Pre-Deployment Testing and Implementation of a Converged Network

After all the planning has been done and the network assessed, the conclusions drawn and assumptions made must be tested to assure that they are correct before moving on with the migration. Testing new applications and hardware in a production environment can be potentially disruptive to a business if things don't go as expected. It is therefore essential that the testing and tweaking be done on an isolated network segment.

The Virtual Network Test Lab

The test lab should mirror the production environment in all aspects. As duplicating a network would be cost prohibitive from a hardware point of view, IT managers have turned to virtualization software to setup and emulate the production network, and to test the effects of MACs before deploying them live. Virtualization pioneer VMWare offers a variety of virtual networking solutions, and in 2004 Microsoft released its Virtual PC (VPC) software.

A virtual network is an invaluable tool that allows IT professionals and system integrators to accurately emulate current and planned network conditions, and to assess the impact of new applications, moves and changes, and new hardware on the network. In the case of implementing convergence, some specific best practices have been adopted as to how testing should proceed in a virtual network test lab.

To ensure the testing environment is realistic, conditions on the production network should be captured over an extended period of time, taking into consideration the busiest time of the day, week, month, and year, as well as average and below average periods of network congestion. Once captured, conditions can be replicated in the test lab, adding the VoIP application and connecting phones or using call generation software to accurately reflect the impact on the network in best, average, and worst case scenarios.

Calls can be analyzed from point to point to monitor QOS as well as for audio quality using human indicators such as the Mean Opinion Score (MOS) rating system. Have end users listen to calls under a variety of conditions to indicate their level of satisfaction with the quality of the call. As mentioned before, bringing the end user into the loop at all stages of deployment will help ensure a successful migration and help avoid mass revolution.

Before, during, and after the rollout, a virtual network test lab can play an important role in testing and optimization. Initial objections by management due to cost considerations should be overcome by the potential ROI realized through enhanced troubleshooting techniques, and the ability to see the effects of MAC before implementation on the production network. A useful tool not just for deploying IP Telephony, but for the network as a whole.

Rolling Out the Converged Network

For smaller businesses, after all the testing is done and confidence is high, it may be advisable to migrate all users in a single sweep. For larger organizations and enterprises with a more complicated network topology or departmental scheme however, they would be well advised to proceed with caution with a partial migration, rolling out the new application on a departmental or regional basis.

The order of migration should be logically considered, and although the pressure may be there, political considerations should not be high on the list of priorities. The last thing any IT manager wants is to roll out the CEO first, to be witness to all the glitches and stumbles that will inevitably occur at the beginning of the migration. Planning the roll out should be based on a number of factors, including but not limited to:

Which department or region has the best qualified support staff, or has taken part in the testing and is most familiar with the system. The early rollouts are the slowest and will for sure have the most problems. Consider the first roll out as a test and let the group closest to the project, perhaps the now converged IT/Communications department, enjoy the fruits of their labor. If rolling out a remote location, choose one with the nearest proximity to the location of the support staff. Subsequent migrations will become easier and the knowledge gained will make for smoother transitions.

The regions or departments that are the most needy should also be considered candidates for early migration. Branches that have sufficient network infrastructure but an aging or outdated PBX system should have the priority, while locations that need significant network upgrades should take care of that first. For example, a public school district may want to deploy IPT to schools that have the capacity first, while simultaneously upgrading the older schools networks to sufficiently accommodate data and voice.

In a campus LAN situation, it may be feasible to migrate by building, where if only one building is involved, migrating by department may be the better solution. For the International Corporation with offices in India, migration may not be possible due to legal considerations of Voice over IP in that country.

You also have the option at this juncture not to migrate. As long as there is sufficient telephony equipment to do the job, it may make better business sense monetarily to postpone migration. As VoIP and IPT mature, competition in the market place grows and prices will eventually fall. Functioning legacy equipment that is not yet fully depreciated could also be reason enough to postpone deployment. If all is well in the organization, there is no reason to hurry; the technology of IPT and Unified Communications will only get better and cheaper over time.