[October 11, 2005] Enterprise IP Convergence: Performance Management Takes Center Stage Tweet By Bruno Zerbib, Director of Product Management, Network Solutions for InfoVista   With affordability and quality greatly improved, IP convergence’s killer voice and video applications (IPTV, unified messaging, VoIP, video conferencing, e-learning) have gained widespread momentum.  As these applications touch end users, however, real-time performance management has taken center stage.    While service providers are increasingly able to roll out new applications for enterprises, pressure is intensified to ensure that users experience these services the right way the first time they are exposed to them.  If service providers fail to deliver an experience that users are accustomed to—toll or broadcast quality—adoption rates will not reach levels necessary to achieve profitability.   Converged services have introduced an unprecedented level of complexity into IT management and require a new approach for managing performance and quality of service (QoS).  Addressing network issues as they unfold is no longer effective for meeting enterprise users’ expectations.  Service providers must embrace an approach to performance management that supports predictive visibility and flexibility to respond to the new requirements and challenges of IP convergence.   IP Convergence:  New Technologies Create New Management Challenges   For decades, service providers have delivered enterprise IP services over a heterogeneous infrastructure, with dedicated networks for voice, data and video applications.  As these networks converge to deliver differentiated and multiple services, new performance management challenges have arisen.   Multicast-based converged applications, for example, are by design subject to inherent performance degradation risks impacting the enterprise end-user relationship.   These performance risks stem from the following characteristics:  multicast traffic is hard to predict; users are mobile by design; the ability to easily join and leave multicast sessions impacts service and multicast applications tend to be delay sensitive.  Performance management tools need to address all of the special circumstances these applications generate through a predictive and proactive approach to management, which resolves performance issues at the source before service degradation occurs.      Adoption of IP convergence applications is also plagued by end-user acceptance issues.  For example, not only are voice applications like VoIP susceptible to delay and jitter, they are also subject to the highest degree of scrutiny from the executive suite.  Service providers will not keep the business of enterprises, unless they can communicate technical performance metrics and service level agreements (SLA) in a way that makes sense not only to the IT department but also to the end business user.  While an SLA might guarantee 99 percent uptime, this technical specification means nothing to an executive who can’t make a simple phone call due to quality issues.   Service-Centric Solution:  Manage Performance as a Service   As IP convergence heightens opportunities and competitive fervor among service providers seeking to accelerate time-to-market for new revenue-generating services, it also comes with new demands from enterprise users.   Enterprise IT organizations are demanding increased network productivity, better management of differentiated QoS levels, and new ways that they can measure, understand and negotiate service levels with their providers.    A service-centric approach to network management is integral to addressing each of these issues and ensuring expected levels of service quality to all constituents in the enterprise.  Service-centric network management provides a snapshot into the performance of connections and components throughout the entire networked infrastructure and empowers IT managers to detect, isolate and resolve network service level problems before they happen.     The challenge of maintaining availability and performance of the service therefore shifts from after-the-fact, event-driven detection and resolution to proactive identification of network events.  Only proactive service-centric performance management can meet this shift, particularly since service degradations occurring in an interdependent environment are by nature difficult to diagnose, isolate and resolve.    There are several service-centric best practices that are key to ensuring effective and efficient delivery of converged services.      Regain domain expertise: During the IT bubble of the 1990s, service providers employed a large number of experts on staff with the resources to devote to performance issues.  After the bust, providers laid off their internal experts losing much of the technical expertise customers had come to expect.  Service provides should seek to regain this expertise by choosing their technology vendors wisely and leveraging the domain knowledge that these firms house.   Obtain a cross-silo view: Delivery of converged applications transcends network, system and application layers and therefore requires a holistic approach to performance management.  Service providers must choose a performance management platform that provides visibility into each of these elements within the IT infrastructure and supports cross-functional teams that span the IT and telecommunications infrastructure.    Measure service levels as close to the end user as possible: Service providers need to measure service levels across a performance continuum, tracking not only hard failures but also the subtleties of service degradation.  End-to-end service-level measurements of both the active and passive varieties are critical for proactive service level assurance and contractual reporting on service level delivery out to the end user.   Proactively optimize infrastructure resources: Research indicates that many IP telephony implementations encounter problems because infrastructure performance management and optimization is not initiated in the beginning stages of a service rollout.  Initial infrastructure optimization avoids problems caused by insufficient capacity planning and ensures the best possible returns on application rollouts by making optimal use of existing bandwidth and IT investments.   The Bottom Line:  Manage Converged Applications as a Service Continued convergence around IP-based networks sets new challenges for performance reporting and service management.  As IP convergence drives the delivery of cross-silo, network-aware services, the optimal path to addressing the unique management challenges surrounding the management of this complex technology paradigm is to manage it “as a service.”   [ Back To TMCnet.com's Homepage ]