Interior Gateway Protcol (IGP) 1/6: Introduction

This is the first post in a 6 part series covering various interior gateway protocols(IGP). Over the next 6 posts we will touch on the most used routing protocols and their pros and cons. The goal is to help you better understand routing and allow you to pick the protocol that is best suited to your network.

First of all, what is an interior gateway protocol(IGP)?

IGP is a routing protocol used to exchange data between nodes within an autonomous system.

Interior gateway protocols fall under 2 categories. Distance-vector routing protocol and Link-state routing protocol.

Distance-vector routing protocol

In distance-vector routing protocols, routers do not have information about a network topology in it’s entirety. Routers only have the details of the neighboring routers and they rely on those routers to route packets efficiently based on their neighbors. Distance-vector protocols use the Bellman-Ford algorithm to determine the best route.

The 3 distance-vector routing protocols we will be going over the next few posts are as follows:

• Routing Information Protocol (RIP)
• Interior Gateway Routing Protocol (IGRP)
• Enhanced Interior Gateway Routing Protocol (EIGRP)

Link-state routing protocol

Opposite of distance-vector protocols, link-state protocols contain information about the entire network topology. Each individual router is able to calculate the best route using the information contained in their local routing table.

The 2 link-state protocols we will look at are as follows:

• Open Shortest Path First (OSPF)
• Intermediate System to Intermediate System (IS-IS)

NetBIOS browsing across subnets not working after PDC upgrade to Windows 2008.

We recently upgraded a Windows 2003 primary domain controller to Windows 2008. Following the upgrade we noticed that NetBIOS browsing across subnets was no longer working. The only hosts that were able to be seen were located in the local subnet. This created some issues for some of the non-networking IT staff. While hosts located on the distant subnets could still be reached via hostname they weren’t appearing in the ‘network’ portion of the Windows explorer.

A brief inspection of the running services on the newly upgraded Windows 2008 primary domain controller showed that the ‘Computer Browser’ service was disabled. Apparently, this service is disabled by default in Windows 2008 Server. Setting the service to run automatically and then starting it resolved the browsing issues.

If you are having similar issues after upgrading your domain controllers, we recommend checking the ‘Computer Browser’ service. Ideally, you would probably want this service running on all of your domain controllers to ensure NetBIOS browsing is possible from any of the subnets on your network.

MPLS Redundancy solutions

A popular network solution a lot of companies who have branch offices are using is MPLS (Multiprotocol Label Switching). There are several MPLS redundancy solutions ranging from costly to cost effective. We won’t go into the technical details of setting it up, but more or less explain the solutions out there.

The ideal solution would be setting up a complete MPLS network using another service provider. This is probably the most costly solution however will give you the greatest results.

Another solution is to have point to point connections to the branch offices. This can be costly as well depending on how many branch offices there are and what other offices they need to have connectivity to.

The most cost effect solution will only give marginal performance depending on the load. You can set up an IPSEC VPN between offices using your typical broadband connections (cable, DSL). I’m actually currently in the process of setting this up for a small office that needs connectivity to a main office in the event that their MPLS link goes down.

Automating the fail-over process for these solutions should be no different than doing it for a dedicated circuit. There are multiple ways of going about it. Dynamic routing, load balancing, and setting different metrics for the 2 different routes are a few ways to do it.

This was just to give you an idea of what types of solutions are out there. In a later post, I will go into detail about the equipment that can be used and some example configurations of the scenarios given the different equipment.

Introduction to the OSI model

Knowledge of the OSI model is crucial in network troubleshooting. Not only is it important to have a good grasp on the OSI model for network troubleshooting, but also if you plan on pursuing any of the Cisco certifications. The OSI model will provide you with an excellent foundation on troubleshooting, giving you a starting point, working your way through the model to find issues in network connectivity.

The Open Systems Interconnection (OSI) model consists of seven layers of network architecture. These layers are the Application, Presentation, Session, Transport, Network, Data-Link, and Physical layers. For the most part if you’re a network administrator, like me, most of your work will be concerned with the first three layers, Physical, Data-Link, and Network.

Now let’s break down the different OSI model layers and what they do:

Layer 7 – Application Layer

The Application layer is generally the closest to the end user. This is where the end user will interact with software in order to communicate across a network. General examples of this layer are telnet,  HTTP, FTP, SMTP.

Layer 6 – Presentation Layer

The Presentation layer is where the data from the Application layer is taken and encapsulated into a form of data that can be used for travel across a network. This layer converts data for use in the Application layer or the Session layer for use in the opposing layer.

Layer 5 – Session Layer

The Session Layer maintains communications or connections between nodes on a network. It establishes, maintains, and terminates. It can operate in full-duplex, half-duplex, or simplex operating modes. Any checkpointing or recovery operations in TCP usually happen on this layer. An example of this layer in action would be an application that uses RPC to execute actions in another address space remotely.

Layer 4 – Transport Layer

The Transport layer is where reliability is controlled through flow control, (de)segmentation, and error control. This is where all data is transfered between end users. Error checking happens on this layer and it can retransmit any failures. Examples of this implementations of this layer are TCP and UDP.

Layer 3 – Network Layer

The Network layer, my favorite, is where all routing happens on a network. This is where data is transferred to it’s destination across one or more networks. This is where routers operate. Data travels across this layer alot like traveling the highway system. Data flows between different routers (hops) to finally make it to it’s destination. The most known implementation of the Network layer is Internet Protocol or IP, as it is most called.

Layer 2 – Data-Link Layer

The Data-Link layer is where data is transfered between nodes on a network. This layer has the capability to detect and correct errors that might have occured on the Physical layer. One of the most well known implementations of this layer is Ethernet. You will find most network switches on this layer. The Data-Link layer provides a connection across the physical link by using vendor assigned hardware MAC address verses assigned IP addresses, like the Network layer does.

Layer 1 – Physical Layer

The Physical layer is where all electrical and physical connections are made. On this layer you will find cabling, hubs, and, network adapters to name a few. Where the Data-Link layer connects multiple nodes, the Physical layer is mostly concerned with connected a single device to the network medium.

That’s a rough introduction to the OSI model. There is a lot more theory of it, but I just wanted to give a quick overview.

All your troubleshooting efforts should start at layer 1, what I like to call ‘checking the basics’. Bypassing layer 1 can cost you alot of time in troubleshooting. I generally make sure cables are plugged in and all devices are powered on. A lot of times you may find a simple easy to correct issue here and save yourself alot of grief in troubleshooting the higher level layers. From there you can just work your way up the layers. This will provide good coverage of possible failures or misconfigurations across the network.

Some network administrators, including my self, talk of a ‘Layer 8′. This is the actual end user themself. I like to call this ‘operator head space and timing’. You should be able to judge and categorize your end users by level of knowledge if you’ve been working with them for some period of time. You can usually skip to ‘Layer 8′ if you’re able to identify the issue with the end user from their complaint. This of course is not part of the Cisco certification process and has more to do with experience in dealing with end users, which you will learn on the job.

How to configure DHCP on a Cisco Router

I prefer running DHCP on a server within the network, but if it comes to it you do have the option of configuring DHCP on a Cisco router. I do believe some of the newer Cisco Catalyst switches have this ability too.

We are going to assume you have a Cisco router already setup and in service on your network.

First thing you will want to do is login to your router and go into config mode. Once in config mode you will want to setup a DHCP pool.

Router(config)# ip dhcp pool newpool

You can replace ‘newpool’ with whatever you choose to name your DHCP pool. Next thing we want to do is tell the router the network and subnet.

Router(dhcp-config)# network x.x.x.x y.y.y.y

Replace x.x.x.x with the network address and y.y.y.y with the subnet mask. Next we will tell the router what DNS domain name the clients will use.

Router(dhcp-config)# domain-name domain.com

Obviously, you will replace ‘domain.com’ with your domain. Now we need to tell the router the primary and secondary DNS server IP addresses to be used in the DHCP scope.

Router(dhcp-config)# dns-server 1.1.1.1 2.2.2.2

Replace 1.1.1.1 with your primary DNS server and 2.2.2.2 with your secondary DNS server. We need to specify the default gateway or router.

Router(dhcp-config)# default-router x.x.x.x

Replace x.x.x.x with your default gateway or router IP address. Next we want to specify the length of the lease of the addresses assigned by the DHCP server.

Router(dhcp-config)# lease 7

Here I have set the lease time to expire in 7 days. Alternatively, you can use the syntax ‘lease DAYS HOURS MINUTES’, replacing those variables with the obvious requirements. You can even go as far as setting it for an infinite perios of time using ‘lease infinite’.

This pretty much sums up setting up DHCP on a Cisco router with IOS. I do want to show you one more thing. Suppose you want to exclude some IP address that you want to reserve for static devices. The following command needs to be performed in the the global configuration mode, so type ‘exit’ and perform the following if you are still at the last step.

Router(config)# ip dhcp excluded-address x.x.x.x y.y.y.y

You will want to replace y.y.y.y with the start address of the range you want to exclude and replace y.y.y.y with the end address. Example: ‘ip dhcp excluded-address 192.168.100.0 192.168.100.100. This will exclude addresses all the way up to 192.168.100.100. Therefore, DHCP will start assigning IP address leases at 192.168.100.101.

Now that we have setup our DHCP scope we can enable to DHCP service.

Router(config)# service dhcp

Now you should be able to grab an IP address via DHCP. Don’t forget to save your configuration.