How to Install and Configure Routers and Switches

Introduction

In this lesson you will learn about aspects of routing and switching when configuring a network. The options and features discussed provide for effective management of a local area network (LAN).

Lesson Objectives

By the end of this lesson, you will be able to:

Identifying the steps involved with installing a network.
Describe the difference in managed and unmanaged network switches.
Describe the options available when configuring routers and switches.
Troubleshoot common router and switch problems.

Network Installation and Configuration: Planning

A network administrator must prepare before setting up a network. The following questions are important to consider and will help map out a plan.

How many hosts will be connected?
How many subnets will be needed?
How will data be forwarded within the network?
How to configure all objects to communicate correctly?
What media will be used?
What services are most important (voice, data, video)?
What is the scope and scalability of the network?
What will the IP address scheme be?

Network Switch

A switch provides for a more efficiently run network. It forwards data to the port on the switch that the destination system is attached to, unlike a hub. A switch prevents loops by using a Spanning Tree Protocol (STP).

Switches filter traffic after updating the MAC table and increases performance and security.

Switches store MAC addresses for each device connected to its ports. Switches have one broadcast domain and multiple collision domains. If a switch does not have the MAC address in its table it floods the switch, sending the data packet to every port.
Spanning Tree Protocol (STP) – places a block on the port that is causing a loop and releases the block when one of the connections (links) goes down.
Hubs send data to all ports. Hubs have one broadcast and one collision domain.

An unmanaged switch allows connected devices to communicate with each other. A managed switch offers advanced features including all capabilities of unmanaged switches. A managed switch allows a network administrator to manage, configure, monitor and diagnose the health of a local area network (LAN).

Advanced features of a managed switch include:

Port mirroring
VLANs
Quality of Service (QoS)
Redundancy

SNMP allows for a Network Administrator to run diagnostics on the switch.

Process Packets

A switch has two ways in which it can process packets. Store-and-forward is when a data packet is analyzed while in memory. If the data packet is error free, it is forwarded to the destination port. Cut-through is when data packets are forwarded straight through with no analysis.

Store-and-forward data packets are analyzed while in memory. If data packet is error free it is forwarded to the destination port.
Cut through data packets are forwarded straight to the destination port with no analysis.

Collision Domain, Port Mirroring

A switch has multiple collision domains and a single broadcast domain.

A broadcast domain sends data to all ports on a switch or hub. A collision domain is where multiple devices attempt to access the media at the same time.

Port mirroring creates a copy of data traffic sent to a port for analysis and diagnosis. Port mirroring can be used for network monitoring.

Each port has its own collision domain. MAC filtering specifies which host machines are allowed to send data packets through a given port on the switch.

Power over Ethernet (PoE) and Trunking

Each port can be configured to function at a specified speed or can be auto-negotiated. Typically ports on a switch can run at 10/100 Mbps. Each port can be Full or Half duplex. Gigabit switches can operate at 1000Mbps and Full duplex.

Power over Ethernet (PoE) enables the switch to supply power to a port connected, PoE capable device (i.e. VoIP phones). Gigabit switches can operate at 1000 Mbps and Full duplex.
Trunking is like bonding, in that multiple ports on a switch are joined to increase the available bandwidth between it and other network devices.

Virtual Local Area Networks

Virtual Local Area Networks (VLANs) can be configured on a network switch to segment or join hosts on the same or different networks.

A switch can be configured to use various ports as a separate LAN or VLAN within the same network segment.
A VLAN can be configured on switches in other LANs, joining the hosts on disjoint networks.
Ports can be trunked (joined) to establish greater bandwidth and link speed between switches.

Hosts that are on separate VLANs cannot communicate with one another; they have different broadcast domains. VLANs give a network administrator the ability to separate one switch into several networks. There are three types of VLANs:

Port-Based – specific ports are assigned to be on one VLAN.
Tag-Based – places a tag on the data packet to determine the destination VLAN.
Protocol-Based – data packets are directed to certain ports based on the protocol being used.

Tag-Based VLANs

Tag-Based VLANs use two protocols:

802.1Q is the IEEE standard for VLAN tagging. This standard allows for switches made by various manufactures to process VLAN tagging.
Inter – switch link (ISL) used to tag data packets for the destination VLAN (Cisco proprietary protocol)
Virtual Trunking Protocol (VTP) is used to propagate changes made to the VLAN definition on one switch to all switches (in a VTP domain) on the LAN (Cisco proprietary protocol). VTP configuration provides the convenience of not having to make configuration changes on each switch.

Quality of Service (QoS)

QoS allows the ability to prioritize which data type (voice, data, or video) takes precedence on the network. Assigning priority to data types ensures available bandwidth for high – bandwidth traffic. Often the QoS for VOIP communications is affected when its priority is low.

The quality of service is better when communications function as expected. Voice and video require greater bandwidth to transmit information. Voice communication can sound muffled and choppy if the QoS is not assigned.

Basics of Routers

Routers provide the ability to connect two separate networks. Routers will not accept or forward broadcast addresses. They will forward data to other routers until the destination is reached.

A switch will pass data packets to a router for destination addresses not belonging to the current network segment. A router will filter and forward data traffic.

The router is usually located on the customer or client side and on the ISP/Telco side. Routers can be configured with different types of interfaces such as Ethernet, fiber, and serial. A router scans the network portion of an IP address to determine whether or not to forward the packet internally or externally.

Network Address Translation (NAT) is an internal operation in the router that translates an internal private IP address into the public address required for sending data out onto the internet.

Port Address Translation (PAT) uses overloading where all private host IP addresses are translated to one public IP address. The router determines where to send traffic based on the port the data was sent from.

The router stores the IP address and port number in a NAT lookup table. Routers use a routing table to determine where to forward data packets. A physical connection (fiber or copper wire) is used to establish communication link between routers. Routers use protocols to send and update routes between one another.

For an example view the following: Displaying the Routing Table.

Routing Protocols

Routers can be located inside of a network and on the edge or perimeter of a network.

Interior Gateway Protocol (IGP) used to transfer routing information within an Autonomous System (AS).
Autonomous System a group of systems with a common attributes (i.e. all network host within a domain or all people with the last name “Smith”).
Border Gateway Protocol (BGP) is used to interconnect Autonomous Systems (AS).

BGP replaced Exterior Gateway Protocol (EGP) to allow for decentralization. Prior to the use of BGP, EGP was a system that used stub gateways to connect to ARPANET core.

Static vs. Dynamic Routing

Static routing requires a network administrator to manually enter the routing table for all routers within their network. In dynamic routing the routers communicate with each other to update their own tables and handle change better.

Three classes of dynamic routing include:

Distance vector builds a list of neighboring routers and sends entire table to neighbors. Each router assigns a distance vector or hop count to each route.
Link state builds list of neighboring routers and each router sends status updates when there is a change in the routing tables. “Hello” data packets are sent amongst all routers to verify connections.
Hybrid provides the benefits of link state and distance vector protocols.

Interior Gateway Protocols (IGPs) can be either link state or distance vector.

Routing Metrics

Routers use various metrics to determine the best route to forward data packets:

Hop counts occur as data passes through a router to its destination. Each Hop represents data reaching another router so the next hop is the next router in the path, to a given destination.
Bandwidth maximum transmission unit (MTU) is the measurement of data capacity between router links.
Costs assign value to distance, bandwidth, time and price of communication links.
Latency network time delay when sending data packets to destination.
Load measures of activity taking place on a router.

Routing Information Protocol

It is important to be familiar with the following terms.

Routing Information Protocol (RIP) is a dynamic routing, distance vector protocol; in which neighboring routers send their entire routing table to one another every 30 seconds (by default).
RIP is only used in networks where all routers and LANs have the same class network address (classful addressing). The maximum hop count is 16 in RIP so it is good for small networks.
RIPv2 provides for classless addressing and Variable Length Subnet Mask (VLSM). RIP does not send the subnet mask for networks while RIPv2 will send the subnet mask with each network ID in common.
Border Gateway Protocol (BGP) is a dynamic routing, distance vector protocol, in which routers from different autonomous systems (AS) share routing tables.
Interior Gateway Routing Protocol (IGRP) is a dynamic routing, distance vector protocol created by Cisco, in which routers with the same AS number share their routing tables every 90 seconds by default. An IGRP is an improvement over RIP and RIPv2. IGRPs have an increased hop count of 255, are classful (no VLSM), include holddowns, and split horizons. Holdowns prevent invalid routes from being reinstated and split horizons prevent information about a route from being sent back the same way that it came.
Open Shortest Path First (OSPF) is a dynamic routing, link state protocol; in which routers initially flood each other with link state advertisements (LSAs). After this initial flood, route updates occur only when there is a change. This routing type requires calculation when the route changes, allows VLSM, and splits network into “areas” for improved management.
IS – IS – is a dynamic routing, link state protocol created by DEC, responsible for keeping tables up to date.
Enhanced IGRP (EIGRP) is a dynamic routing, hybrid protocol created by Cisco, which builds upon IGRP using the following features:
Neighbor Discovery Recovery is used to identify other routers on a directly attached network. “Hello” packets are used to achieve discovery.
Reliable Transport Protocol is used to guarantee delivery of EIGRP communications between routers.
DUAL Finite State Machines tracks all link state advertisements from neighboring routers and is used to prevent looping.
Protocol Dependent Modules works to handle network layer, protocol requirements. (i.e. extracts and formats information from EIGRP packet to be processed by other components.)

Convergence

The goal of all routing protocols is convergence. Convergence is when all routing information has been stored and all routers agree on the best route in which to forward data packets.

Convergence rankings from best to worst include: Enhanced Interior Gateway Routing Protocol (EIGRP), Open Shortest Path First (OSPF), Interior Gateway Routing Protocol (IGRP), and Routing Information Protocol (RIP).

Another important factor in convergence is Administrative Distance. It is the number assigned to the protocol being used. The lower the administrative number the more reliable the protocol.

Router/Switch Troubleshooting

Review the following troubleshooting issues that may occur with routers/switches.

A switching loop that is not blocked via (STP) results in data that cannot be forwarded. Switching loops can result in flooding, where the data is broadcast to all ports continually.
Bad or improper cables connected to a switch or router may result in a bad link, incapable of transmitting data.
Ports and interfaces that are incorrectly configured are incapable of transmitting data.
Switches that have ports configured as VLANs should be assigned correctly in order for the systems/host on those VLANs to communicate.
A power failure will result in bringing a network down until power can be restored via electric generator or restoration of service.
A router with a bad or missing route can result in delayed or non – delivery of data.
Mismatched MTU and MUT black holes occur when data packets exceed the default or modified MTU for a network. When sending data packets through networks an MTU is established, if the data packet is not fragmented it will not be passed through to its destination. When a router processes a data packet that exceeds the MTU for a network it will send an ICMP “destination unreachable” message back to the sending host. If the message is not sent back to the sending host or is blocked by a firewall on the sender’s network, the data packet may be dropped. This causes what is called a black hole.
Bad modules that covert fiber – optic signal into electrical signal can also cause data loss. (i.e. Gigabit interface converter (GBIC) and Small Form factor Pluggable (SFP) are CISCO router plugins used to convert signals.)
Wrong subnet mask can cause miscommunication and prevent the discovery of a network host.
Wrong gateway address will prevent data from being sent between networks.
Duplicate IP addresses will prevent host from accessing the network.
When having trouble resolving a host name (i.e. server1.collin.edu) on a network, DNS may be down or configured incorrectly.

Summary

In this lesson you learned that when installing a network, all network requirements should be reviewed, and then compared to equipment and environmental limitations.

You also gained an understanding about switching and routing and how they play a vital role in making sure data gets from source to destination. Proper configuration of switches and routers enable efficient network performance.

Lastly, troubleshooting common issues within a network help guide and narrow down where an issue is located. For further learning on routing review: Internetworking Technology Handbook and Connected: An Internet Encyclopedia Routing.