Windows Server 2008: Configuring Routing

Routing Fundamentals

When attempting to select a path in a network by which to send data or physical traffic, an administrator has many options available to him. There are a number of ways to send packets from one destination to another based on intermediary hardware or nodes. This can include a number of different hardware devices including bridges, gateways, routers, firewalls, and switches. Even computers with multiple network cards are capable of routing packets. There are different types of routing algorithms or protocols that can be used to organize the signal flow between these devices.

These algorithms rely on what is called a routing metric, a value used by a routing algorithm to determine whether one route should perform better than another. Metrics can include a number of different parameters to judge performance by, as configured by the administrator.

On the simplest level, the system will select an entry from the routing table and use the netmask from that entry (see Figure 1). The system then performs a comparison of this value and the destination address. The resulting value is cross-referenced to the network address in the table entry. If the two values match, the information can arrive at the destination through the gateway in that entry. If the two values do not match, the routing system continues along the routing table to the next entry and performs the same check again. If the “no matching entry” is found on the table, the routing system discards the packet and generates a message notifying the sender that the destination network cannot be reached.

Figure 1. Routing Tables

Otherwise, when a routing table entry is found that matches the network value, the packet is sent based on the information in the table entry via the destination listed. If the destination exists on a portion of the network directly connected to the routing system, the packet is delivered to the destination system. If it does not exist on the same segment, the packet is sent to a gateway system for delivery. This is a very complicated way of describing what is referred to as static routing.

Static Routing

Static routing describes a system that does not implement adaptive routing in its configuration. In these systems, routes through a network are defined by set paths referred to as static routes, which are inserted into the router manually by the system administrator. This is accomplished via the route command, which can be used to manipulate local routing tables. There is no fault tolerance in regards to static routing. Changes to the network or a failure between two statically defined nodes will cause any traffic between those points to not be rerouted. This means any packets that are awaiting transport between the affected paths will be forced to wait for repairs to the failure, or for an updated static route by the administrator. This also leaves open the issue of the request timing out before repairs can be made to the route.

Static routing is considered the simplest form of routing and requires excessive manual processes. It often is the least efficient way of routing in cases where information paths have to be changed frequently. This is also the case for configurations that require a large number of routing devices, because each one must be manually entered. Static routing is also the least preferred method of dealing with outages or down connections, because any route that is configured manually must be reconfigured manually to fix or repair any lost connectivity.

There may be many downsides to static routing, but there are many incidents where a static route is the most logical and efficient method for routing. Static routing is the opposite of dynamic routing, which is a system in which routers will automatically adjust to changes in network topology or traffic. Dynamic routing is used by most modern routers, but some amount of programming is still available for customizing routes if necessary.

As we mentioned earlier, you as an administrator will need to deal with clients and employees of your company attempting to access the network and Internet. The Internet and Local Area Networks (LANs) are referred to as packet switching networks. The idea of packet switching networks is defined by the ability to optimize the use of the channel capacity available in a network. This helps to minimize transmission latency. This also requires the use of specific protocols for directing traffic through them. There are two major classes of routing protocols used in packet switch networking today:

• Distance-vector Routing Protocol A distance-vector routing protocol requires that a router contact and transmit to its neighbors of topology changes to the network. The frequency of this must be periodic and in most instances when a change is detected. Routing Internet Protocol (RIP) is the most popular example of this type of protocol.

• Link State Protocol The simplest explanation of link-state routing is that every node (router) is given a map of the topology of the network. This map is in graph form and shows the connectivity of nodes in the network. Then each individual node calculates the next best hop from every node in the network. This information then forms the routing table for each individual node based on its calculations. No other communication occurs between nodes. The most popular version of this is the OSPF.

Routing Internet Protocol (RIP)

The RIP was once the most commonly used Interior Gateway Protocol (IGP) on internal networks. It was also commonly used on networks connected to the Internet. RIP was used to help routers dynamically adapt to the variety of changes made to network connections. It accomplished this by relaying information about which networks each router had access to, and the distance those networks were from each other.

Although RIP is still actively used and has an important place in some networks, it is generally considered a dying protocol, which has been replaced by other routing protocols such as OSPF. RIP is a distance vector routing protocol that employs the hop count as a routing metric. RIP allows a maximum of 15 hops. The total hold down time for transfer is 180 seconds. Most traffic at the time RIP was commonly used was not significant, so each RIP router had an update time of 30 seconds by default, which was common practice. This proved to be a poor configuration and was later changed to randomized updates.

RIP is limited in a number of ways due to its lack of scalability. It prevents routing loops from continuing indefinitely, by implementing a limit on the number of hops allowed in a path from the source to a destination. It also limits the size of the network that RIP can support by design.

On the other hand, RIP is easier to configure than many other protocols, because it uses one of the smallest amounts of settings of any routing protocols. RIP does not require the use of any parameters on a router, and it can be ideal for small networks. RIP can be configured through the RRAS, which we will discuss later.

Open Shortest Path First (OSPF)

OSPF was the natural successor to the RIP. OSPF protocol is a hierarchical IGP that uses a link state in the individual areas that make up the hierarchy. A link state database (LSDB) creates a tree-image of the network topology. It then sends copies of the LSDB periodically to update all routers in the area of the OSPF network.

OSPF is the most widely used IGP in regards to large enterprise networks. It has a much larger network size range than RIP. The OSPF protocol can determine the best path by communicating with other routers and then saving the routes in their LSDBs securely.

An OSPF network is divided into areas, which contain area identifiers. These identifiers are 32-bit and are usually written in the format of an IP address. Be aware that area identifiers are not IP addresses, and may often times duplicate any IP address without conflict occurring. These areas are logical groupings of routers whose information may be communicated to the rest of the network. There are several types of areas in an OPSPF network:

• Backbone Area The backbone area forms the central hub of an OSPF network. All other areas are connected to it, and inter-area routing happens via routers connected to the backbone area and to their own non-backbone areas. The backbone area distributes all routing information between the non-backbone areas. The backbone must be adjacent to all other areas, but does not need to be physically contiguous. Connectivity can be established and maintained through virtual links. All OSPF areas must connect to the backbone area. This connection, however, can be through a virtual link.

• Stub Area The stub area is an area that does not receive external routes except the default route, but does receive inter-area routes. All routers in the area need to agree they are stub, so that they do not generate types of LSA not appropriate to a stub area. Stub areas do not have the transit attribute and thus cannot be traversed by a virtual link.

• Not-so-stubby area (NSSA) The Not-so-stubby area (NSSA) is a type of stub area that can import autonomous system (AS) external routes and send them to the backbone, but cannot receive AS external routes from the backbone or other areas. The NSSA is a non-proprietary extension of the existing stub area feature, which allows the injection of external routes in a limited fashion into the stub area.