Did you ever hear the term “a dog’s breakfast”? It is a slightly old-fashioned description that the folks over at Princeton University’s WordNet define as “a mess”.And if you’re wondering what a dog’s breakfast has to do with Application Performance Management, check out Figure 1.
In EMA’s latest APM research, survey respondents were asked to indicate which types of applications their IT organizations were delivering as production services and to “check all that apply”. The results reveal an astounding mix of application and technology types. Looking at these numbers from the APM perspective, it’s difficult to imagine how IT organizations keep this mix of diverse services up and running at all.
The average company is utilizing a multiplicity of technologies and architectures to deliver production services. Distributed systems, virtual environments, private cloud, and massively integrated eCommerce applications top the list, although the adoption rates of the various public cloud options grow each year as well. And since performance management assessments rely on metrics gathered from multiple vantage points across the application ecosystem, you have to ask yourself: “How are IT organizations gathering the metrics they need to track performance across these diverse platforms?”
The answer is simple. Many aren’t.
Traditionally, silo managers tapped the budget to buy tools supporting their own silos. Silo tools are accepted as a cost of doing business since networks and databases can’t easily be managed without them.
The process is not as clear cut with application-focused tools. Applications run ACROSS silos and APM buyers tend to have broader, more cross-functional responsibilities than silo buyers. In almost 60% of companies, Directors, C and VP level executives are directly involved in the process. As with decision-making regarding similar cross-functional industry initiatives such as IT Service Management (ITSM) and Service Oriented Architecture (SOA) adoption, APM activities are difficult to fund and carry out without executive leadership. There is an element of self-interest at the executive level as well. Almost every company has applications so business critical that failures have visible and immediate business impact. As a result, when essential services fail, executives more frequently get the call than has been true in the past.
Nevertheless, as Table 1 (below) shows, a surprising number of IT organizations are still on the sidelines in terms of APM investments. Only 35% have actually invested in application-focused Platform solutions; an equal percentage is relying on Change Management tools to identify “what changed?” Perhaps most telling is the fact that there are only a few percentage points separating the adoption rates of Application Management Platforms and that old standby “Homegrown tools”.
|#1||Application Management Enterprise Platform||35%|
|#3||Homegrown Monitoring Tools||32%|
|#6||Application Aware Networking products||19%|
|#7||Network-focused Real User Monitoring (RUM)||19%|
Table 1: Which are the top three (3) application-related management technologies and products that your company currently uses to manage on-premise applications?
Considering the range of technologies and applications shown in Figure 1 in the context of the tools adoption chart shown in Table 1— it becomes clear that the Application Management function is abysmally under-automated in the majority of IT organizations.
The reasons for NOT investing in commercial APM tools, other than cost, tend to be time to value, administrative overhead, lack of in-house skills and a steep learning curve. And to be fair, application management is a complex undertaking. IT specialists staffing Application Management teams must have an understanding of networks, code, databases, middleware, and a host of additional technology components necessary to support execution of complex applications. Tools need similar “skills”; enterprise APM platforms need insight into both silo-focused and application/transaction-focused functions to support both performance assessment and root cause analysis. Because of the high level of application domain expertise which needs to be built in to such tools, they are more expensive to build on the vendor side as well.
There is a happy ending to this story however. SaaS-based Application Management solutions are coming into their own and can eliminate many of the objections IT organizations have had to the APM solutions of the past. In fact, IBM has now released its full APM line “as a service” as part of its SaaS and IT Analytics portfolio (see IBM’s Ben Stern’s blog entry on the latest announcements here: http://bit.ly/1mxtDEO).
These are not watered-down, lightweight APM solutions. With a combination of “behind the firewall” monitoring code communicating with “in the cloud” analytics and reporting capabilities, these are robust, full-blown APM solutions engineered to scale. SaaS-based management tools are emerging as an elegant solution to a long-time challenge. For small companies they can be budget-friendly, while still being full-featured enough for large companies.
Companies of all types and sizes often prefer to move to a “pay as you go” model. In fact, the majority of companies making new tools decisions tell EMA that cloud hosting is now a leading consideration.
Better yet, SaaS-based APM vendors continue to add features and functions. As Ben’s blog indicates, IBM recently added support for Python, PHP and PostgreSQL to its Application Performance Monitoring capabilities.
Readers interested in learning more about the enterprise management industry’s march to the cloud will also enjoy IBM’s Enrico Vannini’s excellent article entitled “The top three challenges with software as a service (SaaS)”.
Any way you look at it, APM as a Service can be a simple, painless way to take advantage of all the benefits world-class APM solutions bring to the table.