In my last column, I discussed the importance of Application Discovery and Dependency Mapping (ADDM) as a tremendously valuable resource for managing and optimizing application and business services, as well as the IT infrastructures, and supporting them, in the age of cloud, agile, and consumer-driven IT. In this column, I'm going to delve more closely into our evaluation criteria — as they applied to the radar, but also as they might apply to you in making an investment choice in ADDM.
However, first, I would like to point out two things that became apparent during the radar reviews with vendors — as appropriate for this discussion.
The first is that the ADDM market has become indeed dynamic — so much so that many of the vendors included were surprised at the positions, strengths and features of others. Some of the progress has been surprising even to me, as I've seen classically “hard-to-deploy” solutions in a few cases become standouts for efficiency and ease of administration. Functional enhancements have also thrown off some vendors looking over their shoulders at others. Perhaps all the more because as ADDM evolves, it often does so “under the covers” of solutions branded with other functions — such as CMDBs, service desks, or application performance solutions.
And this brings up the next point: in selecting an ADDM capability, it's very important to understand exactly what it is, and what it's not. ADDM, across all three use cases (Change Management and Change Impact Analysis, Service Impact and Performance Management, and Service-Aware Asset Management), is an enabler. It is, or should be, a dynamic system of relevance showing how and where application, infrastructure, assets and business values come together — how they are interdependent, and where.
Beyond that, an ADDM capability should be able to incorporate insights into configurations and changing conditions as appropriate for use case requirements. This is asking a lot, and integrations – especially with CMDBs as well as other capabilities for automation, monitoring and configuration management, etc. — are often key to ADDM design.
But ADDM is not necessarily in itself about pure play diagnostics, or software license management, or active configuration or “release management” functions that it ideally supports, but doesn't necessarily perform. Put bluntly, ADDM does not fully equate to application performance management, or configuration management, or asset management — even if it can substantially empower all three.
Having said this much—here are a few additional criteria that we looked at for our evaluations:
- Domain expertise required/skill sets, and credentials handling. These areas have been classic challenges for many ADDM solutions in the past but progress is being made, in some cases dramatically. Not surprisingly, performance-optimized ADDM tends to be more automated, with fewer skill sets or domain expertise required that multi-use-case ADDM. But EMA is seeing progress on all fronts, as vendors learn from both their competitors, and an increasingly impatient IT constituency.
- Opex resources or full-time employees (FTEs) associated with initial deployment, initial application coverage, more robust application coverage, and ongoing administration. Once again, there's a lot of progress on all fronts. After initial credential and other deployment issues have been decided, ongoing administration for your ADDM capability should not be a visible burden.
- On-premise, virtual appliance and SaaS has been another substantial area of growth, with six of our vendors offering SaaS today, and others planning to do so in the future. I should add that in the majority of cases, SaaS is combined with on-premise or, in some cases, appliance options.
- Agent and agentless is once again something to consider. These choices are not purely abstract, but may impact credential requirements, ease of deployment, and level of depth in capturing configuration and other information. EMA would say that everything being equal, having both options can often be a plus.
- Architectural strengths such as scalability, domain breadth and depth, support for public and private cloud, proven integrations — CMDB, automation, service desks, monitoring, analytics, asset or even other discovery tools — are all factors to be considered. EMA also wanted to know the bi-directional story — as ADDM informs and empowers other resources, while also drawing from them for its own enrichment.
- Use case specifics were generally taken in context with the above. Your investment in ADDM as a system of relevance should be very closely tied to its ability to work with and draw from your other investments on a per-use-case basis.
To conclude, I'd like to wrap up with a few quotes from EMA's ADDM Radar Report, which should be in our library within a week:
“Once again, the goal isn't to provide you with linear “winners” and “losers” — but to give you more meaningful insights into design, cost, functionality and administrative tradeoffs across a richly innovative and progressive set of vendors. This is true in many respects for larger, platform solutions just as it is for smaller new comers.”
“This radar has also been built around a premise — that ADDM in many respects provides the key to the future of service management as it begins to unite real-time service performance issues with more process-defined change, configuration and asset management values. Industry dynamics in cloud, including new evolving standards like Topology and Orchestration Specification for Cloud Applications (TOSCA), should actually play to and enrich these trends in the coming years. While the need for IT to have a better “model” of its services as they relate to business values and business outcomes will be yet another driver.”
“EMA looks forward to the time when it can once again do a radar with ADDM solutions even more united than today's in “performance-optimized” and “multi-use-case” — as ADDM service modeling becomes a yet more pervasive and fundamental enabler for optimizing IT services, and the infrastructure investments supporting them, to business results.”
Dennis Drogseth is VP at Enterprise Management Associates (EMA).