Transcript Part I: Introduction - University of Pittsburgh

Final Review
CS1652
Jack Lange
University of Pittsburgh
Final
 Thurs 12th: 12-2PM

This room
 Not Cumulative
 One page of notes
2
Chapter 4: Network Layer
 Network layer service model
 Forwarding vs. Routing
 IP addressing, NAT & DHCP
 Router architecture
 Routing algorithms
 Handling scale - CIDR and BGP
 ICMP, traceroute
3
Network layer service model
 Unreliable, connectionless data delivery
Q: is UDP more reliable than IP?
 Q: do TCP packets receive more special
treatment than UDP packets?

 Host-to-host packet delivery
 Q: Difference from transport layer?
 All systems implement network layer
 End systems + routers
 Q: Why not routers implement transport layer?
4
Forwarding vs. Routing
 Definition?
 Switch vs. router?
 How routing affects forwarding?
 Routing algorithms

Intradomain routing algorithms
• Distance vector, Link state

Interdomain routing algorithms
• BGP
 Longest prefix match
5
IP addressing, DHCP and NAT
 32-bit IPv4 address
 Subnet
part + host part
 Subnet part is used for forwarding decision
 CIDR
 Subnet portion can be an arbitrary size
 Why CIDR?
• IP allocation efficiency & supernetting
 DHCP
 How it works? Where is it useful?
 NAT
 Pros. and Cons.
6
IP datagram format
IP protocol version
number
header length
(bytes)
“type” of data
max number
remaining hops
(decremented at
each router)
upper layer protocol
to deliver payload to
how much overhead
with TCP?
 20 bytes of TCP
 20 bytes of IP
 = 40 bytes + app
layer overhead
32 bits
type of
ver head.
len service
length
fragment
16-bit identifier flgs
offset
time to upper
header
layer
live
checksum
total datagram
length (bytes)
for
fragmentation/
reassembly
32 bit source IP address
32 bit destination IP address
Options (if any)
data
(variable length,
typically a TCP
or UDP segment)
E.g. timestamp,
record route
taken, specify
list of routers
to visit.
7
Router Architecture
Two key router functions:
 Run routing algorithms/protocol (RIP, OSPF, BGP)
 Forwarding datagrams from incoming to outgoing link
4-8
Routers
 Input ports (line cards)
 Forwarding
table lookup – line speed
 Queuing packets if switching fabric is busy
• Head-of-line blocking?
 Switching fabric

Via memory, bus, special interconnection
 Output ports
 Buffering & scheduling
9
Link state algorithm
 Broadcast local link info to all routers
 Dijkstra’s algorithm
 Greedy algorithm
 Compute the least cost path to every node
 Each loop finds at least one node whose least
cost path is found
 Algorithm complexity? O(nlogn)
 Oscillation problem
10
Distance vector algorithm
 Distribute one’s view of network to neighbors
 Bellman-ford algorithm
 Dynamic programming
 Asynchronous update
 Problem?
Count-to-infinity & routing loops
 Possible solution?

11
Border Gateway Protocol (BGP)
 Autonomous System (AS)
 AS number – 16 bit id
 BGP contains full path from src to dest AS
 AS PATH – list of AS numbers
 How to prevent routing loops?
 Hot potato routing?
 One reason for routing path asymmetry
 How to deliver a packet from one AS to another?
 Intradomain (Intra-AS) routing
 Interdomain routing
 Forwarding table (FIB)
12
Chapter 5: Link Layer
 Delivering frames to a direct neighbor
 Error detection and correction
 Sharing a broadcast channel
 Reliable data transfer & flow control
 Hop-by-hop vs. end-to-end
13
Error detection & correction
 Parity checking
 Single
bit vs. two-dimentional bit parity
 Odd/even parity
 Internet checksum – IP/TCP layer
 Why is error checking needed in the upper layer?
 Cyclic Redunancy Check (CRC)
 CRC32 is widely used (e.g., Ethernet)
14
Multiple Access Protocol
 How to share a broadcasting media
 Medium
Acccess Control (MAC) protocol
 Channel Partitioning
 Random Access
 Taking turns
15
Random Access Protocol
 Slotted ALOHA
 Fixed time slot – synchronized
 If collision, retransmit with a probability of p for
each slot
 ALOHA
 No slot synchronization
 CSMA, CSMA/CD, CSMA/CA
 CSMA – sense carrier before sending
 CD – detect collision while sending and cancel it
 CA – avoid collision by getting permission first
 Ethernet and Wi-Fi?
16
MAC addresses
 Ethernet: 48-bit MAC addresses
 Burned
into hardware
 Globally uniquely assigned
 Why not use MAC address instead of IP?
 Address Resolution Protocol (ARP)
 Determining MAC address with IP address
 ARP table = <IP address, MAC address, TTL>
 Broadcasting mechanism (make sure to know!)
17
Ethernet & Switch
 Access protocol: CSMA/CD
 Old
Ethernet hubs used to share access
 Understand the exponential back-off algorithm
 Most current devices are switches
 Connectionless, unreliable
 Frame format
 Preamble, 2 MAC address, type, data, CRC32
 Switch

Make sure you understand self-learning algorithm
18
Chapter 6: Wireless
 Difference from wired environment
 Signal
attenuation
 Hidden terminal problem
 Interference from other sources (phone, microwave)
 SNR (Signal-to-Noise ratio)
 The larger, the better
 BER (Bit error rate)
 CDMA
19
Wi-Fi
 Access protocol : CSMA/CA
 Reserve
the channel first before sending
 No collision detection – why?
 Understand the sending protocol
 Access Point(AP)
 Link layer device (may run DHCP)
 Passive/Active scanning for association
 802.11 frame format

Why we need three MAC addresses?
20
Chapter 7: Multimedia
 Quality of service guarantee
 Providing
performance guarantee required by app
 Current Internet does not directly support it
 Delay sensitive, loss tolerant application
 Video streaming vs. Email?
 What is jitter?
 Multimedia application
Stored streaming
 Live streaming
 Real-time interactive

21
Internet Phone, CDN
 Internet phone
 Network
loss vs. delay loss
 Content distribution networks (CDNs)
 Definition?
 DNS redirection for finding the near server?
22