1.

A.

In layered networking an interface is used to provide service to the layers above
E.g. layer 1 provides a service to transmit bytes over the physical medium

Meanwhile a protocol is used to define rules about the communication on the same layer, but between different hosts
E.g. IP on L3

Diagram is basically the OSI network stack diagram
Encapsulation happens when a packet goes down a layer

B.

C.

Start with /17, subnets need to be /27
This means we get 10 bits to use for the netmask
$2^{10}=1024$, then minus the all zero and all one subnet, we get $1022$

D.

This will not be on the exam

2.

A.

Describe the different kinds of equipment that are being used on the layers 1, 2 and 3 in a network environment and describe their characteristic properties relating to collisions and broadcasts

L1:
Physical layer, e.g. hubs, repeaters, cables
Collisions can occur when multiple devices transmit on the same medium
Broadcasting can be done by hubs
L2:
Ethernet, carried out by a NICs, switches, bridges
Collisions are not a problem due to separate connections and thus collision domains, unlike older hubs
Broadcast can be done to MAC FF:FF:FF:FF:FF:FF, but messages can also be direct
L3:
E.g. IP, carried out by a router
Collisions are a problem for the layer(s) below to solve
Broadcasting can be done to 255.255.255.255

B.

SNAP was not covered

C.

Not really covered, nor relevant for our exam

1. What is the IPv6 prefix for link-local
2. What is the IPv6 prefix for multicast
3. What is the IPv6 prefix for unique local unicast

1. fe80::
2. ff00::/8
3. fc00::/7
4. 2000::/3

D.

Choose a root bridge by:

1. 1. Bridge priority, lower is better, multiples of 4096
2. MAC Address, lower is better
   Priority and MAC concatenated make the bridge ID

Every switch/network chooses a root/designated port through:

1. Lowest cost to the root bridge
2. Neighbour with the lowest bridge ID
3. Lowest port ID

3.

A.

Header question, ignore

B.

Memory question, ignore

C.

D.

1. What is the name in OSPF for a router that is part of multiple areas
2. Name the different OSPF LSAs having type 1-5, mention for each type which type of router originates them

1. Area Border Router (ABR)
2. :
   1. Router LSA, originates from all routers
   2. Network LSA, originates from designated router of that network
   3. Summary LSA, originates from Area Border Router (ABR)
   4. ASBR summary LSA, originates from ABR
   5. External LSA, originates from Autonomous System Border Router (ASBR)

4.

A.

1. Draw a picture of the BGP relationships of an AS with its own providers, peers and customers
2. What filtering should be applied on sessions involved
3. What attribute should be used on incoming routes to distinguish them between the different relationships

1. 
2. You should:
   1. For customers: accept all routes, advertise all routes
   2. For peers: accept their and their customer's routes, advertise your and your customer's routes
   3. Accept all routes, advertise your and your customer's routes
3. To distinguish between relationships, both local preference and communities can be used

B.

1. What does hot potato routing mean and why is this relevant?
2. Which classical attribute is used to prevent your upstream from using hot potato routing? Is it always succesful?
3. What attribute was later introduced to communicate preferences to your upstream or to receive extra routing information from your upstream?

1. Hot potato routing means using the egress point which has the smallest IGP distance. I.e. get the packet out of your network with as short a distance as possible
2. You can use MED to influence this, but it's only a suggestion, you cannot enforce this
3. Communities were introduced to communicate preferred treatment of routes, but they do rely on agreements between ASes on what certain values mean

C.

Was touched upon ultra shortly in the lecture

1. Why is iBGP scaling needed
2. What methods are there and how do they work

1. Routes learned through iBGP are not advertised to other iBGP nodes, therefore you must maintain a full mesh to advertise all routes to all nodes, this is not scalable
2. Solutions
   1. Route reflectors: break split horizon and re-broadcast learned routes anyway
      Hosts only peer with RR, not with one another
   2. Alternatively, split your AS into multiple internal ASes, these are only visible internally
      These internal ASes use eBGP to communicate between one another, meaning no full mesh is needed as eBGP allows re-sharing routes

D.

Not covered