CS335 Networking &

Network Administration

Tuesday, May 11, 2010

ARP – Address resolution protocol

Translates IP address into a hardware address Physical network hardware does not know how to locate

a computer from its protocol address Known as address resolution

ARP

Can only resolve hardware addresses for machines on the local physical network

Address Resolution

Three techniques of address resolution Table lookup – stored in a table in memory

Table lookup

For less than a dozen hosts sequential search suffices

In larger networks this requires excessive CPU cycles

Hashing – general purpose data structure Direct indexing

Table lookup

Direct indexing – uses the host address as an index into the array

Address resolution

Closed-form computation Used when the network interface can be

assigned specific hardware addresses Computed by a single Boolean and operation Hardware_address = ip_address & 0xff When a computer connects to a network that

uses this, resolution is trivial

Address resolution

Message exchange Computers exchange messages across network

to resolve an address 3 types of address resolution

Table lookup Closed form computation Dynamic message exchange

Address resolution

ARP

ARP standard defines 2 basic message types Request – contains an IP address and requests

the hardware address Response – has both the IP address and the

hardware address

ARP message delivery

ARP message format

Although the ARP message format is sufficiently general to allow arbitrary protocol and hardware addresses, ARP is always used to bind a 32 bit IP address to a 48 bit Ethernet address

ARP is encapsulated directly in a hardware frame

Identifying ARP frames

The type field in the frame header specifies that the frame contains an ARP message

ARP caching

ARP software extracts and saves the information

Uses small table of bindings in memory Checks cache first before broadcasting an

ARP request Improves the efficiency of network traffic

Higher levels use protocol addressing

IP

TCP/IP includes both connectionless and connection-oriented services

Routers can connect heterogeneous networks so they cannot transmit a copy of a frame that arrives on one network across another

IP is a hardware independent packet format

IP datagram

Size of a datagram is determined by the application that sends the data

Similar to format of a frame Uses IP addresses in header Can contain as little as a single octet of data or at most

64K octets

Forwarding IP datagrams

Next hop – either the destination or the next router

IP addresses and routing tables

Routing

Destination and Next-Hop addresses

The destination address in a datagram header always refers to the ultimate destination

When a router forwards the datagram to another router the address of the next hop does not appear in the datagram header.

Best effort delivery

IP uses best-effort to describe the service Doesn’t guarantee that it will handle:

Datagram duplication Delayed or out-of-order delivery Corruption of data Datagram loss

Additional layers of protocol software handle these errors

IP Datagram Header

Each field has a fixed size

Encapsulation

Network hardware doesn’t understand datagram format or IP addressing

Network understands its own frame format and heterogeneous networks may have different formats

IP datagram is encapsulated in a frame

Encapsulation

Frame type field uses the value reserved for IP

Receiver knows the data area contains IP datagram

Uses a frame address for next hop obtained by ARP

Transmission across an internet

When a datagram arrives in a network frame the receiver extracts the datagram from the frame data and discards the frame header. Frame headers don’t accumulate on the trip.

MTU – Maximum transmission unit

Each hardware technology has a limit to the amount of data in a frame

Datagram must be smaller than the MTU or it can’t be encapsulated for transmission

Fragmentation

In a internet with heterogeneous networks, MTU restrictions can be a problem

Routers fragment or divide a datagram into smaller pieces to meet the MTU

Fragmentation

Each fragment uses the IP datagram format but carries only part of the data

Flags field of the header indicates whether it is fragment or a complete datagram

Reassembly

Process of creating a copy of the original datagram from fragments

Fragment with the final data has an additional bit set in header so receiver knows all fragments have arrived

Ultimate destination host reassembles fragments so the routers

Identifying a datagram

IP doesn’t guarantee delivery Fragments can be lost or arrive out of order Sender places a unique identification number

in the identification field of outgoing datagram When a router fragments, the identification

number and source IP address determines to which datagram a fragment belongs

Fragment offset field tells a receiver how to order fragments

Fragment loss Fragments can be delayed or lost IP specifies a maximum time to hold fragments if they are delayed When a fragment arrives receiver starts a timer, if all arrive before

timer runs out, datagram is reassembled, otherwise they are discarded

No way for receiver to tell sender what fragments didn’t arrive Sender doesn’t know about fragmentation Resent packets may take a different path with different

fragmentation Fragments can be fragmented in case of an even smaller MTU

Future IP

Let’s go to the lab and research IP v6!