Lecture 1: 24th September 2019 Flashcards
IP Mobility
What is interworking?
Moving between different networks which provide Internet connections.
Why does IP need to enable interworking today?
There are multiple different types of networks which people use, particularly on mobile devices as they physically move, which people should be able to use and seamlessly switch between.
Why was IP not designed to allow for mobility?
There was originally only one network to which computers connected, and all computers were static so would only ever connect to one network.
What is IP mobility?
Mechanisms that allow a device connected to the Internet to move freely between different IP networks (possibly with different underlying technologies) while maintaining IP connectivity in a transparent way to upper layers.
How are data flows uniquely identified?
With IP addresses that resolve a point of attachment to the Internet network, and port numbers that provide resolution within a node.
What is the entanglement of network layers with IP? Why is it a problem?
Protocols at each layer of the 5 layer model all use IP addresses in naming which removes modularity and the abstraction we aim to achieve.
What happens to a TCP connection when the IP or port of a host changes? When might this happen?
It is destroyed as they are identified by the IP addresses and port numbers of both hosts. This can happen when moving between different IP networks.
What is Mobile IP?
An IETF-standardised solution to the IP mobility problem. It is designed to allow mobile device users to move from one network to another while maintaining a permanent IP address.
How does Mobile IP work?
Mobile hosts (MHes) have a starting home network (HN) and a home agent (HA) and want to move to a foreign network (FN) with a foreign agent (FA). An agent acts like a proxy and forwards packets to the MH. When an MH moves to an FN, its HA finds the FA with advertising/solicitation. The MH completes registration with the FN using DAD. The MH then updates its HA with its new care-of-address (CoA), which points to the FA. Any packets for MH that go to the HN are forwarded by the HA to the FA, which de-encapsulates them and forwards them to the MH. The MH only keeps the last HA and HN as well as its FN and FA: a max of 1 proxy.
What are some problems with Mobile IP?
It has a number of security issues, such as enabling reflectin attacks since you can spoof source IDs. End-to-end semantics are lost with proxy tunnelling. Proxies also offer a performance bottleneck, single point of failure and attack. They may also allow traffic monitoring and MH tracking.
How does Mobile IPv6 improve on Mobile IP?
Hosts can effectively be their own agents, authentication and privacy are improved, Binding Updates (BUs) to send CoA to remote end systems improve address and route optimisation, and SLAAC allows nodes to have their own addresses.
What is the fundamental issue with Mobile IP?
It uses the same method of addressing as in normal IP, which means the entanglement problem is still present.
What is ILNP?
Identifier-Locator Network Protocol is a network protocol designed to separate the two functions of network addresses, the identification of network endpoints, and assisting routing by separating topological information from node identity.
Give an overview of how ILNP works.
128 bit IPv6 address is split into 64 bit locator address and 64 bit node identifier. The locator address is topologically significant, identifies the subnetwork a host is in, and is used for routing and forwarding. The node identifier is not topologically significant, names a node not an interface, and is used only by upper layers in the layered model.
How do TCP and IP packets change from IP to ILNP?
In IP, TCP packets have IP addresses and port numbers of each host, and IP packets have the IP addresses of each host. In ILNP, TCP packets have node identifiers and IP packets have locators.
