Distributed systems and applications running across a mixture of dedicated servers, virtualized components, and cloud services is the new normal. This growing fragmentation of the enterprise IT environment is creating new application performance monitoring challenges. Each service provider creates their own reporting. Application vendors have yet another. Internal IT departments have custom metrics that must be collected. These issues have not even addressed the need to actual manage the applications. If there is no visibility on end-to-end performance, how can there be management? To rephrase the common moniker, how can you manage what you cannot monitor?
Application performance management (APM) is a service within information technology systems management centering on monitoring and managing the performance and availability of applications. Application performance management systems monitor and report on transactions and information delivery by an application in a defined infrastructure.
APM traditionally has included several functional areas related to performance, including end-user experience reporting, application runtime discovery and monitoring, transaction profiling, component-level monitoring inside specific applications and data analytics.
Hybrid IT is the result of combining internal and external services, usually from a combination of internal and public clouds as well as physical, dedicated datacenters, in support of a business outcome per Gartner. A recent study found that 45% of companies surveyed had at least one “hybrid Cloud” application and the estimated deployment of hybrid Cloud applications could be as high as 70%.
Hybrid IT relies on new technologies and approaches towards traditional management to connect clouds, sophisticated approaches to data classification and identity, and service oriented architecture.
As a result the role of IT and its practitioners is undergoing significant change which includes deploying and supporting applications, services and processes that cross each of these traditionally “siloed” elements.
Hybrid IT will now be the “status quo” for IT deployments. Application and infrastructure providers alike are starting to recognize that there will be no single “IT department” for even a small or mid-size enterprise
The Hybrid APM Opportunity
IT Service providers – traditional and cloud providers alike – have an opportunity to create new offerings to address the new hybrid IT delivery and operational model. However it is not as simple as deploying the same tools under the same rules.
Legacy APM was developed and is traditionally deployed to mirror legacy enterprise software: siloed, dedicated infrastructure, mostly centralized to one datacenter, which translates to commonly defined frameworks for compute environments, network parameters, routing tables, and firewall rules.
Deploying Hybrid APM requires a deep understanding of applications and infrastructure at each installation and touch point, how managed service providers work together (and against each other) to provide solutions.
Additionally it requires being able to implement tools that not only monitor and report but give flexibility and automation into management. Some of the new, emerging use cases related to Hybrid APM include:
- Monitoring and reporting on storage application services and storage systems platform OS’s to ensure agreed to availability of data regardless of geography.
- Inclusion as a service for PaaS to report on end-user and runtime experiences for users and developers. This use of APM can be used for incident management and also predictive analysis of how a specific application would perform after made generally available (GA).
- Monitoring of ERP systems spread across multiple data centers and clouds to ensure application performance and report on bottlenecks.
- APM rules can be constructed to expand capacity, reroute application data and network traffic based on prebuilt and created models.
- HRaaS applications that integrate to legacy systems such as payroll or benefits can be managed through a “single pane of glass.”
Getting Started with Hybrid APM
Initial definitions should be built based on current state. This should include functions the businesses are looking for in order to provide services to the enterprise and the customers. Divided into applications and infrastructure they should include end-user experiences, run-time architecture, transaction management, and system monitoring.
Next, an integration framework of tools and APIs must be developed to support performance requirements and key performance indicators (KPIs). To deliver a unified dashboard reporting and management requirements should be documented and implemented.
Finally, the operational model should be implemented which allow for continued service improvement and insure the ability to report and resolve incidents. This can and should be proactive as well as reactive management, financial and business operations, as well as capacity planning and management are necessary to have a true hybrid APM.
None of these steps are for the faint of heart and require expertise outside the traditional IT delivery and operations team. Working with all of the vendors involved to drive delivery of this level of complexity requires new IT skills. But once completely deployed the CIO will have a fully integrated view of the enterprise which not only reports on performance but allows him/her to drive financial and operational control like never before.