Title: Beyond Buzzwords: Intrinsic Network Concepts

Abstract:

Network protocol design is not a nice, clean science, where what gets deployed is always the best possible design. Instead, designs are influenced by issues such as politics, general confusion, and backward compatibility. Statements get made, and repeated, until it never occurs to anyone to question whether they're true. Mistakes get made, and rather than backing up and fixing them, kludges are introduced to make things sort of work. This talk introduces people to the fascinating and frustrating field of network protocols. The focus of this tutorial is mostly on the lower layers, but the main message is how to reason about networks. For example, a switch makes a forwarding decision by examining fields in the packet header and looking up information in a forwarding table. What are the tradeoffs between having the forwarding table indexed by destination address, "label" (a field that changes at each hop) or "flow"? What are the tradeoffs between having the forwarding table calculated remotely by a central node vs calculated by the switches through a distributed algorithm?