IP Mobility

Question 1

Why do we need mobility support for ip?

Answer 1

While we do have mobile systems today (3G, 4G), they are not sufficient. Not available everywhere, bad reception, faults, downtime etc.

Question 2

IP was not designed for mobility. What about it makes it no suited? 2 things

Answer 2

an IP addr is an interface identifier, which means there is layer entanglement. its present in applications, transport (tcp using ip + port), network (ip itself) and interface

the ip address indicates a point of attachment to the network, TIED to an interface. movement of the host means new ipv4 addr and routing info

Question 3

describe the 7 step process of mobile ip.

Answer 3

1) MH (with HoA) arrives at FN, and locates FA
(using agent adverMsements / solicitaMon).

2) MH completes registraMon procedure with FA.
Performs Duplicate Address DetecMon (DAD).

3) MH updates HA with its new CoA (i.e. the FA).

4) Host A now tries to contact MH. Packets for MH are
intercepted by HA (HA acts as a proxy).

5) HA tunnels the packets from Host A to the CoA for MH (i.e. the FA)
6) The FA de-encapsulates the inner IP packet and transmits the packet locally to MH.
7) The packets from MH to Host A are sent directly from the FN.

Improved in Mobile IPv6:
• mobile host can act as its own FA
• use of Binding Update to send CoA to HostA

Question 4

why does mobile ip suck? 2

Answer 4

MH has two addresses:
– HoA: (semi-)permanent, acts as an idenLfier.
– CoA: transient, acts as a topological locator.
so End-to-end semanMcs are lost:
– proxy with tunnelling.

 Proxy is a potential weak point:
– performance boqleneck.
– single-point-of-failure.
– security aqack point.
– traffic monitoring and remote tracking of MH (privacy).

Question 5

pros (5) and cons (3) of mipv4

Answer 5

�Transparent to non-mobile
(aware) hosts.
�Does not break/change
existing IP addressing and
routing.
�Can be introduced into the
network as required
(incrementally).
�Normal (unicast) routers do
not need to be modified.
� Does not affect DNS usage.

Complexity:
– use of addresses.
– use of agents (proxies).
Asymmetric routing:
– could be inefficient.
– path characteristics.
– higher layer protocol
operation (e.g. TCP).
Security and privacy:
– firewall configuration.
– authentication.
– end-to-end security.
– tracking of users.

Question 6

what does mipv6 bring?

Answer 6

Auto-configuration:
– stateless address auto-configuration to find CoA for use at the FN.
– neighbour discovery to find default router.
• No FA required to support mobility:
– MH takes care of home address (HoA) and care-of address (CoA) itself.
• Need dynamic DNS update support for incoming connections:
– or application-specific rendezvous.
• Address and route optimisation:
– optimistic DAD (assume DAD will succeed).
– send CoA to remote end-system using a Binding Update (BU).
• IPv6 Binding Update:
– similar function to ILNPv6 Locator Update.
• Security:
– authentication and privacy

Question 7

how does ilnp solve the entangment issues?

Answer 7

application layer uses fqdn, transport uses identifier (+ port no), network uses locator, interface has a dynamic mapping

Question 8

how can ilnp fit in the existing internet

Answer 8

Can be seen as a set of ‘extensions’ to IPv6:
• Uses same packet format as IPv6 in network core.
• IPv6 core routers do not need to change.
• Incrementally deployable on IPv6 core.
• Backwards compatible with IPv6.
• Split 128-bit IPv6 address:
• 64-bit Locator (L) - network name.
• 64-bit Identifier (I) - node name.
• Could also be retro-fitted to IPv4 (but messy).

uses the same syntax and semantics as ipv6 routing prefix so ipv6 core routers work as normal

the ipv6 addresses (source and dest) are split into id and locat. but they get reead the same! wowwee

• ILNPv6 on-the-wire is the same as IPv6 on-the- wire, end-to-end

Question 9

how are the identifier and locator used in ilnp?

Answer 9

Locator, 64 bits, L64:
• Topologically significant.
• Names a (sub)network
(same as today's network prefix).
• L64 used only for routing and forwarding.

 Node Identifier, 64 bits, NID:
• Is not topologically significant.
• Names a logical/virtual/physical node, does not
name (bind to) an interface.
– NID used only by upper layers.

Question 10

why is ilnp lit? cons?

Answer 10

No NATs needed.
• No tunnels needed.
• No proxies / middleboxes needed.
• No changes to routing needed.
• Harmonised functionality in the end-system:
– mobility without agents or proxies.
– mobility and multihoming together (duality).
– multihoming without extra routing state.
– end-to-end packet-level security.
– support for wide-area VM-image mobility.

needs DNS enhancements :c
Need new DNS Resource Records, e.g.:
• NID – 64-bit IdenMfier value, EUI-64 syntax.
• L64 – 64-bit Locator value.
• DNS lookup will return:
• 1 or more NID records, 1 or more L64 records.
• For mulMple NID and L64 RRs, use preference bits.

Question 11

how is ilnp network handoff liter?

Answer 11

no need for DAD check, instead it just sends a locator update

Question 12

what r some considerations about ilnp?

Answer 12

No globally routeable interface name, which may
impact some applicaMons such as SNMP.
• Some legacy applicaMons may break:
– use of address in applicaMon code or configuraMon.
• No new security issues created:
– potenMal for enhanced end-to-end security.
– potenMal for beqer locaMon privacy.
– potenMal for beqer idenMty privacy.