Factors to Consider in Choosing a Switch
Although there are many factors that must be considered when selecting a switch, the next topic will explore two: cost and interface characteristics.
Factors to Consider in Choosing a Router
When selecting a router, we need to match the characteristics of the router to its purpose. Similar to the switch, cost and interface types and speeds must be considered as well. Additional factors for choosing a router include:
Types of Media
Choosing the cables necessary to make a successful LAN or WAN connection requires consideration of the different media types. As you recall, there are many different Physical layer implementations that support multiple media types:
Each media type has its advantages and disadvantages. Some of the factors to consider are:
Types of Interfaces
In an Ethernet LAN, devices use one of two types of UTP interfaces - MDI or MDIX.
The MDI (media-dependent interface) uses the normal Ethernet pinout. Pins 1 and 2 are used for transmitting and pins 3 and 6 are used for receiving. Devices such as computers, servers, or routers will have MDI connections.
The devices that provide LAN connectivity - usually hubs or switches - typically use MDIX (media-dependent interface, crossover) connections. The MDIX cables swap the transmit pairs internally. This swapping allows the end devices to be connected to the hub or switch using a straight-through cable.
Typically, when connecting different types of devices, use a straight-through cable. And when connecting the same type of device, use a crossover cable.
Many devices allow the UTP Ethernet port to be set to MDI or MDIX. This can be done in one of three ways, depending on the features of the device:
Data Communications Equipment and Data Terminal Equipment
The following terms describe the types of devices that maintain the link between a sending and a receiving device:
Data Communications Equipment (DCE) - A device that supplies the clocking services to another device. Typically, this device is at the WAN access provider end of the link.
Data Circuit-Terminal Equipment (DTE) - A device that receives clocking services from another device and adjusts accordingly. Typically, this device is at the WAN customer or user end of the link.
If a serial connection is made directly to a service provider or to a device that provides signal clocking such as a channel service unit/data service unit (CSU/DSU), the router is considered to be data terminal equipment (DTE) and will use a DTE serial cable.
Be aware that there will be occasions, especially in our labs, when the local router is required to provide the clock rate and will therefore use a data communications equipment (DCE) cable.
DCEs and DTEs are used in WAN connections. The communication via a WAN connection is maintained by providing a clock rate that is acceptable to both the sending and the receiving device. In most cases, the telco or ISP provides the clocking service that synchronizes the transmitted signal.
For example, if a device connected via a WAN link is sending its signal at 1.544 Mbps, each receiving device must use a clock, sending out a sample signal every 1/1,544,000th of a second. The timing in this case is extremely short. The devices must be able to synchronize to the signal that is sent and received very quickly.
By assigning a clock rate to the router, the timing is set. This allows a router to adjust the speed of its communication operations, thereby synchronizing with the devices connected to it.
A CSU/DSU is a device used to make the physical connection between data networks and WAN provider's circuits.
To develop an addressing scheme for a network, start with determining the total number of hosts. Consider every device that will require an IP address, now and in the future.
The end devices requiring an IP address include:
Network devices requiring an IP address include:
Network devices requiring an IP address for management include:
There may be other devices on a network requiring an IP address. Add them to this list and estimate how many addresses will be needed to account for growth in the network as more devices are added.
Once the total number of hosts - current and future - has been determined, consider the range of addresses available and where they fit within the given network address.
Next, determine if all hosts will be part of the same network, or whether the network as a whole will be divided into separate subnets.
Recall that the number of hosts on one network or subnet is calculated using the formula 2 to the nth power minus (2^n - 2), where n is the number of bits available as host bits. Recall also that we subtract two addresses - the network address and the network broadcast address - cannot be assigned to hosts.
There are many reasons to divide a network into subnets:
To assist troubleshooting and expedite adding new hosts to the network, use addresses that fit a common pattern across all subnets. Each of these different device types should be allocated to a logical block of addresses within the address range of the network.
Some of the different categories for hosts are:
For example, when allocating an IP address to a router interface that is the gateway for a LAN, it is common practice to use the first (lowest) or last (highest) address within the subnet range. This consistent approach aids in configuration and troubleshooting.
Similarly, when assigning addresses to devices that manage other devices, using a consistent pattern within a subnet makes these addresses easily recognizable. For example, in the figure, addresses with 64 - 127 in the octets always represent the general users. A network administrator monitoring or adding security can do so for all addresses ending in these values.
Roll over the device groupings in the figure for an example of how to allocate addresses based on device categories.
In addition, remember to document your IP addressing scheme on paper. This will be an important aid in troubleshooting and evolving the network.