| Martin had always treated the POC and pilot like they were deployments. In this way, issues of manageability and scalability could be discussed with the vendor while the pilot went on. And, when it came time to do a deployment, all these issues had already been addressed. Jason looked at the calendar and realized that it was Wednesday afternoon on the second to last day of the trip. He had many outstanding issues to still address, the least of which was defining the resources needed for support and deployment. Deploying an executive was out of the question. |
The Methodology As the pilot comes online, the issues of scalability and manageability come to the forefront. During the POC, the user population was relatively small and easy to manage in a distributed way. Also, the infrastructure needs were minimal. Now that there may be many more users, scalability and manageability will be tested . To help in recognizing these needs, the following questions should be asked: -
Do we need additional infrastructure ? ” As the pilot starts consuming the current infrastructure, it will become clear whether the current infrastructure will support the pilot and possible future rollout. Is the authentication server or database server able to handle the authentication demands? Are the networks able to handle the extra bandwidth, especially at peak times of the day? Answers to these types of questions will allow the project team to decide if the infrastructure needs to be upgraded for a rollout to occur. By finding this out now, the infrastructure can be upgraded in a planned and controlled way. -
What changes do we need to make to the installation package ? ” It may become evident that certain tweaks are necessary to make it run smoother. Special installation cases may be identified, and original assumptions may be proved wrong. If possible, these changes should be made to the package during the pilot deployment and tested on the outstanding pilot desktops. -
Is the user initialization scaling ? ” For the end-users to start using the biometric system, they need to initialized . This initialization may range from user creation in a database to the addition of a schema to an existing LDAP directory. If the current methods for implementation seem cumbersome or time-consuming for the pilot, then they must be modified to scale for a potential rollout. -
Do we need additional support personnel ? ” As the pilot is deployed to many more users, a general level of user support will set in. There will be an initial surge in support requests at the start of the pilot. As this declines over time, there will be a general leveling off in calls. This new level needs to be sustained. This information can be extrapolated to the projected size of the deployment, and will in turn provide the number of support personnel required to move forward. -
Can this be centrally managed ? ” This is very important for general deployment. It will not be practical to visit every desktop or part of the company for general management of the biometric solution, so the biometric solution being tested needs to be centrally managed. If it is not currently feasible to do this, is the vendor willing to support centralized management? What mitigating steps can be taken internally until centralized support is provided? The sooner in the pilot that issues of scalability and manageability can be addressed, the easier it will be to prepare for deployment. In addition, scalability and manageability are more than likely important success criteria that need to be met. By gathering this information early on, it can be used to help in the evaluation. |
