Configuring Windows Server 2003 for LAN Routing (part 4) - Exploring LAN Routing Scenarios

Exploring LAN Routing Scenarios

You can use routers in many different topologies and network configurations. When you configure a server running Routing And Remote Access as a router, you can specify the following:

• The protocols to be routed (IP or AppleTalk) by the router

• Routing protocols (RIP or OSPF) for each protocol to be routed

• LAN or wide area network (WAN) media (network adapters, modems, or other dial-up equipment)

Simple Routing Scenario

Figure 14 shows a simple network configuration with a server running Routing And Remote Access and connecting two LAN segments (Network A and Network B). In this configuration, routing protocols are not necessary, and static routes need not be added manually because the router is directly connected to all the networks to which it needs to route packets.

Multiple-Router Scenario

Figure 15 shows a more complex router configuration. In this scenario, three networks (Networks A, B, and C) are connected by two routers (Routers 1 and 2). Router 1 is directly connected to Networks A and B, and Router 2 is directly connected to Networks B and C. Router 1 must notify Router 2 that Network A can be reached through Router 1, and Router 2 must notify Router 1 that Network C can be reached through Router 2. This information is automatically communicated by means of routing protocols such as RIP or OSPF. When a user on Network A wants to communicate with a user on Network C, the user’s computer on Network A forwards the packet to Router 1. Router 1 then forwards the packet to Router 2. Router 2 then forwards the packet to the user’s computer on Network C.

Figure 15. Routing with routing protocols or static routes

Without the use of routing protocols, a network administrator has to enter static routes into the routing tables of Router 1 and Router 2. Although static routes work in simple networks, they are difficult to implement on a large scale. In addition, static routes do not automatically adapt to changes in the internetwork topology.

Understanding Static Routes

Static routed networks do not use routing protocols such as RIP or OSPF to communicate routing information between routers. A static routed IP environment is best suited to small, single-path, static IP internetworks. For best results, the internetwork should be limited to fewer than 10 subnets. In addition, these subnets should be arranged consecutively (in a straight line) so that traffic pathways are predictable. A final guideline for static routing is that the topology for internetworks relying on static routing should not change over time. Figure 16 illustrates such a network.

In the example provided in Figure 16, Router C-B can see all computers on Subnets C and B. When Router C-B receives a packet destined for an address outside Subnet C or B, it forwards this packet along the default route to Router B-A. Because all computers outside Subnets B or C lie in the direction of the default route, static routes need not be added to the routing table on Router C-B.

However, for Router B-A, which sees all computers on Subnets B and A, the computers on Subnet C do not lie in the direction of the default route. If Router B-A receives a packet destined for Subnet C, it incorrectly forwards the packet to Router A-Int unless instructed to do otherwise. Adding a static route to the routing table on Router B-A, as shown in Figure 17, allows Router B-A to properly direct traffic destined for Subnet C toward Router C-B.

Figure 17. Adding a static route

Router A-Int, meanwhile, can see the computers only on Subnet A and the computer connected upstream at the Internet service provider (ISP). Because the default route directs traffic toward the Internet, traffic destined for Subnets C and B is improperly forwarded to the Internet without static routes instructing Router A-Int to handle this traffic differently. Figure 18 illustrates these routes. Note that the static route for Subnet C can point toward only the neighboring router, Router B-A, which itself does not see Subnet C. This static route, then, relies on the static route configured on router B-A that in turn directs appropriate traffic to Router C-B.

Figure 18. Adding more static

Adding Static Routes

You can add static routes through either the Routing And Remote Access console or a command prompt.

To add a static route in Routing And Remote Access, complete the following steps:

To add a static IP route at a command prompt, open a command prompt and type the following:

route add destination mask netmask gateway metric costmetric if interface

Specifying the metric is optional, and the interface is chosen automatically if you do not include it. For example, to add a static route to the 10.0.0.0 network that uses a netmask of 255.0.0.0 and a gateway of 192.168.0.1, you need only type the following at a command prompt: route add 10.0.0.0 mask 255.0.0.0 192.168.0.1.

Also note the following when using the Route command:

• Static routes added with the Route command are not persistent unless the –p switch is used.

  As in the following example, use the –p switch with the Route command if you want the static route to remain in effect even after the routing computer is rebooted:

  route add –p 10.0.0.0 mask 255.0.0.0 192.168.0.1

• Use the Route Delete command to delete a route that you have added. After route delete, you need only type as much as is necessary to distinguish the route:

  route delete 10.0.0.0

• Static routes that you add by means of the Route command are not listed as static routes in the Routing And Remote Access console. However, you can view them along with other routes when you use the Routing And Remote Access console to view the routing table.

• Interfaces in the Route command are designated by a hexadecimal number, not by the interface address. You can view the number assigned to each interface in the Interface List section of the Route Print output:

  c:\>route print
  IPv4 Route Table
  ===========================================================================
  Interface List
  0x1 ........................... MS TCP Loopback interface
  0x2 ...00 50 ba 40 5c 73 ...... D-Link DFE-530TX+ PCI Adapter
  0x10003 ...00 53 45 00 00 00 ...... WAN (PPP/SLIP) Interface
  0x20004 ...00 53 45 00 00 00 ...... WAN (PPP/SLIP) Interface
  

  To refer to one of these interfaces, you can either type the hexadecimal number as shown (including the “Ox” prefix), or convert the hexadecimal number to a decimal value. Note also that hexadecimal values 0x1-0x9 are equivalent to 1–9 in decimal.

  As a result, the following two routes are equivalent:

  route add 192.168.2.0 mask 255.255.255.0 192.168.1.2 IF 0x2
  route add 192.168.2.0 mask 255.255.255.0 192.168.1.2 IF 2
  

  The following two routes are also equivalent, given that the hexadecimal value 0x10003 is equivalent to the decimal value of 65539:

  route add 207.46.2.0 mask 255.255.255.0 207.46.1.2 IF 0x10003
  route add 207.46.2.0 mask 255.255.255.0 207.46.1.2 IF 65539
  

Advantages of Static Routing

Static routing is advantageous in small networks for which configuring a few static routes is simpler than configuring a dynamic routing protocol.

Besides this ease of deployment, another advantage of static routing is that static routes are less resource-intensive than are dynamic routing protocols. They do not require communication among routers, and this feature makes static routes preferable to dynamic routing protocols for low-bandwidth WAN links.

A final advantage of static routing is that, unlike dynamic routing protocols, static routes provide support for unnumbered connections: connections in which one or both of the connecting logical interfaces (usually in a demand-dial connection) fail to obtain an IP address. Typically, demand-dial routing operates through numbered connections, which means that both the calling and answering router obtain an IP address from each other and assign these IP addresses to the logical endpoints of the point-to-point connection.

When this process fails in Windows Server 2003, the connecting logical interfaces are normally assigned Automatic Private IP Addressing (APIPA) addresses. Unnumbered connections thus occur only when either router does not support APIPA.

Disadvantages of Static Routing

The main disadvantage of static routing is that it is a feasible means of maintaining only small routed networks. As a network grows, the administrative cost of maintaining static routes quickly outweighs the cost of implementing and maintaining a dynamic routing protocol.

A second disadvantage of static routing is the lack of fault tolerance. If a route is improperly configured, connectivity is disrupted until the problem is diagnosed and fixed.

Static Routing Design Considerations

To prevent problems, you should consider the following design issues before you implement static routing.

Peripheral Router Configuration

Peripheral routers by definition have only one neighboring router. Found at the periphery of networks, they connect the outer subnets of an organization to its network backbone. To simplify configuration, you should normally configure a peripheral router with a default route that points to its neighboring router. In Figure 16 shown earlier, Router C-B is a peripheral router. Its default route should therefore point to the neighboring router, Router B-A.

Default Routes and Routing Loops

You should avoid configuring two neighboring routers with default routes that point to each other. A default route passes all traffic that is not on a directly connected network to a specified router. Two routers that have default routes pointing to each other will produce a routing loop for traffic with an unreachable destination.