After years of investing in VoIP technology, service providers are seeing a ramp in subscriber numbers. Likewise, many enterprise networks have at least pilot-stage VoIP deployments. Yet organizations continue to face significant risks due to the relative immaturity of VoIP management methodology. As with many emerging technologies, focus on manageability lags behind the drive for deployment. But, because telephony is a sensitive, mission-critical service tied closely to business processes, a comprehensive VoIP management strategy is critical, particularly as deployments expand beyond the trial stage.
A sound overall VoIP management strategy should be:
- Multi-layered: Management must encompass all IP network infrastructure involved in delivering the VoIP service, including individual network device status, VoIP signaling and call-quality monitoring and visibility into the behavior of the underlying IP network layer.
- Continuous: The VoIP network must be measured continuously, in real time, and throughout the cycle of the VoIP application, with accurate event history captured for effective forensic analysis.
- Deterministic: Beyond information on call quality, IT managers need data on the root causes of problems, so that immediate corrective action can be taken and recurrences prevented.
An important yet often overlooked aspect of VoIP manageability is management of the underlying IP network. Even when VoIP components themselves are behaving flawlessly, complex Layer 3 dynamics can bedevil VoIP quality. This is because VoIP causes a fundamental shift in network utilization by using many more paths through the network than legacy data applications. Most legacy applications, be they categorized as "client server," "web-enabled," etc., function primarily on a client-server basis, where many client hosts access relatively few application servers. These applications traverse a preset number of paths that go primarily to the Internet and a few key data centers within the enterprise.
Not so with VoIP. Though the call-setup traffic between phones and gateways is client-server in nature, actual call traffic moves peer to peer, utilizing many more paths across the network – and dramatically complicating manageability. A trade publication asserts that, "While a number of factors can degrade voice clarity . . . by far the number one culprit is the underlying IP network itself." This is backed up by a recent study (done by the University of Michigan and Sprint) finding that 59% of network problems are caused by routing or related Layer 3 issues such as routing misconfigurations, protocol software bugs and faulty design. If these issues are cause for concern in networks where legacy data applications utilize a fairly stable set of paths, how much greater will their impact be when the networks add VoIP, and network paths that previously were seldom used are suddenly mission-critical? Even in a carrier network carefully engineered to deliver toll-quality voice services, studies have shown that routing and other Layer 3 issues can wreak havoc with VoIP call quality and availability.
Recognizing these issues, VoIP equipment vendors recommend that, to effectively troubleshoot VoIP call problems, users have an accurate Layer 3 network map network and be able to identify the precise network path(s) that a call traverses. However, neither traditional SNMP-based network management solutions, which monitor the status of (and detect failures in) individual network devices, nor VoIP application management solutions, can provide this map, since neither has real-time or historical visibility into the dynamic routing and Layer 3 topology in IP networks. As a result, network engineers are often hard-pressed to explain, let alone prevent and resolve, Layer 3 and routing problems plaguing their organizations' VoIP deployments.
A new technology called route analytics addresses the need for better management of the underlying IP network for VoIP deployments. Route analytics provides visibility and automated analysis of the dynamic, logical network-layer elements – e.g., IP network addresses, routes, routing events, routing protocols (OSPF, IS-IS, EIGRP, BGP), routing attributes and policies – that determine the IP network's dynamic topology, which in turn controls network behavior and traffic flow. Route analysis tools are a key component of both pre-deployment network assessments, to determine network "readiness" for VoIP; and post-deployment network monitoring, to discover, diagnose and resolve VoIP problems.
Route analytics solutions listen passively to all routing exchanges on the network and deliver a "router's eye view" of Layer 3 activity, providing previously unavailable intelligence on the real-time, network-wide behavior for use in VoIP planning and diagnosis. Route analytics provide accurate and timely notification of Layer 3 issues such as route flaps, missing or invalid routes that can cause levels of latency, jitter, packet loss and unordered packets that are unacceptable in VoIP operation. It validates routing operations in real time, before, during and after network maintenance, to prevent configuration errors and service interruptions that can quickly drive up VoIP costs. And it provides a historical record of all Layer 3 topology events and changes, which dramatically improves forensic analysis and troubleshooting, and works with diagnostic and planning tools to help network engineers quickly get to the root causes of VoIP problems.
The advent of route analytics brings to organizations a means to ensure comprehensive manageability, reducing the risk of compromising the business and technical cases used to justify the deployment of VoIP in the first place.
About the author:
Alex Henthorn-Iwane has been senior director of marketing at Packet
Design, Inc., since 2004. Previously he was senior director of product
management and marketing at CoSine Communications, Inc., a provider of network-based IP Services platforms enabling rapid delivery of
communication services by service providers; there he worked with a
number of global broadband providers on next-generation broadband
architectures. Before that he was senior director of product and
program management at Corona Networks. He has also held product
management positions with Lucent Technologies and Livingston
Enterprises (acquired by Lucent); and systems engineering management
posts with Fibronics America. Henthorn-Iwane, who has been a principal
member of the DSL Forum, holds a B.A. degree from the University of
California at Berkeley.
