers all the way up into various cloud computing platforms. And just for good measure, we've added a layer of virtualization that all too often blinds us to what is happening at the physical layer of the infrastructure.

Once upon a time we could count on essentially sampling these applications at various specified moments to get a picture of the overall performance environment. But today the business notices every application glitch. Unfortunately, trying to recreate various application anomalies can drive even the most experienced IT organization to distraction. So the issue seems to be that IT organizations need more sophisticated application performance management tools that are capable to tracking application interactions in real time across various tiers of application infrastructure.

A good example of a company trying to reinvent the way we think about application performance management is dynaTrace. Founded by a team that cut their teeth at Wily Technology, which was acquired by CA, dynaTrace is trying to pioneer the concept of continuous application performance management that relies on what company officials describe as 'bottom-up approach.' What that translates to in practical reality is that dynaTrace has found a way to continuously monitor all application events using agent software distributed throughout the enterprise. Those pieces of agent software then send small bits of information about the status of the overall environment back to a central PurePath console, pictured here:

The benefit of this approach is that the IT organization doesn't have to go to the trouble of trying to recreate an application anomaly.  All the required data is already captured in the system. This also substantially reduces the amount of time spent on 'blamestorming' because it becomes immediately apparent where the problem is with any given application.

To back up its claims, dynaTrace recently moved to give away a free AJAX edition of its software. More than a few IT organizations are having trouble with AJAX applications as they continue to scale in the enterprise. dynaTrace is hoping that this move with entice customers to buy the commercial versions of its software that are highly tuned for Java and Micosoft.Net applications.

Whether IT organizations adopt dynaTrace or some other new approach to application performance management, the reality is that mounting business pressures are creating a pressing need for IT organizations to resolve application issues quickly. And frequently, that will mean starting over when it comes to application performance management.



Add Comment      Leave a comment on this blog post
Dec 4, 2009 11:12 AM Kobi Korsah Kobi Korsah  says:
It is true that e-customers interact with increasingly sophisticated services; unwittingly reaching through powerful and increasingly complex web applications into the rich layers of IT sediment. Without an early warning system alerting the organisation to myriad customer interaction problems the words �tip�, �iceberg�, and �Titanic� spring to mind. In my view it is even more vital to be able to gain and then seamlessly correlate a real understanding of the business impact of ALL aspects of real user experiences with a deeper visibility into ALL aspects of that evolved environment � even as that experience takes place. So as the focus on retaining customers and recruiting new ones sharpens - continually charting the full customer experience alongside a deep and meaningful understanding of web application performance will be vital to success. Explore this subject www.ca.com/apm. Reply
Jan 5, 2010 7:01 PM Greg Howard Greg Howard  says:
Another key point in rethinking APM is ensuring that the tool works as well in production as it does in development. It should add only a matter of a few percentage points to the overall overhead; otherwise, it will be an excellent developer's tool but fail the IT operations and infrastructure professionals whose "beeper goes off" when the application fails in production. Reply
Sep 12, 2012 3:17 AM William Louth William Louth  says:
Sorry but dynaTrace is a monitoring solution not a management solution. Once humans are involved it is not truly "continuous" at least when we are talking about the performance of software and in particular deployments in the cloud. And the only management solution that Ops do today when it comes to applications is kill the runtime because they simply have not got the same control mechanism they have at the network. http://www.jinspired.com/site/the-complexity-and-challenges-of-it-management-within-the-cloud trace diagnostics have not truly delivered and less so when we consider the scale and rate of change in the cloud (environment) and the applications (codebase). We need to rethink this approach now. http://www.jinspired.com/site/jxinsight-simz-time-space-data-in-application-performance-monitoring Reply

Post a comment

 

 

 

 


(Maximum characters: 1200). You have 1200 characters left.

 

null
null

 

Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.