What happens after PBX?
- May
- 09
If you are looking at integrating telecom into IT at your company, then the following is for you. For decades companies procured and maintained separate systems, typically called a PBX, to provide reliable dial tone to an organization. As telecom and IT organizations and their budgets got merged, CIOs not only want to reduce cost, but standardize operations and datacenter requirements on common best practices. With legacy PBX systems that is not possible, so that a new solution is required to provide reliable dial tone as an IT application. Here are the eight steps to consider for a successful migration.
1) Assess your current environment: Situations in different companies are much more alike than you might think. Most large enterprise that operate several larger locations and potentially many smaller locations sit on a widely distributed and typically complex installation to provide dialtone to each location and every employee. Requirements for access to the PSTN grew over time with many telco lines connecting into these locations. Local or regional clusters were built, not allowing users to seamlessly roam between locations. If your company went through acquisitions, you typically look at a multi-vendor environment. In most cases software and hardware releases are not current and in order to be able to keep your vendor’s support contract there is a need for constant costly upgrades. Licensing is a jungle and an inhibitor to harmonizing the system and adapt it to changing needs in your company. Completely different skills and operating procedures are required to maintain the telephony platform.
2) Start with an enterprise wide SIP core: Do not fall into the trap of assessing one complete system against another complete system from a different vendor. Successful migration into the post-PBX era requires a different layering and a departure from the vertically integrated model of the past. The foundation for every scalable, flexible, and cost effective system is an open SIP core that can be deployed enterprise wide or globally, offering not only flexible call routing, but user roaming without boundaries. Complete 100% standards compliance and openness should be the core requirement. In a horizontally layered infrastructure model you will want to be able to add applications that connect to this SIP core later from different vendors, and therefore you have to insist on complete adherence to the SIP standard and openness of the solution. This is the most critical foundation of your new system.
3) Consider your IT best practices: What skills, infrastructure, and best practices do you have already established in your organization in support of other business critical IT infrastructure and applications? Have you moved into a private cloud or IaaS infrastructure? What level of outsourcing is in your plans? How do you do monitoring and reporting, backup & restore, and what are your DR requirements and practices? Are you on a Linux or Windows stack? What skills do you have in your organization to manage such an application? Answers to these questions are critical as you assess not only your needs, but also your future cost.
4) Take stake of your network infrastructure: Especially if this is your first move into communications over IP or VoIP, a network and plumbing infrastructure assessment is an absolute must. Your new system will only perform as well as your network does and user experience critically depends on things such as QoS management to reserve bandwidth and guarantee low jitter and delay. Today’s routed and switched networks are perfectly capable to deliver everything you need, provided they are properly configured and you have implemented a good network topology. In addition, you will have to assess your DNS and DHCP infrastructure as a modern SIP core heavily relies on DNS to load-share and failover.
5) Migration requirements: The transition to a new system does not happen overnight, especially not in a large organization. Therefore, you will need a carefully crafted migration strategy that allows you to migrate one location at a time, one department at a time, or one service at a time while maintaining interoperability and feature and user experience transparency during the migration. Most legacy systems either do not provide SIP interfaces, or provide SIP capabilities not up to par with respect to standards compliance or features. Therefore, different options for interoperability with legacy systems have to be considered.
6) What to do with phones? Phones represent the largest percentage of your new system cost and therefore you want to carefully assess your options. If you are on an old TDM system it is likely that phones used by your employees are amortized and up for renewal. For newer systems, such as Cisco CallManager implementations, you can consider re-using these phones. An alternative strategy used more and more includes a ‘bring your own device’ program where employees might want to use tablets and smartphones as end devices instead of desk phones. With a standards compliant and open SIP core you have the flexibility to consider end devices from a variety of vendors including all kinds of apps running on tablets and smartphones, letting your organization participate in and profit from the accelerating level of innovation in this field.
7) Application and business process integration: We all remember how cumbersome and expensive CTI based integration used to be in the past. As a critical requirement your new system needs to offer a comprehensive set of Web Services based interfaces. With that communications enablement of applications and business processes becomes easy. Every organization uses a distinct set of critical applications to conduct its business. The new communications solution needs to be integrated with these applications, offering communications services right in the same browser based user interface as these other applications. This can now easily be done. In addition you will want user profile integration, social graph integration, common activity streams, and a cohesive experience as users go from one application to another.
8) Add up your cost: Cost matters to every organization, small and large. The variance on solution pricing is huge and the general rule is that the more vendor specific and vertically integrated a solution is, the more it costs. That might seem counterintuitive to a customer, but it makes perfect sense to a vendor. As solutions are more and more looked at as a service, operating cost is often more important than upfront CAPEX. Lowest possible OPEX depends on the degree you can apply established IT best practices to both operations as well as hardware requirements. The pricing model is critical too. Do you buy a fully featured seat, or does the vendor itemize pricing and charge for every single thing separately? What operating system, database, and middleware stack does the solution run on? There is a significant different in price between a free Linux stack and a vendor specific stack with a commercial database such as Oracle or Microsoft SQL. Do you have to build regional clusters or can you create one global cluster instead. We have shown that for e.g. a four location global deployment the difference in the number of servers required to build the solution can be as much as 6x. This translates directly into operating cost and not just CAPEX.
We hope the above steps are helpful in your assessment of your dialtone solution requirements. Times have changed and the decade old vertically integrated model is giving way to an IT application as we enter the post-PBX era.
