With the proliferation of composite applications for cloud and mobility, monitoring individual components of the application delivery chain is no longer an effective way to assure user experience. IT organizations must evolve toward a holistic, more collaborative methodology based on a service-delivery principle that is more aligned with corporate strategy.
The more business processes come to depend on multiple applications and the underlying infrastructure, the more susceptible they are to performance degradation. Unfortunately, most enterprises still monitor and manage user experience from traditional technology domain silos, such as server, network, application, operating system or security. As computing and processes continue to shift from legacy architecture, this approach only perpetuates an ineffective, costly and politically-charged environment.
Key drivers necessitating change include widespread adoption of virtualization technologies and associated virtual machine (VM) migration, cloud balancing between public, hybrid and private cloud environments, and the traffic explosion of latency-sensitive applications such as streaming video and voice-over-IP (VoIP).
The migration toward IaaS providers such as Amazon, Google and Microsoft underscore the need for holistic user experience assurance across multiple data centers, which are increasingly beyond the corporate firewall.
Moreover, as video joins VoIP as a primary traffic generator competing for bandwidth on enterprise networks, users and upper management will become increasingly intolerant of poor performance.
By having different tools for monitoring data, VoIP and video traffic, enterprise IT silos experience rising cost, complexity and mean time to repair. Traditionally, IT has used delay, jitter and packet loss as proxies for network performance. Legacy network performance management (NPM) tools were augmented with WAN optimization technology to accelerate traffic between data center and branch office user.
A more granular approach is to look at application payload and measuring the quality of voice and video communications. For unified communications (UC), this includes monitoring signaling between the UC components.
Meanwhile, conventional application performance management (APM) tools monitor performance of individual servers rather than across the application delivery chain – from the web front end through business logic processes to the database. While synthetic transactions provide a clearer view into user experience, they tend to add overhead. They also do not experience the same network latencies that are common to branch office networks since they originate in the same data center as the application server. Finally, being synthetic, they are not representative of “live” production transactions.
Service delivery must be unified across the different IT silos to enable visibility across all applications, services, locations and devices. Truly holistic end-to-end user experience assurance must also map resource and application dependencies. It needs to have a single view of all components that support a service.
In order to achieve this, data has to be assimilated from network service providers and cloud service providers in addition to data from within the enterprise. Correlation and analytics engines must include key performance indicators (KPIs) as guideposts to align with critical business processes.
Through a holistic approach, the level of granularity can also be adjusted to the person viewing the performance of the service or the network. For example, a business user’s requirements will differ from an operations manager, which in turn will be different from a network engineer.
A unified platform integrates full visibility from the network’s vantage point, which touches service and cloud providers, with packet-level transaction tracing granularity. The platform includes visualization for mapping resource interdependencies as well as real-time and historical data analytics capabilities.
Taking a holistic unified approach to user experience assurance enables IT to identify service degradation faster, and before the end user does. The result is improved ROI throughout the organization though reduced costs and higher productivity.
Optimizing performance of services and users also allows IT to evolve toward a process-oriented service delivery philosophy. In doing so, IT also aligns more closely with strategic initiatives of an increasingly data-driven enterprise. This is all the more important as big data swamps the enterprise. It is why I suggested in a recent article that user experience assurance should be big data job number one.