1. CS4470 Computer Networking Protocols
7/6/2018
CS Computer Networking Protocols
8. Network layer 2
Huiping Guo
Department of Computer Science
California State University, Los Angeles

2. Outline What’s insider a router IPv4 datagrams
Datagram transmission Paradigm
Fragmentation and reassembly
8. Network layer CS470

3. Router architecture overview
two key router functions:
run routing algorithms/protocol (RIP, OSPF, BGP)
forwarding datagrams from incoming to outgoing link
forwarding tables computed,
pushed to input ports
routing
processor
routing, management
control plane (software)
forwarding data plane (hardware)
high-seed
switching
fabric
router input ports
router output ports
8. Network layer CS470

4. Router architecture overview
Router control plane
Runs the routing protocols
Responds to the attached links that go up or down
Performs management functions
Implemented in software
Execute on routing processor
Router forwarding plane
Consists of input ports, output ports and switching fabric
Implements forwarding function
Implemented in hardware
8. Network layer CS470

5. Input port functions decentralized switching: physical layer:
lookup,
forwarding
queueing
line
termination
link
layer
protocol
(receive)
switch
fabric
physical layer:
bit-level reception
data link layer:
e.g., Ethernet
decentralized switching:
given datagram dest., lookup output port using forwarding table in input port memory (“match plus action”)
goal: complete input port processing at ‘line speed’
queuing: if datagrams arrive faster than forwarding rate into switch fabric
8. Network layer CS470

6. Switching fabrics
transfer packet from input buffer to appropriate output buffer
switching rate: rate at which packets can be transfer from inputs to outputs
often measured as multiple of input/output line rate
N inputs: switching rate N times line rate desirable
three types of switching fabrics
memory
memory
bus
crossbar
8. Network layer CS470

7. Switching via memory first generation routers:
traditional computers with switching under direct control of CPU
packet copied to system’s memory
speed limited by memory bandwidth
Memory bandwidth B, forwarding throughput <=B/2
2 packets cannot be forwarded at the same time
input
port
(e.g.,
Ethernet)
memory
output
system bus
8. Network layer CS470

8. Switching via a bus datagram from input port memory
to output port memory via a shared bus without intervention of the routing processor
bus contention: switching speed limited by bus bandwidth
32 Gbps bus, Cisco 5600: sufficient speed for access and enterprise routers
bus
8. Network layer CS470

9. Switching via interconnection network
overcome bus bandwidth limitations
A crossbar switch
an interconnection network consisting of 2N buses that connect N input ports to N output ports
Can forward multiple packets in parallel
Cisco 12000: switches 60 Gbps through the interconnection network
crossbar
8. Network layer CS470

10. Output ports
datagram
buffer
queueing
switch
fabric
link
layer
protocol
(send)
line
termination
buffering required when datagrams arrive from fabric faster than the transmission rate
scheduling discipline chooses among queued datagrams for transmission
Datagram (packets) can be lost due to congestion, lack of buffers
Priority scheduling – who gets best performance, network neutrality
8. Network layer CS470

11. Why does queuing occurs?
Packet queues my form at both the input ports and output ports
The location and extent of queuing depend on
Traffic load
Speed of the switching fabric
Line speed
Main reasons of packet loss
As the queues grow large, the router’s memory can eventually be exhausted.
Packet loss will occur when no memory is available to store the arriving packets
8. Network layer CS470

12. Output port queueing
at t, 3 packets
from input to output
one packet time later
switch
fabric
buffering when arrival rate via switch exceeds output line speed
queueing (delay) and loss due to output port buffer overflow!
8. Network layer CS470

13. Input port queuing
fabric slower than input ports combined -> queueing may occur at input queues
queueing delay and loss due to input buffer overflow!
Head-of-the-Line (HOL) blocking: queued datagram at front of queue prevents others in queue from moving forward
one packet time later: green packet experiences HOL blocking
switch
fabric
switch
fabric
output port contention:
only one red datagram can be transferred. lower red packet is blocked
8. Network layer CS470

14. The Internet Network layer
Host, router network layer functions:
Transport layer: TCP, UDP
IP protocol
addressing conventions
datagram format
datagram handling conventions
Routing protocols
path selection
RIP, OSPF, BGP
Network
layer
forwarding
table
ICMP protocol
error reporting
router “signaling”
Link layer
physical layer
8. Network layer CS470

15. IP datagrams Motivation: universal service
Because it can connect heterogeneous networks, a router cannot transmit a copy of a frame that arrives on one network across another.
To accommodate heterogeneity, an internet must define a hardware-independent datagram format.
8. Network layer CS470

16. IP datagrams Created and understood only by software
Contains sender and destination addresses
Size depends on data being carried
Called IP datagram
header
Data
datagram
8. Network layer CS470

17. IPv4 datagram format
Version
HLen
TOS
Length ID
Flags
Offset
TTL
Protocol
Checksum
SourceAddr
DestinationAddr
Options (variable)
Pad
(variable) 4 8 16 19 31
Data
Every datagram contains enough information for the network to forward it from source to destination.
Question: Why is Version on the top of the datagram?
Aligned on 32-bit boundaries.
8. Network layer CS470

18. IPv4 datagram format (cont.)
Version
HLen
TOS
Length ID
Flags
Offset
TTL
Protocol
Checksum
SourceAddr
DestinationAddr
Options (variable)
Pad
(variable) 4 8 16 19 31
Data
HLen: length of header(the number of 32-bit words).
TOS: type of service.
Length: total length of datagram(in bytes) (header+data).
Questions:
Why HLen?
What is the minimum, maximum length of a datagram? Will the underlying physical network support it?
Aligned on 32-bit boundaries.
8. Network layer CS470

19. Type of Service D T R C Precedence TOS Not used in IPv4
7/6/2018
Type of Service
D: Minimize delay R: Maximize reliability
T: Maximize throughput C: Minimize cost D T R C
Precedence
TOS
Not used in IPv4
TOS bits
Description
Protocol
0000
Normal
ICMP
0001
Minimize cost
NNTP
0010
Maximize reliability
SNMP
0100
Maximize throughput
FTP (data)
1000
Minimize delay
FTP(control)
DNS
NNTP: network lay3 protocol
8. Network layer CS470

20. IPv4 datagram format (cont.)
7/6/2018
IPv4 datagram format (cont.)
Version
HLen
TOS
Length ID
Flags
Offset
TTL
Protocol
Checksum
SourceAddr
DestinationAddr
Options (variable)
Pad
(variable) 4 8 16 19 31
Data
TTL: time to live (hop counts)
ensure data don’t circulate for ever
Usually set to 32 or 64
Protocol: identity of upper-level protocol to which IP passes this datagram.
6:TCP 17: UDP
Checksum: applies to header
Questions:
1. Will different routers see the same checksum in the same datagram?
Aligned on 32-bit boundaries.
8. Network layer CS470

21. IPv4 datagram format (cont.)
Version
HLen
TOS
Length ID
Flags
Offset
TTL
Protocol
Checksum
SourceAddr
DestinationAddr
Options (variable)
Pad
(variable) 4 8 16 19 31
Data
SourceAddr
DestinationAddr: used for data delivery.
Options: header length determines if any are set.
Aligned on 32-bit boundaries.
8. Network layer CS470

22. Internet Transmission Paradigm (General Case)
Source host
Forms datagram
Includes destination address
Sends to nearest router
Intermediate routers
Forward datagram to next router
Final router
Delivers to destination host
8. Network layer CS470

23. Datagram Transmission
Datagram sent across conventional network
From source host and router
Between intermediate routers
From final router to destination host
Encapsulation needed
Network hardware does not recognize
Datagram format
IP addresses
8. Network layer CS470

24. Illustration Of IP Encapsulation
Entire datagram treated like data
Frame type identifies contents as IP datagram
Frame destination address gives next hop
8. Network layer CS470

25. Frame And Datagram Destination Addresses
Frame address
Hardware (MAC) address
Next hop
Datagram address
IP address
Ultimate destination
8. Network layer CS470

26. Frames and Datagrams Datagram survives entire trip across Internet
Frame only survives one hop
A datagram is encapsulated in a frame for transmission across a physical network.
The destination address in the frame is the address of the next hop to which the datagram should be sent
the address is obtained by translating the IP address of the next hop to an equivalent hardware address.
8. Network layer CS470

27. Illustration Of Frame Headers Used For Datagram Transmission
Each hop extracts datagram and discards frame
8. Network layer CS470

28. Exercise
In LAN1, LAN2 and LAN3, A sends a datagram to F, what’s the datagram’s source IP address, destination’s IP address, corresponding frame’s source MAC address and destination’s MAC address? 1 2 3
8. Network layer CS470

29. Maximum Frame Size
Each network technology imposes maximum frame data size
Called Maximum Transmission Unit (MTU)
The maximum amount of data that a link layer frame can carry
MTUs differ
Internet
Can contain heterogeneous technologies
Must accommodate multiple MTUs
8. Network layer CS470

30. Illustration Of How Two MTUs Cause A Problem For IP
Host 1
Creates datagram for Host 2
Create a frame of 1500 bytes
Transmits the frame across network 1
Router R
Receives frame over network 1
Must send frame over network 2
Employs fragmentation
8. Network layer CS470

31. Fragmentation and Reassembly
Ethernet: max. frame length = 1500 bytes.
FDDI: max. frame length = 4500 bytes.
MTU
Question: How does the underlying network technology affect IP? How does one choose the maximum datagram size for IP?
Design options: (1) Choose IP datagram max. length to be compatible with MTU. (2) Fragment IP datagram into multiple datagrams
8. Network layer CS470

32. Fragmentation and Reassembly
large IP datagram divided (“fragmented”) within net
one datagram becomes several datagrams
“reassembled” only at final destination
IP header bits used to identify, order related fragments
fragmentation:
in: one large datagram
out: 3 smaller datagrams
8. Network layer CS470

33. Fragmentation and Reassembly
When to re-assemble
At destination only
Results in datagrams getting smaller as data traverses internet
Why not re-assemble at intermediate routers?
Need large buffers at routers
Buffers may fill with fragments
All fragments must go through same router
Routers have enough work to do already
8. Network layer CS470

34. IP Fragmentation IP re-assembles at destination only
Uses fields in header
Data Unit Identifier (ID)
When a datagram is created, the sender stamps the datagram with an ID as well as source and destination address
The sender increments the ID of each datagram it sends
Offset
Position of fragment of user data in original datagram
In multiples of 64 bits (8 bytes)
flag
Indicates that this is not the last fragment
0 if a fragment is the last one
8. Network layer CS470

35. Fragmentation and Reassembly
1. Each fragment is itself a self-contained IP datagram that is transmitted over a sequence of physical networks, independent of other fragments.
2. Each IP datagram is reencapsulated for each physical network over which it travels.
8. Network layer CS470

36. Fragmentation and Reassembly
Ident = x
Start of header
Rest of header
1400 data bytes
Offset = 0
(b)
512 data bytes 1
Offset = 64
376 data bytes
Offset = 128
Some nitty-gritty details
If a fragment is lost, the receiver still tries to assemble the datagram… When this fails, it has to clean up the memory that was used in the attempt.
If hosts can do “path MTU discovery”, then IP fragmentation can be avoided altogether.
8. Network layer CS470

37. Fragmenting Fragments
A fragment may encounter a subsequent network with even smaller MTU
Router fragments the fragment to fit
Resulting (sub)fragments look just like original fragments (except for size)
No need to reassemble hierarchically; (sub)fragments include position in original datagram
8. Network layer CS470

38. Dealing with Failure Re-assembly may fail if some fragments get lost
Need to detect failure
Re-assembly time out
Assigned to first fragment to arrive
If timeout expires before all fragments arrive, discard partial data
Loss of any fragments means loss of entire datagram!
8. Network layer CS470

39. Fragmentation & ReassemblyID
=77
offset =0
flag
length
=4000
fragflag =1
=1500
=185
=370
=1040
One large datagram becomes
several smaller datagrams
Example
A datagram of 4000 byte (header+data)
Assume ID=77
MTU = 1500 bytes
1480 bytes in data field
offset =
1480/8
8. Network layer CS470