Today, we have a guest post from Atchison Frazer, CMO for Xangati.

A key component of the corporate network is the Application Delivery Controller (ADC), which selectively removes burdensome workloads from application and web servers to improve performance and to balance loads across servers.

But as networks and platforms become increasingly complex – with virtualized datacenters now extending into the hybrid cloud, and software-defined networking (SDN) gaining currency – two capabilities have become indispensable:

  1. The ability to rapidly “coalesce” key performance indicators of the entire IT infrastructure
  2. The ability to automatically analyze interdependencies that reveal dynamic remediation options, which enhance time-to-resolution and user productivity
  3. Managing secure, endpoint access to application workloads that are being balanced

Sophisticated ADCs achieve these goals by performing health status checks of applications and underlying servers. This trend has occurred as load balancers have moved from:

  • A network-only concentration to a focus also on application awareness
  • Support for a particular protocol to translation of multiple protocols
  • Server load balancing to broader support for a greater number of deployment scenarios, ranging from on-premise to hybrid cloud to NFV (network functions virtualization)

The traditional hardware load balancer, however, has been limited in delivering these capabilities because it forced a physical-device view of the environment and lacked the flexibility to provide end-to-end insights within hybrid-cloud topologies.

The migration of ADCs to a more flexible software-centric format was driven by the overall trend toward virtualization of network, compute and storage.

The virtualization of the ADC within an overarching virtualized or hybrid cloud scenario forced a re-examination of traditional networking architecture. When conventional hardware-centric load balancers needed to be physically close to the application server, they unnecessarily forced location-dependent restraints on network administrators, creating a static and uniform situation.

The more progressive view is that applications and underlying servers must be location-independent, and so too must the critical data that determine performance be location-independent.

This results in a very distributed and dynamic environment by its nature, whereby the virtual ADC, or “vADC,” exists closer to the virtual objects and is flexible enough to move when those objects move for reuse and scale, all while ensuring high availability of applications to authorized users from any location and to any device. Indeed, the marketplace is embracing this trend, fueling vADC growth that now represents 49 percent of the overall market, according to Gartner’s latest Magic Quadrant.

Capturing and visualizing critical performance data need to follow the same standard of flexibility and versatility. Thus a converged infrastructure that avoids intrusive network taps and agents – which have historically been required to access critical application flows – is key.

Working synergistically with a powerful data analytics processing and analysis engine, the vADC can be leveraged, in a sense, as a master tap because it is uniquely situated in the data path, brokering end-to-end application traffic at very high speeds.

The vADC’s ability to expose collected application and server data in an open, standardized format so that the data can be easily consumed by third-party performance monitoring and analytics solutions is critical. Also required is the flexibility that allows key performance data to be correlated, analyzed and acted upon along with interactional analysis of intelligence from other devices in the application infrastructure.

Only with this flexibility can IT personnel gain greater insight into how the entire virtualized infrastructure is performing, and take advantage of proactive remediation to prevent potential performance bottlenecks.

Remediation in this context is just what it implies: how the vADC-monitoring fabric responds to remediate or, better yet, prevent performance issues in the environment. However, there are many layers to virtual infrastructure (VI) performance. There are three common forms of VI performance remediation: recommended, assisted and automated, but none are very useful without real-time data capture, aggregation and visualization capabilities, combined with predictive analysis and prescriptive remedial capabilities. Only full data integration of the VI performance management platform with the vADC uniquely provides all three.

VI Performance Management Integration with vADCs 

Integrating VI performance management with a vADC is critical to extending the functionality of the vADC, which on its own may be limited in how much of the entire IT infrastructure it has visibility to. As an example, integration of the VI performance management platform with Citrix NetScaler VPX, and with Citrix XenApp and XenDesktop for application-aware optimization, can make it possible for VI admins and business users to predict and remediate:

  • Security problems between virtual desktops and servers
  • Multiple desktops overloading one XenDesktop
  • Communications problems between multiple XenDesktops and a single data store

The advantages of providing XenDesktop-based virtual desktops to employees of Swett & Crawford, a commercial insurance provider, included the heightened security that it provides as well as the flexibility for employees to work on any device any time. To reap those benefits–according to CIO Travis Conley–it was vital that the centralized computing infrastructure run flawlessly at all times in terms of availability and performance metrics monitored by Xangati.

Deep integration of VI performance monitoring with the vADC – all the way through to specific applications, such as VDI – enables VI admins to identify resource contention issues and potential VM-boot storms at all levels, including potential QoS issues for end-users, and also provides specific, recommended remedial actions (heuristics informed by algorithms and machine-learned best practices) to ensure performance optimization.

Providing deep integration with the vADC affords three other key considerations for accuracy in resolution of performance issues:

  • Real-time data, contention diagnostics and remedial action views into the entire IT infrastructure:  this capability is critical to problem identification, and for remediation before adverse affects occur and before they ripple through the infrastructure
  • Application health awareness:  this is functional important for understanding how to resolve device contention or other high-availability issues across all traditional silos
  • Analytics: by collecting data from all assets running in the organization’s environment in real-time, the VI performance management system can make intelligent recommendations for immediate problem resolution, and offer business users a predictable level of service assurance

When the analytics engine resides in memory, much faster data capture and problem resolution are possible, and remediation can be applied before issues cause infrastructure performance problems or downtime. Furthermore, when the analytics engine also considers data from other instances of the same problem, the system can correlate that data with currently observed data to increase the accuracy of service assurance outcomes.

Load balancers have traditionally provided valuable network services for ensuring application uptime, high server performance, and secure application access. However, today the ability to extract intelligence and link it to other interdependencies in the IT environment has become increasingly important, enabling VI admins to benefit from more meaningful decision-analysis data that can result in insightful remedial action.


Xangati’s infrastructure management software provides end-to-end VI performance monitoring and service assurance analytics for determining SLA targets.