All good things must come to an end, and so it is with Networking 101. This installment, we take a look back at everything we covered in our series.
We designed the series with the belief that everyone in networking should understand all of these concepts, at least within the space constraints we gave ourselves. We wanted to relate information applicable to all parties involved in maintaining or making network decisions, and in such a way that seasoned veterans down where the rubber meets the road could brush up on forgotten basics, and managers could get a sense of what those veterans are talking about.
We began with an overview of IP Addresses. The idea was to provide a basis for the way subnets work by giving readers the tools to understand the features and limitations of 32-bit numbers. We also discussed what the terminology around IP was all about, including the differences between multicast, unicast and broadcast addresses.
We also wanted to provide plain English discussions of the topics everyone hears about, but may not fully understand. In both installments on subnetting, we did a thorough overview of Subnets and CIDR, but the second delved into some Subnetting Examples and an IPv6 overview. The most important take-aways from the IP address and subnet discussion were:
- CIDR IP addresses have a host and network portion. The netmask specifies the number of bits that the network portion uses, and the rest are for the host.
- Subnets are created by the simple act of moving the divider up and down the 32-bit number.
- IPv6 addresses are the same with regards to slicing them into subnets, and as long as you remember the rules of address representation, confusion can be kept to a minimum.
Up the OSI Stack
These articles provided the background necessary to understand all other aspects of networking. The next step began the trip up the OSI stack, with an introduction to how layers work. The OSI model works well, but plenty of network training material fails to mention exactly how these layers work together.
- Unless you’re a router, data coming up the stack is for you, and data going down the stack is being sent by you.
- Layer two data is called a frame, and doesn’t involve IP addresses. IP addresses and packets are layer 3, MAC addresses are layer 2!
The journey up the OSI stack began at Layer 2, the Data Link layer. When managing a network, Layer 2 issues seem to crop up more often than you’d think, so this is a very important section. Likewise, we brought the spanning tree protocol into the loop (pun intended). Spanning tree provides a means to control loops in such a way that allows you to have an Ethernet network that will "fail over" in the case of downed links. Spanning tree is a bit complex, but necessary, and its concepts relate to many routing protocols too.
Before moving on to layer 3, we dedicated an entire article to ICMP, because it lives in-between layers 2 and 3. ICMP is vital to proper routing and packet delivery, and there are many aspects of ICMP that go unnoticed. It isn’t just ping.
Layer 3, IP, began the next part. IP is unreliable. When IP packets are lost it’s up to the higher-level protocols to realize this and request retransmissions. It’s very important to understand IP fragmentation as well, because firewall and network connection decisions can impact the Internet Protocol in strange and unexpected ways.
We covered TCP in two parts: basics and a more in-depth discussion. Most applications use TCP, and troubleshooting TCP sometimes requires looking at packet dumps and figuring out what went wrong. Flow control in TCP can also be impacted by management decisions, so an understanding of congestion control and TCP windows is quite relevant.
The trip up the stack was basically concluded at that point, since Layer 7 applications are just that: random applications. Layers 5 and 6 don’t exist, and just add to the confusion.
Of Governance and Protocols
Routing protocols were next in the list, but first we diverged a bit into a quick talk about how the Internet works, with Internet Governance. ICANN’s role, IANA and RIR roles, and what the IETF and IAB actually do were clarified. Understanding the impact of news items should be a bit easier, once people understand a bit about how the Internet operates.
The routing portions began with a good overview of what routing is, explaining the theory of routing. Routers send packets toward their destination, normally by shipping them toward a router that knows a bit more about the destination topology. It’s important that the decision makers understand the limitations and features between the two types of protocols, link-state and vector-distance.
The most widely used internal routing protocol was examined in two parts, OSPF one and two. The concept of “areas” in OSPF are very important, both from the designer’s and an implementer’s perspective. Very subtle, but serious routing issues can result from a poorly designed OSPF network. Due to the sheer complexity of OSPF, two parts were used to explain it in enough detail to allow for informed architectural decisions to be made.
Internet Routing came in three parts: how routing and peering work, BGP and iBGP. The most important takeaway from this portion was that there is no such thing as a default route in the Internet. BGP operates very differently when compared to other routing protocols, so the first article dealt mostly with the conceptual protocol-level aspects of BGP. The iBGP protocol is simply BGP used internally as a mechanism to exchange BGP information between multiple BGP border routers (on the inside). The iBGP article really glues together the concepts of autonomous systems and BGP routing.
Multicast routing was the final routing topic, then two other miscellaneous topics were added: Understand Tunnels and NAT is not what you think it is. Tunneling itself is sometimes very complicated to conceptualize. Many people struggle when setting up VPN connections, so we felt that some level of discussion about tunnels was required.
And that’s the series that was. Feedback from readers was very positive, and it encouraged us to look at taking Networking 101 to the next level. We hope to see you there.