Assigning Action Items

In the next phase of the problem-solving process, a system needed to be designed around Windows Media, and an implementation plan formulated. Action items were assigned to two internal groups. The IT department was asked to draw up a network topology that included Windows Media, and to decide how it was to be implemented. The Media department was to make recommendations on how to produce the content.

Once the system was in place, IT would maintain the servers and infrastructure. They would use the Enterprise Deployment Pack for Windows Media Player to package the Player and settings into a simple one-click setup program. Then they would add a section to their Web site from which users could download the Player and obtain setup help.

The groups decided to deploy Windows Media in three stages. At the end of each stage, a fully operational system would be in place. Subsequent stages would add on to what was already there without disrupting the operation of the preceding stage.

The company had several reasons for deploying in stages. First, there was no critical reason for pulling out all the stops and doing a large-scale deployment that would require lots of resources. The current systems were working. There was no reason to stop using them right away or to disrupt their use. Also, the new system was not merely a hardware upgrade. Deployment would affect workflow patterns, schedules, and corporate culture. The Windows Media-based system could be blended in with the current system over a reasonable amount of time.

Second, managers wanted immediate results. They didn’t want to wait until the system was completely finished before using Windows Media. This makes sense; completing every area of Windows Media deployment takes time and resources. If deployment begins with the origin servers, there is no need to wait until all of the cache/proxy servers are configured before the company can start streaming in Toronto. Also, there were certain points during deployment when slowdowns could occur, such as when configuring routers and switches to pass multicast packets. The process of configuring the infrastructure could take place during and after the time when on-demand streaming was being implemented.

Third, IT knew from experience that unexpected things could happen. A phased approach would make it easier for IT to locate trouble spots. Rather than build the entire system and then test it, it made more sense for them to build, test, and release continuously as they proceeded—making sure a setup was successful before adding on to it.