Q&A: Application Performance Management
Enterprise Systems recently talked with Compuware’s John Van Siclen about one of the fastest growing segments of the IT industry -- application performance management.
Why is application performance management (APM) getting so much attention from IT, and what metrics should IT be tracking? To learn more about APM, we turned to John Van Siclen, general manager of Compuware's application performance management (APM) business.
Enterprise Strategies: What's behind the interest in application performance management?
John Van Siclen: In leading enterprises, there is a growing realization that applications are the business. Applications are what end users -- be they customers, employees, or partners -- interact with. Executives are finally understanding the direct impact that sub-optimal performing applications have on their businesses.
Additionally, applications are now delivered through an increasingly complex application delivery chain that has become mobile and distributed, reliant on third parties, Content Delivery Networks, the cloud and increasingly difficult to analyze.
This is fundamentally why APM has become crucial throughout organizations, often from IT leadership right up to the C-suite. APM is now the principal IT discipline that enables organizations to assure the availability and optimize the performance and value of business-critical applications.
Gartner estimates that at present, 20 percent of the global 2000 are trying to reconstruct the whole of their IT operational process frameworks in a way that accords the monitoring and management of applications, rather than infrastructure, a central place. We believe this is because speedy, properly functioning applications directly relate to bottom-line revenue, loyalty, and overall operational productivity. Moreover, in its latest “Magic Quadrant for Application Performance Monitoring,” Gartner also cites that APM “has become a central focus for IT operations.”
Can you explain the differences between “traditional APM” and “new generation APM”?
A profound change is taking place in the APM market. Older, traditional APM tools no longer offer the capabilities needed to address today's ever-widening IT ecosystems.
Traditional APM tools focus on component health as well as reactive alerting after a problem has occurred. They can't see across the full application delivery chain to identify performance issues and root causes proactively (when they can be fixed before user impact). Also, traditional APM tools focus on monitoring at the data center which makes them blind to user interactivity and third-party services, which make up a growing amount of end-user activity in modern applications.
In contrast, new generation APM systems start their monitoring as close to the end user as possible, in the browser itself, so that availability and performance is exactly as the user experiences it. Perhaps most challenging for IT is that these legacy APM tools are very difficult to use and must be tweaked every time an application or the application environment changes. Modern APM is built to auto-discover applications, auto-adapt to change, and auto-baseline for very fast time-to-value and low cost of ownership over time.
The new generation of APM is also business savvy, giving executives real-time performance and availability, visibility with dashboards that show how application performance relates directly to business impact -- conversion rates, abandonment rates, revenue, and so on. To address today’s ever-growing IT complexity, companies need to move away from a piecemeal, tools-based approach to a more integrated approach that offers a complete view of performance across the entire application delivery chain from the user’s click or gesture on the screen through the cloud, across server tiers to data sources of record and back.
What exactly should IT be measuring? What are the most useful metrics?
Businesses need to adapt to the realities of today’s marketplace. End users are increasingly impatient and demand application performance at lightning speed. This applies to consumers as well as business users where faster time-to-value is critical. As a result, APM driven by insight into the end-user experience is a must-have.
Business transaction management is another essential area that can provide insight to companies. Business owners are requesting more than just uptime stats and page views to make informed decisions. They want to know if their applications are meeting service-level targets for each customer, how performance is affecting revenue, and how marketing campaigns are affecting usage.
Modern APM gives business owners this insight needed to make informed decisions, and delivers IT the transaction-pure detail -- all the way down to an offending line of code at the root of a performance issue -- which they need in order to optimize application performance and assure stability and scalability.
What has changed?
I could talk for hours about how new generation APM differs from legacy tools that previously defined the APM space, but let me give you some key differences.
First, APM now delivers unified insight across the entire application delivery chain. With this visibility, organizations can measure app performance from user to cloud to data center to the mainframe and back again. The system can integrate every essential form of APM (real user, synthetic, multi-tier, deep app monitoring, and transaction tracing) as well as advanced analytics for all applications, including Web, non-Web, mobile, streaming, cloud, homegrown, and third-party apps. It’s hard to manage what you can’t see -- but now IT can have visibility into everything without having to go to dozens of tools and try to humanly correlate the disparate data.
Second, deep 24/7 application visibility proactively prevents issues from impacting users. New generation APM can now trace and capture every production transaction 24/7 -- with less than 2 percent overhead (traditional APM was well known for 20 to 30 percent overhead just to run basic statistical data capture). Following the path of the user captures all technical details and their business context changes everything. Issues can now be proactively identified by the system, with one-click down to the line of code to identify root causes and enable immediate resolution. Accelerating mean-time-to-repair (MTTR) in this way cuts weeks to months off old methods of problem reproduction then resolution.
Third, APM can now provide unprecedented visibility into cloud environments, enabling organizations to leverage the cloud with new confidence. Whether public, private, or hybrid, by following the actual path of the application not only can you identify and manage issues in ways not possible before, you can also measure cloud services to ensure performance and service-level agreement (SLA) compliance. Visibility in the cloud is essential for organizations deploying both standard and big data solutions.
Fourth, mobile is changing everything and is a driving force behind “complexity at the edge.” Users are accessing applications from a myriad of Web browsers, mobile browsers, mobile sites, and native applications from an unprecedented number of devices. Along with the mobile explosion has come a change in end users’ mobile experience expectations. Mobile users may be willing to trade some functionality for the convenience of mobility but won’t sacrifice speed or availability. Modern APM addresses this and optimizes mobile performance to keep businesses competitive.
What best practices can you suggest to measure performance?
The overriding best practice is to focus on the end user. This is the view through which the top companies manage and look at application performance. The application is what the end user interacts with, so companies should monitor the entire path of an application -- from the edge of the Internet through the cloud to the data center all the way back to the data source of record, including the mainframe if applicable. A unified APM strategy may sound daunting, but with modern APM systems, it’s very fast and easy thanks to embedded analytics intuitive interactivity design.
Another best practice organizations are adopting is utilizing the APM platform to support the entire application life-cycle -- development, test, and production. By using a life cycle approach, redundant, time-consuming tasks are automated and business processes across traditional silos can be significantly improved. This means faster applications, faster time-to-market for new apps or enhanced apps, elimination of issues before they become problems in production, and reduced, of course, dramatically faster MTTR. The business value of a life-cycle approach can’t be underestimated.
IT budgets are tight, so justifying a new APM solution might be difficult. How can you tie an APM system directly to an enterprise's bottom line?
Actually, justifying investment in a new generation APM solutions is easy. Businesses are continuing to make large investments in business-critical applications and the environments upon which they run; investing in an APM system to better manage and optimize these business applications makes sense: it drives value, lowers TCO, and maximizes productivity and profitability. With such a fast time-to-value and high multi-dimensional business impact, it’s not a difficult financial exercise to justify the spend and companies are beginning to understand this. Right now IT organizations are growing their budgeting for APM at three times the rate of overall IT spending.
Let me give you some tangible examples. A decade ago, we were content to wait 10 seconds or more for an application or Web page to load. The industry standard is now under two seconds. For every second of wait time beyond that, the business is impacted and revenue is lost. Although dollar impact varies from business to business, The Aberdeen Group estimates an e-commerce Web site can lose seven percent of sales conversions for every added second of wait time. Performance matters.
Financial benefits of APM go far beyond Web performance. ROI can be found in virtually any business function that is app-driven. Slowdowns and downtime in ERP systems, for example, can lead to hundreds of hours in lost productivity and sales -- a problem one of our Fortune 500 customers had before investing in new generation APM.
Significant ROI can be found in legacy mainframe environments as well. New generation APM is now being used for optimizing the interaction between distributed applications and the mainframe to reduce resource consumption and MIPS. Lower resource and MIPS consumption helps organizations delay or avoid costly CPU upgrades, eliminating both hardware and software licensing fee increases that accompany mainframe upgrades.
What products or services does Compuware offer for APM?
Compuware APM optimizes the performance of Web, non-Web, mobile, streaming, and cloud applications. Our new generation of application performance management is driven by end-user experience and provides unified coverage of the application delivery chain -- from the user through the cloud to the data center into the mainframe and back -- with deep, 24x7 application visibility by tracing every transaction in production from user click to code line. We have a comprehensive, life-cycle-oriented APM cloud solution and provide APM for mobile Web and native applications.