Download Chapter 5

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Deep packet inspection wikipedia , lookup

Point-to-Point Protocol over Ethernet wikipedia , lookup

Asynchronous Transfer Mode wikipedia , lookup

Zero-configuration networking wikipedia , lookup

IEEE 1355 wikipedia , lookup

Multiprotocol Label Switching wikipedia , lookup

Piggybacking (Internet access) wikipedia , lookup

Computer network wikipedia , lookup

Wake-on-LAN wikipedia , lookup

Network tap wikipedia , lookup

List of wireless community networks by region wikipedia , lookup

Recursive InterNetwork Architecture (RINA) wikipedia , lookup

Airborne Networking wikipedia , lookup

Cracking of wireless networks wikipedia , lookup

CAN bus wikipedia , lookup

Internet protocol suite wikipedia , lookup

Packet switching wikipedia , lookup

UniPro protocol stack wikipedia , lookup

Transcript
CSCI 4550/8556
Computer Networks
Comer, Chapter 21:
IP Encapsulation, Fragmentation, and
Reassembly
Datagram Transmission and Frames
The IP internet layer:
constructs a datagram,
determines the next hop, and
hands the datagram to the network interface layer.
The network interface layer:
binds the next hop address to the hardware address, and
prepares the datagram for transmission.
But, the hardware frame doesn’t understand IP; how is the
datagram transmitted?
Encapsulation
The network interface layer encapsulates each IP datagram
in the data area in the hardware frame
The hardware ignores the IP datagram format.
The standards for encapsulation describe the details.
The standard defines the data type for an IP datagram, as
well as others (e.g., ARP)
The receiving protocol stack interprets the data area based
on the frame type.
Encapsulation Across Multiple Hops
Each router in the path
from the source to the
destination:
unencapsulates incoming
datagram from the frame;
processes the datagram determines next hop; and
encapsulates the datagram
in an outgoing frame.
Datagram may be
encapsulated in a different
hardware format at each
hop.
The datagram itself is
(almost!) unchanged.
Maximum Transmission Unit: MTU
Every hardware technology specification includes
the definition of the maximum size of a frame’s
data area.
This size is called the maximum transmission unit
(MTU).
Any datagram encapsulated in a hardware frame
must be no larger than the MTU for that hardware.
MTU and Datagram Transmission
IP datagrams can be larger than most hardware MTUs. For
example
IP: 216 – 1, or 65535 bytes
Ethernet: 1500 bytes
Token ring: 2048 or 4096 bytes
The source can simply limit an IP datagram’s size to be no
larger than the local MTU.
Thus we must pass information about the local MTU up to
TCP for TCP segments.
But what about UDP datagrams? Aren’t these supposed to
be treated as single “virtual frames?”
MTU and Heterogeneous Networks
An internet may have networks with different
MTUs.
Suppose a “downstream” network has a smaller
MTU than the local network. How is the hardware
frame size handled in this case?
Fragmentation
It is obvious that the larger frame cannot fit in the smaller
frame.
One technique to solve the problem is to limit a datagram’s
size to the smallest MTU of any network in the internet.
IP uses fragmentation - datagrams can be split into pieces
to fit into a network with a smaller MTU than the current
network.
A router can detect that a datagram is larger than the next
network’s MTU.
It then splits the datagram into pieces, each of which is no
larger than the outbound network’s MTU.
Fragmentation (Details)
Each fragment is an independent datagram.
It includes all IP datagram header fields.
A bit in the header indicates the datagram is a fragment.
Other fields have information for reconstructing original datagram.
FRAGMENT OFFSET gives original location of fragment
A router uses the local MTU to compute the size of each fragment.
It puts part of the data from original datagram in each fragment.
Additional information is placed in the header.
Datagram Reassembly
Reconstruction of the original datagram is called
reassembly .
The ultimate destination performs reassembly.
Fragments may arrive out of order; a header bit identifies
the fragment containing the end of the data from the
original datagram.
In the example, fragment 3 would be identified as the last
fragment.
Fragment Identification
Inidivudal fragments must be associated with the original
datagram.
The IDENT field in each fragment matches the IDENT field
in the original datagram.
Fragments from different datagrams can arrive out of order
and still be sorted out by :
matching them using the IDENT field values;
recognizing the last fragment from the bit set in its header; and
recognizing when all fragments in an original datagram have been
received.
Fragment Loss
IP may drop a fragment (for many reasons).
What happens to the original datagram in this case?
A. The destination drops the entire original datagram.
How does the destination identify a lost fragment?
A timer is associated with each datagram.
If the timer expires before all fragments arrive, one or more
fragments are assumed lost, and …
The entire datagram is dropped.
The source (in the application layer protocol) is assumed to
retransmit if necessary.
Fragmenting a Fragment
A fragment may encounter a subsequent network
with an even smaller MTU.
The router fragments the fragment to fit that
smaller MTU.
The resulting “subfragments” look just like the
original fragments, except for their sizes.
There is no need to reassemble fragments
hierarchically; “subfragments” include their position
in original datagram in a header field.
Summary
IP uses encapsulation to transmit datagrams in
hardware frames.
Network technologies have an MTU that limits the
amount of data that a single frame can carry.
IP uses fragmentation to carry datagrams larger
than the network MTU.