IBM SmartCloud Application Performance Management: Shifting Application Performance into Higher Gear

Cloud computing poses new puzzles for application performance management

Service Management in Action

Composite application management is no simple feat—and as cloud momentum increases and cloud architectures are increasingly adopted, it's only getting more complex.

You can get a sense of the difficulty of optimizing composite applications just by considering the complex array of resources and systems they typically involve. If an application involves five systems, six databases, and various instances of runtime Java applications, how quickly can you pinpoint performance issues when they occur? Bottlenecks could stem from any number of potential root causes. And when different IT domains are managed by different managers or entire teams, pursuing cross-team collaboration to resolve application performance issues can itself be a challenging matter.

With the advent of private cloud architectures, things get more complex still. In theory, clouds leverage fluid resource allocation and dynamic virtual server creation automatically, via policies, to support the highest levels of application and service availability. In practice, the fact that new servers are continually being created and eliminated in a cloud, and resources are allocated in a manner difficult to predict in advance, often makes composite application management even more challenging than it is in conventional architectures.

Going beyond a technical perspective, there's the human implementation of problem solving to consider as well. It's important to be able to support performance management goals in a job-specific context—one that acknowledges the fact that different team members need different kinds of information, and presents that information in a clear, intuitive, and role-specific way in every case.

IBM's got smart, cloud-savvy APM solutions in the works

"Through the project dashboard on the Service Management Connect site, you can see the current release plan, download and try recent beta releases, and respond with your own thoughts and requests—and IBM Tivoli encourages you to do exactly that."

For these and related reasons, IBM is developing a new solution: IBM SmartCloud Application Performance Management. Building on the SmartCloud portfolio, which also includes provisioning, monitoring, and control desk elements, the new solution will provide enhanced visibility specifically into application performance per se—giving organizations the power to understand both when and why performance has degraded, and the information needed to improve matters in a business-prioritized fashion.

The new solution is also notable for another reason: it's one of the first new Tivoli solutions to be created transparently and collaboratively with customers. Through the project dashboard on the Service Management Connect site, you can see the current release plan, download and try recent beta releases, and respond with your own thoughts and requests—and IBM Tivoli encourages you to do exactly that.

By helping IBM better understand exactly what you need from a cloud-savvy application performance management solution, you'll be contributing to the business value you ultimately get from it—a strong incentive indeed.

Five different ways to assess application performance—and holistically manage applications for higher value

Wouldn't it be great if you could detect performance problems quickly enough to minimize—or even preclude—their business impact? The new offering promises just that via a range of linked, integrated capabilities.

They include:

  • Discovery: Accurately establishing the different elements that comprise the composite application, and their relationships to one another, even in the dynamic environment of the cloud, all without requiring manual processes or ongoing oversight.
  • End-user experience: Understanding how the application looks and works from the ultimate end-user's perspective, so that that experience can be optimized accordingly.
  • Transaction tracing: Stepping through the logical sequence by which the composite application works during an actual business transaction.
  • Diagnostics: Isolating performance bottlenecks/outages to root causes, and driving an accelerated response designed to minimize the business impact created as a result.
  • Analytics: Evaluating different circumstances/conditions that affect application performance in order to uncover hidden trends and patterns that can subsequently lead to technical optimization and a superior business outcome.

As just one example of the integrated possibilities, consider this scenario: analytics discover a problematic trend and generate an alert to managers, who then use diagnostics to trace its root cause, and discovery to assess other services/applications that might be affected.

Support for many kinds of applications and technical domains

How are these features implemented? The solution supports both agent and agentless technologies; both can draw information from different elements of the cloud as required, combining that information in a single, elegant topology view.

Also included: a template designed to support complex J2EE, and SAP applications with built-in industry best practices. And right out of the box, this offering will also provide recommended metrics for various application elements and protocols, such as HTTP clusters, WAS clusters, DB2 clusters, and SAP—all based on the specific type of application being delivered.

Together, these capabilities can simplify and clarify exactly what's happening as composite applications execute in a cloud—and allow organizations to capitalize better on the full promise and economic benefits of the cloud, which is after all supposed to be a largely autonomous, yet optimized, platform of IT service delivery. By giving organizations the information they need to keep composite applications running smoothly, the new solution will drive up the business value the cloud creates while driving down costs, and risks, of many kinds.

Also significant is the fact that the solution can present not just information, but information in a granular way—as 10 actionable insights, perhaps, instead of a thousand metrics. That means organizations can dive into the technical detail of performance problems at any given depth—if they need a great deal of detail for something like code analysis, that will be available, but if it's a simpler matter, such as the need to stop and restart a particular virtual server, that can be easily established.

New configurable Web 2.0-style dashboards: All the job-specific information you need, simply presented

The emphasis on easy is also apparent when you consider the way in which all this information is presented. IBM SmartCloud Application Performance Management will include brand new Web 2.0-style user dashboards designed to provide the most straightforward, intuitive possible depiction of key data and performance metrics. It's through these that managers can drill down into technical detail, discovering and researching performance issues quickly, easily, and cost-effectively.

These dashboards provide information on key topics like availability, performance, technical composition, utilization, and capacity. You can consider these three different lenses by which to view cloud-hosted composite applications. Each lens reveals a different aspect of application performance, and the dashboards allow you to choose the kind of insight you need—or combine them as required.

They also support role-customizable dashboards. The kinds of information a line of business manager needs to know concerning an application, for instance, are likely to be quite different from the kinds of information an IT manager or business executive needs to know.

One particularly important job-based distinction: IT operations vs. development. Because both groups are supported, both can leverage the new solution, and the performance data it yields, to pursue essential tasks.

So IT operations team members can check up-and-running cloud applications for performance issues, resolving problems when they occur; meanwhile, IT development team members can use the same solution to assess new applications prior to their release into production, minimizing problems that might occur under different conditions well in advance. The overall result is that applications can be released into production more quickly, and once there, will be less likely to have to revert back to an earlier version due to unacceptable issues discovered at that point.

Another attractive aspect of the new user interface is the fact that it is supported on a variety of different platforms, including mobile platforms, such as the iPad and iPhone. That means both smart phones and tablets can be used to oversee and improve application performance remotely—powerful, hands-on control whenever, and wherever, managers happen to need it.

Any type of IT application infrastructure is supported—and the solution will continue to work as that infrastructure evolves

And while the new solution is specifically targeted at private cloud architectures—as the name suggests—it's important to note that it also supports earlier, less sophisticated architectures just as well.

Whether your architecture is best described as traditional (a mix of distributed and/or mainframe systems), virtualized, a private cloud, or a hybrid cloud model—or any combination of the above—IBM SmartCloud Application Performance Management will support it, helping you understand and eliminate performance-threatening bottlenecks in every case.

That implies not just a broad range of compatibility with different organizational infrastructures, but also a degree of upgrade insurance. That is, if your infrastructure expands and evolves over time, the solution will continue to support it, despite the fact that the delta of change may be very extensive.

Download the beta, try the online demo, and contribute your own feedback to the development team

Interested in evaluating IBM SmartCloud Application Performance Management for yourself? As of April 3, the interface of the solution is in beta form, and can be downloaded and experienced at will. The most recent version includes three-tier application dashboards:

  • An application summary dashboard
  • A J2EE application dashboard composed of widgets that address HTTP, WAS, WMQ, and DB2
  • Component dashboards, also for HTTP, WAS, WMQ, and DB2

The first two can both be edited and saved, so you can get a sense of the solution's forthcoming customization capabilities.

And for a quick test drive of the new user interface, be sure to try out the Web-based demo IBM is hosting for free online—accessible from any standard browser, including those running on smart mobile devices.

As suggested earlier, IBM Tivoli is openly soliciting user feedback on this solution. The more information users contribute, the more complete and well-tailored the release version will be to those users' interests. Users are also encouraged to feel free to pass on this information to anyone else they think might be interested in optimizing application performance inside cloud architectures in the smartest, most intuitive way.

Additional information

Recent Articles

Contact IBM

Considering a purchase?

Innovate2012 - The premier event for software and systems innovation

Innovate 2012

Innovate 2012 is the software innovation conference focused on helping you transform software innovation and accelerate better business outcomes. Sign up for the Collaborative Development and Operations track to learn how to design, deliver and manage innovative services more quickly and effectively. Don't miss Innovate 2012, June 3-7, 2012.

Register today

Subscribe to the Service Management in Action RSS feed.

Help with subscribing