This lesson examines some of the core Windows NT network services and configurations that will affect your Windows 2000 infrastructure design or that might necessitate fundamental redesign prior to deployment. The physical and logical network topology of the Windows NT network and the services it offers to clients will dictate such design issues as these:
After this lesson, you will be able to
Estimated lesson time: 25 minutes
Prior to migration, you should ensure that all diagrams of your network environment are current. Layouts displaying both the physical network and the logical network are required. The physical network diagram shows how the machines are actually connected and must contain the following details:
Figure 4.1 shows an example of a physical network diagram and the required types of information.
Figure 4.1 Physical network diagram
The logical network diagram should provide the following information:
Figure 4.2 provides an example of a logical network diagram. Compare the information it contains to the physical network diagram in Figure 4.1.
Figure 4.2 Logical network diagram
A fundamental understanding of your current Windows NT domain structure will help determine the following design factors:
Understanding your physical network infrastructure will help with such Windows 2000 design issues as these:
Detailed information on the server roles listed next can be found in the MCSE Training Kit—Microsoft Windows 2000 Active Directory Services. If you don't understand these special server roles, you should read the Active Directory Services book or attend a course on Active Directory directory services before considering a migration from Windows NT. The strategic physical placement of these servers will facilitate faster delivery of network services.
Server types that exist in Windows NT and are carried over to Windows 2000 include:
Server types that are new and unique to Windows 2000 include:
Remember that several of these services can be held on the same server, so the infrastructure design issues might not seem as enormous. However, do try to include servers that can act as backups for fault-tolerance purposes. For example, once your new Windows 2000 infrastructure has been established, you should plan for a server that could take on the role of PDC emulator in case the current one for that domain fails or in case you need to take it offline for routine maintenance such as a memory or hard disk upgrade.
Examining the physical and logical network arrangements will give you an idea about where high amounts of network traffic are likely to occur. This can be confirmed by gathering information on the bandwidth usage and then analyzing what would happen if extra traffic from the migration or the new infrastructure were added to that area of the network. You should perform this analysis when you know the network is busy—for example, at the start and end of a work day.
You can use the Systems Management Server (SMS) Network Monitor component of Microsoft BackOffice to collect data on network loading and activity. The data will then allow you to establish a baseline requirement for network and server performance and the amount of spare capacity that will be available for use during the migration.
When assessing network traffic for design considerations, remember to build in capacity for other services in addition to Windows 2000, such as Microsoft Exchange Server 2000, workstation management services such as SMS, and online video training.
Having established the network's baseline requirement under normal loading, you can consider the additional traffic that will be imposed during the migration and the bandwidth requirements of the final system. Two volumes of the Microsoft Windows 2000 Server Resource Kit (Microsoft Windows 2000 Server Deployment Planning Guide and Microsoft Windows 2000 Server Distributed Systems Guide), provide information about bandwidth requirements for replication and authentication traffic within Active Directory.
The failure of a server or network connection can seriously affect the configuration's overall performance as replication and authentication traffic is redirected. You should consider the impact of such failures as part of the analysis of the network infrastructure, both during and after the upgrade, and you should make contingency planning a part of your recommendations.
Unless the Big Bang approach is being adopted (in other words, the complete conversion of your Windows NT environment overnight), many of the network services just mentioned will need to be available to users during the migration. Hence, part of your migration plan must include how to minimize potential interruptions to the numerous network services that are running. For example, how will you maintain DHCP services while migrating the DHCP server so that users' systems will still be able to access the network? Your migration plan should minimize the impact on the business by keeping as many of the network services running as possible (see Chapter 12, "Business Continuity," for further information). The communications team should keep users informed of the teething problems they'll face during this period.
In this lesson, you learned to plan server locations by having an understanding of the physical and logical network topologies. You also looked at how the various network services need to be maintained during and after migration to keep the business running.