2016
1. Theory
A.
I. (4p)
What is the difference between an interface and a protocol?
Same as exam 2015
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
II. (2p)
How are they used for network communication?
Protocols and interface enable separation of concerns, where each layer focuses mainly on one responsibility
When going down, data is encapsulated, e.g. L3 IP gets wrapped in L2 Ethernet packet
When going up the layers, data is de-capsulated again
E.g. layer 1: actual transmission
L2: Peer-to-peer routing
B. (6p)
Same as 2015#B.
C.
I.
Name four different kinds of IPv6 Addresses inside which an IPv4 address can be embedded
I don't expect this kind of question to be asked
- IPv4 compatible IPv6 addresses:
::xxxx:xxxx
So an IPv6 small enough to also be used as an IPv4 - Mapped addresses:
::ffff:xxxx:xxxx - ?
- ?
II.
For the purpose of IPv4/IPv6 algorithmic translation, a service provider has set aside the IPv6 prefix 2016:0322:deed::/48. What is the resulting IPv6 address when embedding the IPv4 address 201.60.32.2
201 -> C9
60 -> 3C
32 -> 20
2 -> 02
2016:0322:deed:0000:C93C:2002
D.
Consider the subnet of the IPv6 space given by the union of the prefixes
::/28000::/24000::/36000::/3c000::/4d000::/4
Describe this same subnet as the union of prefixes with the maximum possible aggregation
Not a clue tbh
2. Link Layer
A.
I.
Not relevant
II.
What does VLAN abbreviate and in what specification is it defined?
VLAN stands for for Virtual Local Area Network
It is defined in 802.1q
III.
This type of question will not be on the exam
B.
I.
Give an example of an L2 MAC
How can it be used in IPv6 for the local link
Example: E8-9C-25-7D-EC-6E
Can be used in SLAAC to generate the bit of the address after the link local prefix
If duplicate address detection finds this address to be free, it can then be used
II.
What is the equivalent for ARP in IPv6
The equivalent in IPv6 is the Neighbour Discovery protocol, which is built on ICMPv6
C.
I.
What is the IPv6 prefix used for link local addressing
fe80::/10
II.
What is the IPv6 prefix used for unique local unicast addressing
fc00::/8
This was not in any of the lectures
III.
What is the IPv6 prefix used for multicast
ff00::/8
IV.
How is IPv6 multicast traffic restricted to the link layer?
Not covered in the lectures
D.
See STP & RSTP
3. Distance Vector and Link State Routing
A.
What is the fundamental difference between RIP and OSPF as ar as the theoretical foundation of these routing protocols is concerned
RIP:
- Distance Vector, based on Bellman-Ford
- Routers only know hops as metric
- Each router has a limited view of the total topology, only knows distance and next hop, not full path
OSPF: - Link state routing, based on Dijkstra
- Router use bandwidth as metric
- Each router has a complete, consistent view of the topology
B.
What are the equation for the centralised Bellman-Ford algorithm?
Not relevant
C.
Do Dijkstra
D.
What is the difference between a neighbour and an adjacency in OSPF
A neighbour is a directly connected OSPF speaker, discovered via Hello packets
An adjacency is formed after both parties have agreed to and successfully finished exchanging LSDB information
4. Path Vector Routing and Internet eXchanges
A.
I.
What are the usual import and export policies of ISPs with respect to prefixes from providers, peers, customers and the ISP itself?
Customer: advertise & accept everything
Peer & provider: advertise your own and your customer's routes, but not those learned from peers or providers, still accept everything, while giving preference to routes from your customers, so we do not take a detour
II.
If a certain AS has 4 external BGP peerings, what can you say about the number of internal BGP peerings? Suppose no use is made of router reflectors or other optimisations
Assuming the question means that an AS has 4 bgp edge routers
Since BGP requires a full internal mesh, since routes learned through iBGP do not get forwarded, this would mean it needs to have
The end result would be 6 internal peerings
B.
What does hot potato routing mean and why is this relevant?
Hot potato routing means using the BGP egress point which has the lowest IGP distance to the source
It is relevant as using hot potato routing can also (indirectly) burn you, by forcing another AS to use a non-preferred (slower) link
This can be remedied by setting MED, and having the other party respect it (since you cannot force them)
C.
I.
What is the difference between a route reflector and a route server?
A route reflector sits within an AS, making BGP more scalable by breaking the iBGP split horizon to remove the need for a full mesh
A route server on the other hand lives at an Internet eXchange
It only does eBGP and, similarly to a route reflector, removes the need for a full mesh, which in this case means not having to peer with each AS at the IX
Instead, you peer with the route server, through which (according to your policies) prefixes are shared and received
Important note: the route server does NOT carry traffic, it only shares prefixes
D.
The AMS-IX is supposed to be a layer 2 platform. Explain how it is possible the first thing you connect to on the AMS-IX is a router
AMS-IX operates as one large ethernet domain
So the first thing you connect is YOUR router to THEIR switch