The Road to Network Convergence: More Delays, Disruptions, and Shelved Projects than Enterprise IT Success Stories

So, what’s causing the delays, disruptions, and shelved convergence projects? Why are enterprise IT organizations struggling to launch converged applications and services from a pilot to a production environment? There are many reasons, but perhaps the most obvious one is simply the management complexity inherent in converged applications. In the 40+ year history of enterprise-based information technology, very few IT organizations have mastered and assimilated a technological innovation as multi-faceted as convergence. In this article, we will discuss why many enterprise IT organizations experience delays and disruptions in their deployment of converged applications and we will provide a plan of action: a set of convergence best practices, which will enable the enterprise IT organization to succeed on the Road to Network Convergence.

What’s causing the delays, disruptions and shelved projects?

Convergence projects like IP telephony, perhaps the most in-demand converged application, require planning, planning, and more planning. Believe it or not, the majority of enterprises who experienced problems deploying converged applications in a production environment failed somewhere during the planning phase. Perhaps they started with an incomplete network assessment, or used an existing rather than a specialized tool to monitor the converged application, or they eliminated some “best practices” because they seemed expensive and unnecessary. In any case, these unintentional oversights resulted in problems like poor call quality and dropped calls in the production environment. Their lack of planning ultimately resulted in lost credibility with their end users and a black eye for the IT organization. So, the first rule of convergence is: IP Convergence isn’t guaranteed, it results from a plan of action based in knowledge of the enabling technologies, and in the complexity of the business critical applications and services provided to the end user. To successfully deploy a converged application in a production environment, IT organizations need to do their homework, plan, and follow a set of industry best practices.

Convergence also requires a new organizational culture to succeed. Separate application, systems, and network silos in an IT organization will negatively impact the quality delivered. To meet the 99.999% application availability/reliability, which users have come to expect from the PSTN phone service, the enterprise IT organization needs to adopt the strategy and mindset of the group who set the five-nines standard, the service provider. Service provider’s keep and grow their customer base by delivering optimal service quality. Their organizations are cross-silo by nature, with a team of application network and systems people working together to ensure service performance. To succeed with convergence, enterprise IT organizations require cross-silo teams to administer and support convergence best practices. Every person in the enterprise IT organization needs to focus on the quality of the service delivered; is it performing as expected, is capacity available to launch new applications and services, and is there a quick and efficient way to troubleshoot problems?

It’s easy to overlook the importance of the network in delivering converged applications and services. In the past, network performance problems were usually fixed by simply adding more bandwidth. As enterprise IT organizations began deploying converged application and services, they continued with the status quo; more bandwidth was added and it may have fixed the immediate problem, but at some point, issues began to reappear. What these IT organizations soon came to realize was that the network wasn’t simply a second string player on the convergence team, it was the quarterback. Perhaps the single most important component of a converged, cross-silo (application – system – network) service, because it was the delivery mechanism. And without the delivery mechanism working as planned, everything else was irrelevant. Whether ongoing or intermittent, problems with network performance play the greatest role in limiting the number of enterprises who have successfully launched converged applications into production. So, how was the network overlooked? In most cases, the appropriate mechanisms were not put in place to allow the enterprise IT organization to monitor and manage performance.

What many IT organizations lack is a management platform, which provides service-based view of the network, enabling them to monitor and manage network performance. As enterprise networks evolve from data only networks to converged data, voice, and video networks, their quality of service (QoS) strategy must evolve as well. Unlike data, which is elastic and can bounce back from network delays and congestion, voice is rigid, and it will experience quality degradation if packets are delayed or discarded. Converged networks require QoS policies in place to determine which traffic gets priority. Although many enterprise IT organizations have QoS enabled on their network, they have no easy way of monitoring it, verifying that it is configured correctly, or if it is performing as intended.

Another area where many enterprise IT organizations fell short, was in the collection and analysis of data that accurately reflects the end-user quality of experience. These metrics provide real insight into how well the application or service is performing. For IP telephony, metrics including call quality, service availability, and service reliability best represent the overall end-user experience. Other metrics such as dial tone delay, post dial delay, call setup time, and call completion ratio provide additional perspectives on the user experience. These metrics need to be collected in real-time and presented in a manner which provides insight into how the service is perceived by the end user. Real-time data supports the ongoing performance management of the service, provides immediate feedback on the end-user experience, and can play a critical role in troubleshooting performance anomalies.

Lastly, there are a considerable number of enterprise IT organizations who planned effectively, implemented cultural changes, accounted for the importance of the network, and monitored end-user performance during the pilot phase, but launched into production using a management platform which didn’t scale. A resource or element-based view works fine for fifty, or five hundred users, but try and manage application or service performance for thousands of users in the same fashion, and you will find it slow, cumbersome, and ineffective at best. To successfully deploy converged applications and services to tens of thousands of users, enterprise IT organizations need a management platform which will provide visibility into both service and infrastructure and the right level of aggregation, logical grouping and drilldown capabilities, enabling them to quickly and easily detect performance anomalies and preempt service disruptions.

Plan of Action: Convergence Best Practices

First and foremost, the distributed nature of converged applications requires a service-centric approach to management, which means shifting from traditional resource-based management to managing from a business or service perspective. A critical aspect of this is focusing on performance, which literally means ensuring the application or service is available, reliable, and responsive. Most enterprise IT organizations have experience implementing performance management, however it was most likely done in combination with fault management, where performance took back seat to fault in terms of priorities. With converged real-time business critical applications like IP telephony, the reverse approach needs to occur. Enterprise IT organizations can’t afford to be responding to service faults or outages, when it is too late. Instead, they need to get ahead of the problem using a proactive performance management strategy to ensure service quality.

Adopting a service centric approach to managing converged services requires cross-silo visibility into the underlying applications, systems and networks, which support the service. To do this enterprise IT organizations need a management platform, like InfoVista’s VistaInsight ® solutions, which link the underlying infrastructure resources to the service and to the business unit or customer using the service. This top down (business unit – service – system – resource) view enables the IT organization to track service or application quality, prioritize problem resolution based on business drivers, and more effectively manage service performance.

Another convergence best practice is the collection of service quality metrics, which are used to evaluate and compare the actual service delivery and performance with the business goals. The business requirements of converged service consists of delivering agreed levels of service quality, availability, and reliability. A converged service like IP telephony requires a sophisticated data collection strategy to monitor and report on end-to-end metrics like call quality (MOS), reliability and availability. To support this, the performance management platform needs to collect and display transaction-based service quality data for both end-to-end (e.g. between different users, locations and business units) and aggregated views (e.g. for an individual end user, location or business unit). In addition, time aggregated views, which provide daily, weekly, and monthly reports of key metrics are needed to fulfill service level reporting requirements. Specific instrumentation (such as probes or agents) may be required to collect end–to-end data or it may be abstracted from the underlying performance data collected from the systems and resources, which comprise the service (for example, a Call Detail Record, or CDR, from a call processing server). Data collection occurs using both active measurements, which enable capacity planning, baselining, and trending (used to pinpoint service anomalies), as well as passive measurements, which accurately reflect the end user’s service quality experience. Best practices support using active and passive measurements in combination, in order to meet specific business requirements.

As part of their best practices, IT organizations need to be aware of how a converged service fits into the overall service portfolio. The performance of a converged service may be affected by many of the same issues that affect data-intensive applications and services, like security, right-sizing, quality of experience, and processing time. IT organizations will want to routinely monitor these aspects of their converged services.

Once a service has been launched into production, the enterprise IT organization can perform service level optimization to assure the continued delivery of consistent service quality. This ongoing, day-to-day performance management strategy includes several key tasks:

• Proactive detection and alerting of service degradations for operations teams to use in pre-emptive problem resolution.
• Accurate measurement and reporting of service and resource utilization through critical metrics.
• Forecast capacity and days to outage at both service and resource levels based on historical trends of use.
• Capacity review to plan for service expansion and support the launch of new services.
• Adjustments to the QoS strategy for high priority voice traffic to maintain service levels.

Conclusion

Today’s enterprises face many challenges when deploying converged services, however, these challenges can be met and minimized through the use of an effective performance management strategy, which includes aligning the IT organization’s structure with the service delivery, understanding the management requirements of the enabling technology, and recognizing and executing a cross-silo performance management plan. In addition to lowering costs, a service-centric management strategy can help IT organizations harness the value of converged networks and provide a clear vision and business justification for infrastructure upgrades. It can also provide the IT organization with a workable plan for deploying new applications and services through a unified management framework. Lastly, a service-centric management strategy not only supports the standardization of management practices, but also recognizes the need to adapt to the specifics of IP services and users – from both business and IT.

About InfoVista ®

InfoVista provides service-centric performance management solutions to both service providers and enterprise customers that are uniquely suited to address the increasingly demanding challenges of our customers in delivering service excellence. The company’s experience serving both sectors provides tremendous cross-functional synergy and uniquely positions InfoVista as a valued partner to help them transform their technology assets into strategic business value.