One of the hazards of being an Analyst is that it brings out your inner geek. In completing the latest research on “End-to-End Application Management in the Age of Cloud”, it was necessary for me to hole up with a gallon of coffee and a bag of chocolate-covered coffee beans and figure out what Cloud REALLY means to the application management world.
Totally wired and ten pounds heavier, I finally emerged with an updated version of EMA’s End-to-End Application Management Semantic Model. It includes a new Cloud Analytics layer, which is somewhat theoretical but will hopefully materialize relatively quickly as Cloud services evolve and mature.
Long story short: the vast majority of today’s Application Performance Management (APM) and monitoring tools were developed to be hosted on premise and to monitor/manage on-premise services. As enterprises move to the Cloud, what kinds of capabilities are necessary to monitor and troubleshoot this new “end-to-end” application environment?
Answering this question is particularly relevant in view of the findings in the associated survey. Forty-nine percent (49%) of companies surveyed are already hosting tiered applications that span on-premise and Cloud/SaaS. Thirty-five percent (35%) are integrating multiple SaaS services.
The bottom line? Enterprises have already deployed complex Cloud services. While Cloud in all its forms is mainstream, traditional monitoring, management, and troubleshooting tools may not have the visibility to Public Cloud necessary to perform these functions.
In the Public Cloud, visibility is at the whim of the Cloud provider. Although providers such as Amazon and Rackspace are offering APIs for their environments, they are primarily aimed at adminisration versus mangement. Some SaaS providers have created online dashboards that show red/green uptime and performance metrics at a high level. However many IT organizations are seeking to consolidate, versus proliferate, management dashboards, and few of these solutions integrate to APM dashboards used in production.
When it comes to tracking transactions end-to-end across on-premise and Cloud, there are not many options as yet. Without instrumentation, Public Cloud becomes a black hole, management-wise, and few Cloud monitoring options integrate to traditional APM toolsets as yet.
End User Experience (EUE) solutions offer an immediate answer by tracking overall performance and availability from the perspective of the user. However even EUE tools can’t penetrate the Cloud (unless it is instrumented by the provider). This makes it difficult to troubleshoot transactions executing across multiple Cloud platforms, and those deployed as integrated SaaS-to SaaS applications.
After reading this far, you can probably see why I needed massive infusions of caffeine and chocolate to think this through—it is a veritable can of worms. If you are up for the “long story long” (versus the long story short) version of this discussion, check out my latest free Advisory Note on the EMA site (located at: http://www.enterprisemanagement.com/research/asset.php/2071/Managing-Applications-in-the–Public-and-Private-Cloud). The paper is titled “Managing Applications in the Public and Private Cloud”, and is drawn from the longer research paper published earlier this month. It includes before and after graphics of the End-to-End Application Management Semantic Model, along with an explanation of how EMA sees Cloud monitoring and management evolving over time. Check it out when you’re ready for a geek-fest.