Planning for network software

Start with the list of business activities to develop a list of applications that require network resources. These types of applications range from universal applications, such as e-mail and Web serving, to specialized applications, such as video conferencing and voice over IP. Depending on the list of applications and the number of users that require access to the applications, your network resource needs can vary widely. Therefore, start by developing a list of network applications to determine the scope of your network plan.

You need to obtain a network number and register your domain with an accredited domain registration service for your network to communicate with the Internet. The Internet Corporation of Assigned Names and Numbers (ICANN) maintains a list of accredited registrars in your area. Your entire logical connection scheme, including the IP addresses and the host names that are associated with the addresses, starts with the assigned name and number for your company.

For information about how to obtain an assigned name and number, see the ICANN Information Web site.

Your hosts, or end user workstations, need IP addresses that are based on your assigned name and number. The scheme that you devise needs to scale to your business needs and enable easy management, thus an addressing scheme is an important part of your network software plan. The basic scheme involves assigning a unique IP address and host name to every host in your network. This allows the applications to look up the address and the host name to send the appropriate host the needed messages.

Depending on the size and scale of your network, you must determine whether you want to use static or dynamic addressing. To manage a growing network, most networks use dynamic addressing for workstations and static addressing for servers. Dynamic addressing automates IP address assignment, which can significantly reduce management overhead. Dynamic addressing is performed through a Dynamic Host Configuration Protocol (DHCP) server. Depending on your operating environment, there are a variety of tools that can make setting up and managing a DHCP server relatively easy.

To learn about how to plan your IP addressing scheme, see the chapter about address, name, and network management in the IP Network Design Guide.

Build an IP address and host name database

You will need to keep track of all of your IP addresses and host names that are in your network to efficiently manage organizational change and troubleshoot bandwidth and latency problems. Start by creating a list of all of the IP addresses and host names on your network. There are research tools available to help you create a database that relates IP addresses or host names to individual machine names on your network topology. The machine names are typically identified by the medium access control (MAC) address on the network card of the host.

To learn about how to build an IP address and host name space, see the chapter about address, name, and network management in the IP Network Design Guide.

Some form of directory services is required to manage host names and addresses, and their interrelations in network domains and zones. You have many options that are related to the type of name service that you use. If you use static addressing, use a Domain Name System (DNS). The DNS provides applications with a mapping of IP addresses to host names wherever they reside in various network domains and zones. It also provides other vital information to the application. If you use DHCP to dynamically assign addresses, your hosts have a different IP address every time they start up and initialize a connection to the DHCP server. To manage the names and addresses in a dynamic addressing system, you need to plan for a dynamic DNS.

To learn about how to plan for name management, see the chapter about address, name, and network management in the IP Network Design Guide.

In large networks, with routers, a lot of overhead is created with a large number of hosts. To improve performance and manageability, the host number (typically the second half of an IP address) can be subdivided into a subnet number and a host number to provide a second logical network. The second network is called a subnetwork or subnet. You need to determine your subnet scheme as part of your network software plan. A good scheme can ensure that each router in your network is performing well and that your hosts do not need to send messages through too many hops to get them to their destinations.

To learn about how to plan for subnets, see the chapter about address, name, and network management in the IP Network Design Guide.

You need to create administrative subdivisions to improve manageability, to isolate sensitive information, and to ensure adequate resources to mission-critical applications. Typically, businesses are divided into a front end and a back end. The front end includes a basic file and print network, which also provides e-mail and Web browsing to users. The back end contains all of the transaction processing for the business, such as automated accounting, shipping, receiving, Web and e-commerce workloads, and other types of electronic data interchange (EDI).

Your plan for administrative subdivisions must include the platforms on which to run the subdivisions. The front end may not need to meet the rigorous requirements of performance, availability, and security that the back end requires.

IBM operating systems and server families include every type of administrative subdivision. IBM systems allow you to have a Windows® or Linux® front end with a mixed back end of AIX, i5/OS®, and Linux workloads. Consult with your solution assurance marketing representative to design an administrative subdivision scheme for your business.

To learn about how to plan for administrative subdivisions, see the chapter about address, name, and network management in the IP Network Design Guide.