This page regroups all the network classes I gave with my broken english at ITAM and also in France for the MTIA/RSIE double degree.
The first class we saw Layer 2 bridging with Spanning Tree and also how to enhance bridging scalability with QinQ and MACinMAC. The resulting architecture is a way to isolate bridges in the core network from external network. We will see the same kind of architecture at layer 3 when we will see MPLS and more precisely the interaction between BGP and MPLS.
Then we will see basic things concerning Internet, especially on the addresses, the big change in 1993 when the network went from a flat addressing plan to a hierarchical one with CIDR. We will also see how private addresses can be used and the different kind of NAT used to move from private addresses to a single public address.
We made an exercise on a book shop addressing plan.
In this second part, we continue to explore CIDR addressing scheme and we start introducing IPv6 addresses. We introduce briefly the different kind of addresses and look how global addresses are built.
In the third part, we discuss the different way to build IPv6 Interface Identifier (IID) and look more deeply to IPv6 Link-Local addresses (how to scope them and their use in routing protocols). We also look at multicast addresses and how IPv6 multicast solicited are built. They will be useful when we will study Neighbor Discovery.
In this fourth part (second day of classes), we look at anycast addresses and then we study the IPv6 packet header format. We see how extensions work and that the order is important. We then discuss on the use of extensions in operational network and conclude that except fragmentation and IPsec extensions the others not frequently deployed in production network.
In the fifth part we see ICMPv6 and different protocol using it, such as Path MTU discovery and in more detail Neighbor Discovery Protocol (NDP). We discuss if DHCPv6 are needed to allocate addresses or if we can use only NDP. We also study how to secure NDP.
We then look at routing protocol in general and why they are useful in network. We will see the different king of routing protocols and study RIP, ans star looking at Link State algorithm.
In the sixth part, we look at OSPF and more precisely we look at division into areas and how reliable flooding is done. We also look at database synchronization between routers.
On the seventh part, we see how we can move progressively from RIP to OSPF on a simple example. This illustrate the fact that it is difficult to use two routing protocols since metrics are different.
We then look at IS-IS routing protocol, we see the similarities and differences with OSPF. It is also a good way to introduce OSI vocabulary.
On the heigth part, we see MPLS architecture and the interaction with IGP. We see the concept of LSR (Label Switch Router). We will only study how LSP (Label Switching Path) are built with LDP.
We also start to see Autonomous System and how peering is described in whois database. We also see on a simple example the impact of announcing or not a route to another AS. We conclude that routing is naturally asymmetric in the Internet.
In the ninth part, we study more deeply peering between ASes and we start looking at BGP protocol. We see some basic attributes such as AS_PATH, MED and Local pref.
In the tenth part we continue to explore BGP and see Community Attribute and how we can modify routing strategies from a remote transit AS. We also look at all the way NEXT-HOP is used. We also look at some live example of BGP announcement by surfing on Looking Glass, even if during the class this exercise was not so easy and forced me to edit a lot the video.
Finally the last day we saw VPN with BGP and some way for providers to integrate IPv6 without modifying their core network, or vice versa build an IPv6 core network to carry IPv4 traffic from customer. We also see L3 VPN in details.
We end the class with the following exercise http://c2.touta.in/?p=410
In Rennes, we continue classes by looking at the transition between IPv4 and IPv6. In a first classe, we saw basic classification for transition, or at least integration of IPv6 in existing networks. We viewed v6ops classification and basic mechanism such as tunneling and 6PE.
Last part for integration strategies, we talk about 6rd, CGN and 4rd, we also discuss how IPv6 can be introduced in large companies by the use od proxies and ALG.
For the last class on WSN, the sound was not recorded, so I will just put the slides and some interesting videos of the M2M seminar are also available on line at this address: http://videos.rennes.inria.fr/seminaire_Irisa/Journees-ObjetsCommunicants/