End-to-end service availability support: Theory and application

Abstract

To ensure that Alcatel-Lucent solutions deliver the end-to-end service availability and quality expected by service providers, new capabilities are required to model, measure, and evaluate the robustness and effectiveness of the solutions over their lifecycles. This paper describes the implementation of a proactive end-to-end service availability support platform that meets the needs above for network technologies and services such as IP Multimedia Subsystem (IMS) and Internet Protocol television (IPTV). Key innovations include a model that predicts end-to-end service availability across an entire solution, near-real time secure collection and centralization of all available network and service data, and a data model that derives the critical end-toend service availability metrics from this data. These three elements were combined and implemented. The platform enables operations staff to receive early warning indications of impending issues with each service. It also helps technical assistance centers to rapidly troubleshoot critical problems. We present lessons learned from practical application of these methods to IPTV solutions. © 2010 Alcatel-Lucent.voice-centric circuit-based networks. In the past, data services have seldom been required to provide the levels of quality and reliability that are required now that services like TV and other applications are carried over them. Subscriber satisfaction depends on a service as a whole working correctly, with high quality and uninterrupted availability. Accordingly, service providers must establish an environment for efficient, entiresolution support, as opposed to maintenance of individual product robustness.In addition to network complexity, lack of experience with next-generation technologies is a major stumbling block for providers planning large-scale IntroductionA significant challenge telecommunications companies face today is management of end-to-end service availability (SA) and quality. Twenty-first century networks are technologically capable of delivering a wide range of Internet Protocol (IP)-based services to subscribers, with few limitations on the types of applications available. Moreover, deployments are expanding impressively in scale: for example, in December 2008, AT&T announced its one-millionth Internet Protocol television (IPTV) subscriber in a 16-state, 79-market offering [6]. The abundance of applications and their explosive growth require a radically different operational approach from that of twentieth century deployments. Problems are difficult to diagnose, and the impact of issues is difficult to assess. Providers often must rely on subscriber calls to detect service problems-even problems deep inside the networkand use various ad hoc, manually intensive methods to address complaints. This approach is unsustainable as services expand to support millions of subscribers.In this paper, as a solution to these problems, we present the concept of end-to-end service availability support (E2E SAS). E2E SAS combines a wide range of techniques th...