Download Internet Design Principles (Cont.) and Link Layer

Introduction to Communication
Networks –67594
Dr. Michael Schapira
Some of the slides were taken from Prof. Scott Shenker, UC Berkeley
Physical Communication
• Communication goes down to physical network
• Then from network peer to peer
• Then up to relevant layer
Application
Transport
Network
Datalink
Physical
Host A
Protocol
Network
Datalink
Physical
Router
Application
Transport
Network
Datalink
Physical
Host B
2
Characterizing the Layers (OSI)
• Service: what a layer does
• Service interface: how to access the service
– Interface for layer above
• Protocol: how peers communicate
– Protocol interface: set of rules and formats that govern the
communication between network elements
– Determines how the peers achieve the service
– Does not govern implementation on a single machine, but
how the layer is implemented between machines
• Examples: layer can have many implementations
– This allows innovation!
3
The Internet Hourglass
SMTP HTTP
DNS
TCP
Applications
NTP
Transport
UDP
IP
Waist
Data Link
Ethernet
Copper
SONET
Fiber
802.11
Radio
Physical
The Hourglass Model
There is just one network-layer protocol, IP.
The “narrow waist” facilitates interoperability.
4
Implications of Hourglass
• Allows arbitrary L2 networks to interoperate
– Any networks that support IP can exchange packets
• Allows applications to function on all networks
– Applications that can run on IP can use any network
• Supports innovations above and below IP
– Innovations can be independent, done in parallel
• But changing IP itself, i.e., IPv6, very hard
5
Back to Modularity
• Modularity in programming:
– Set of abstractions
• Modularity in networking:
– Distributed nature of system requires deciding
where abstractions are implemented
6
Three Basic Architectural Decisions
• How to break system into modules
• Where modules are implemented
• Where state is stored
7
Breaking into Modules
• That is described by layering
8
Placing Network Functionality
• Hugely influential paper: “End-to-End Arguments
in System Design” by Saltzer, Reed, and Clark
(‘84)
– End-to-end principle
• Basic observation: some types of network
functionality can only be correctly implemented
end-to-end
• Because of this, end hosts:
– Can satisfy the requirement without network’s help
– Must do so, since can’t rely on network’s help
• Thus, don’t need to implement them in network
9
Example: Reliable File Transfer
Host A
Host B
Appl.
OS
Appl.
OK
OS
• Solution 1: make each step reliable, and string
them together to make reliable end-to-end
process
• Solution 2: allow steps to be unreliable, but do
end-to-end check and try again if necessary
10
Discussion
• Solution 1 cannot be made perfectly reliable
– What happens if a network element misbehaves?
– Receiver has to do the check anyway!
• Solution 2 can also fail, but only if the end system itself
fails (i.e., doesn’t follow its own protocol)
• Solution 2 only relies on what it can control
– The endpoint behavior
• Solution 1 requires endpoints trust other elements
– That’s not what reliable means!
11
Robust (From Clark’s Paper)
• As long as the network is not partitioned,
two endpoints should be able to
communicate
• Failures (except network partition) should
not interfere with endpoint semantics
12
Question?
• Should you ever implement reliability in network?
• Perhaps, if needed for reasonable efficiency
– Don’t aim for perfect reliability, but ok to reduce error rate
• If individual links fail 10% of the time, and are
traversing 10 links, then E2E error rate is 65%
• Implementing one retransmission on links
– Link error rate reduced to 1%, E2E error rate is 9.5%
13
Back to the End-to-End Principle
Implementing such functionality in the network:
• Doesn’t reduce host implementation complexity
• Does increase network complexity
• Probably imposes delay and overhead on all
applications, even if they don’t need functionality
• However, implementing in network can enhance
performance in some cases
– E.g., very lossy link
14
Conservative Interpretation of
E2E
• Don’t implement a function at the lower
levels of the system unless it can be
completely implemented at this level
• Unless you can relieve the burden from
hosts, don’t bother
15
Radical Interpretation of E2E
• Don’t implement anything in the network
that can be implemented correctly by the
hosts
– E.g., multicast
• Make network layer absolutely minimal
– E2E principle trumps performance issues
– Increases flexibility, since lower layers stay
simple
16
Important life lessons (partial)
• Flexibility often more important than
performance
• Architect for flexibility, engineer for
performance
17
Moderate Interpretation
• Think twice before implementing
functionality in the network
• If hosts can implement functionality
correctly, implement it in a lower layer only
as a performance enhancement
• But do so only if it does not impose burden
on applications that do not require that
functionality
18
What Does E2E Principle Ignore?
• There are other stakeholders besides users
– ISP might care about the good operation of their
network
– Various commercial entities
– Money-chain might require insertion into the
network
• The need for middlebox functionality
– Some functions that, for management reasons, are
more easily done in the network.
19
Three Basic Architectural Decisions
• How to break system into modules
• Where modules are implemented
• Where state is stored
20
Fate-Sharing
• Note that E2E principles relied on “fatesharing”
– Invariants break only when endpoints themselves
break
– Minimize dependence on other network elements
• This should dictate placement of storage
21
General Principle: Fate-Sharing
• When storing state in a distributed system, colocate it with entities that rely on that state
• Only way failure can cause loss of the critical
state is if the entity that cares about it also fails ...
– … in which case it doesn’t matter
• Often argues for keeping network state at end
hosts rather than inside routers
– In keeping with End-to-End principle
– E.g., packet-switching rather than circuit-switching
– E.g., NFS file handles, HTTP “cookies”
22
Decisions and Their Principles
• How to break system into modules
– Dictated by Layering
• Where modules are implemented
– Dictated by End-to-End Principle
• Where state is stored
– Dictated by Fate-Sharing
23
Reminder: Tasks in Networking
(bottom up)
•
•
•
•
Electrons on wire
Bits on wire
Packets on wire
Deliver packets across local network
– Local addresses
Not in this course
(mostly)
The following few
weeks
• Deliver packets across country
– Global addresses
• Ensure that packets get there
• Do something with the data
Later in the course
24
Chapter 5
Link Layer and LANs
A note on the use of these ppt slides:
We’re making these slides freely available to all (faculty, students, readers).
They’re in PowerPoint form so you can add, modify, and delete slides
(including this one) and slide content to suit your needs. They obviously
represent a lot of work on our part. In return for use, we only ask the
following:
 If you use these slides (e.g., in a class) in substantially unaltered form,
that you mention their source (after all, we’d like people to use our book!)
 If you post any slides in substantially unaltered form on a www site, that
you note that they are adapted from (or perhaps identical to) our slides, and
note our copyright of this material.
Computer Networking:
A Top Down Approach
5th edition.
Jim Kurose, Keith Ross
Addison-Wesley, April
2009.
Thanks and enjoy! JFK/KWR
All material copyright 1996-2009
J.F Kurose and K.W. Ross, All Rights Reserved
5: DataLink Layer
5-25
The Data Link Layer
Our goals:
r understand principles behind data link layer
services:
m
m
m
m
error detection, correction
sharing a broadcast channel: multiple access
link layer addressing
reliable data transfer, flow control
r instantiation and implementation of various link
layer technologies
5: DataLink Layer
5-26
Link Layer
r 5.1 Introduction and
r
r
r
r
services
5.2 Error detection
and correction
5.3Multiple access
protocols
5.4 Link-layer
Addressing
5.5 Ethernet
r 5.6 Link-layer switches
r 5.7 PPP
r 5.8 Link virtualization:
ATM, MPLS
5: DataLink Layer
5-27
Some Terminology
r
r
hosts and routers are nodes
communication channels that
connect adjacent nodes along
communication path are links
m
m
m
r
wired links
wireless links
LANs
layer-2 packet is a frame,
encapsulates datagram
data-link layer has responsibility of
transferring datagram from one node
to adjacent node over a link
5: DataLink Layer
5-28
Link layer: context
r datagram transferred by
different link protocols
over different links:
m
e.g., Ethernet on first link,
frame relay on
intermediate links, 802.11
on last link
r each link protocol
provides different
services
m
e.g., may or may not
provide reliability over link
transportation analogy
r
trip from Honolulu to
Jerusalem
m taxi: Honolulu to HNL
m plane: HNL to EWR
m plane: EWR to TLV
m
train☺: TLV to Jerusalem
r passenger = datagram
r transport segment =
communication link
r transportation mode =
link layer protocol
r travel agent = routing
algorithm
5: DataLink Layer
5-29
Link Layer Services
r framing, link access:
m
m
m
encapsulate datagram into frame, adding header, trailer
channel access if shared medium
“MAC” addresses used in frame headers to identify
source, dest
• different from IP address!
r reliable delivery between adjacent nodes
m seldom used on low bit-error link (fiber, some twisted
pair)
m wireless links: high error rates
• Q: how this corresponds with the E2E principle?
5: DataLink Layer
5-30
Link Layer Services (more)
r flow control:
m
pacing between adjacent sending and receiving nodes
r error detection:
m
m
errors caused by signal attenuation, noise.
receiver detects presence of errors:
• signals sender for retransmission or drops frame
r error correction:
m
receiver identifies and corrects bit error(s) without
resorting to retransmission
r half-duplex and full-duplex
m with half duplex, nodes at both ends of link can transmit,
but not at same time
5: DataLink Layer
5-31
Where is the link layer implemented?
r in each and every host
r link layer implemented in
“adaptor” (aka network
interface card NIC)
m
m
Ethernet card, PCMCI
card, 802.11 card
implements link, physical
layer
r attaches into host’s
system buses
r combination of
hardware, software,
firmware
host schematic
application
transport
network
link
cpu
memory
controller
link
physical
host
bus
(e.g., PCI)
physical
transmission
network adapter
card
5: DataLink Layer
5-32
Adaptors Communicating
datagram
datagram
controller
controller
receiving host
sending host
datagram
frame
r sending side:
m encapsulates datagram in
frame
m adds error checking bits,
rdt, flow control, etc.
r receiving side
m looks for errors, flow
control, etc.
m extracts datagram, passes
to upper layer at receiving
side
5: DataLink Layer
5-33
Link Layer
r 5.1 Introduction and
r
r
r
r
services
5.2 Error detection
and correction
5.3Multiple access
protocols
5.4 Link-layer
Addressing
5.5 Ethernet
r 5.6 Link-layer switches
r 5.7 PPP
r 5.8 Link Virtualization:
ATM. MPLS
5: DataLink Layer
5-34
Error Detection
EDC= Error Detection and Correction bits (redundancy)
D = Data protected by error checking, may include header fields
• Error detection not 100% reliable!
• protocol may miss some errors, but rarely
• larger EDC field yields better detection and correction
otherwise
5: DataLink Layer
5-35
Parity Checking
Single Bit Parity:
Detect single bit errors
Two Dimensional Bit Parity:
Detect and correct single bit errors
0
0
5: DataLink Layer
5-36
Internet checksum (review)
Goal: detect “errors” (e.g., flipped bits) in
transmitted packet (note: used at transport layer
only)
Receiver:
Sender:
r
r
r
treat segment contents
as sequence of 16-bit
integers
checksum: addition (1’s
complement sum) of
segment contents
sender puts checksum
value into a checksum
field
r
r
compute checksum of
received segment
check if computed checksum
equals checksum field value:
m NO - error detected
m YES - no error detected.
But maybe errors
nonetheless?
5: DataLink Layer
5-37
Checksumming: Cyclic Redundancy Check
r
r
r
view data bits, D, as a binary number
choose r+1 bit pattern (generator), G
goal: choose r CRC bits, R, such that
m
m
m
r
<D,R> exactly divisible by G (modulo 2)
receiver knows G, divides <D,R> by G. If non-zero remainder:
error detected!
can detect all burst errors less than r+1 bits
widely used in practice (Ethernet, 802.11 WiFi, ATM)
5: DataLink Layer
5-38
CRC Example
Want:
D.2r XOR R = nG
equivalently:
D.2r = nG XOR R
equivalently:
if we divide D.2r by
G, want remainder R
R = remainder[
D.2r
G
]
5: DataLink Layer
5-39
Information Theory
r Finding the optimal code under given
constraints is a very well-researched area,
which is part of Information Theory
m
r
Pioneered by Claude Shannon 1916-2001;
first information theory paper in 1948
Very active area of research even today
m
E.g., network coding
r Beside what we saw so far, out of the
scope of this course
5: DataLink Layer
5-40
Link Layer
r 5.1 Introduction and
r
r
r
r
services
5.2 Error detection
and correction
5.3Multiple access
protocols
5.4 Link-layer
Addressing
5.5 Ethernet
r 5.6 Link-layer switches
r 5.7 PPP
r 5.8 Link Virtualization:
ATM, MPLS
5: DataLink Layer
5-41
Multiple Access Links and Protocols
Two types of “links”:
r point-to-point
m PPP for dial-up access
m point-to-point link between Ethernet switch and host
r broadcast (shared wire or medium)
m old-fashioned Ethernet
m upstream HFC
m 802.11 wireless LAN
shared wire (e.g.,
cabled Ethernet)
shared RF
(e.g., 802.11 WiFi)
shared RF
(satellite)
humans at a
cocktail party
(shared air, acoustical)
5: DataLink Layer
5-42
Multiple Access protocols
r single shared broadcast channel
r two or more simultaneous transmissions by nodes:
interference
m
collision if node receives two or more signals at the same time
multiple access protocol
r distributed algorithm that determines how nodes
share channel, i.e., determine when node can transmit
r communication about channel sharing must use channel
itself!
m
no out-of-band channel for coordination
5: DataLink Layer
5-43
Ideal Multiple Access Protocol
Broadcast channel of rate R bps
1. when one node wants to transmit, it can send at
rate R.
2. when M nodes want to transmit, each can send at
average rate R/M
3. fully decentralized:
m
m
no special node to coordinate transmissions
no synchronization of clocks, slots
4. simple
5: DataLink Layer
5-44
MAC Protocols: a taxonomy
Three broad classes:
r Channel Partitioning
m
m
divide channel into smaller “pieces” (time slots,
frequency, code)
allocate piece to node for exclusive use
r Random Access
m channel not divided, allow collisions
m “recover” from collisions
r “Taking turns”
m nodes take turns, but nodes with more to send can take
longer turns
5: DataLink Layer
5-45
Channel Partitioning MAC protocols: TDMA
TDMA: time division multiple access
r access to channel in "rounds"
r each station gets fixed length slot (length =
packet transmission time) in each round
r unused slots go idle
r example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6
idle
6-slot
frame
1
3
4
1
3
4
5: DataLink Layer
5-46
Channel Partitioning MAC protocols: FDMA
FDMA: frequency division multiple access
r channel spectrum divided into frequency bands
r each station assigned fixed frequency band
r unused transmission time in frequency bands go idle
r example: 6-station LAN, 1,3,4 have pkt, frequency
FDM cable
frequency bands
bands 2,5,6 idle
5: DataLink Layer
5-47
Random Access Protocols
r When node has packet to send
m transmit at full channel data rate R.
m no a priori coordination among nodes
r two or more transmitting nodes ➜ “collision”,
r random access MAC protocol specifies:
m how to detect collisions
m how to recover from collisions (e.g., via delayed
retransmissions)
r Examples of random access MAC protocols:
m slotted ALOHA
m ALOHA
m CSMA, CSMA/CD, CSMA/CA
5: DataLink Layer
5-48
Slotted ALOHA
Assumptions:
r all frames same size
r time divided into equal
size slots (time to
transmit 1 frame)
r nodes start to transmit
only at slot beginning
r nodes are synchronized
r if 2 or more nodes
transmit in slot, all
nodes detect collision
Operation:
r when node obtains fresh
frame, transmits in next
slot
m if no collision: node can
send new frame in next
slot
m if collision: node
retransmits frame in
each subsequent slot
with prob. p until
success
5: DataLink Layer
5-49
Slotted ALOHA
Pros
r single active node can
continuously transmit
at full rate of channel
r highly decentralized:
only slots in nodes
need to be in sync
r simple
Cons
r collisions, wasting slots
r idle slots
r nodes may be able to
detect collision in less
than time to transmit
packet
r clock synchronization
5: DataLink Layer
5-50
Goodput vs. Throughput
r Goodput: long-run
fraction of
successful slots
(many nodes, all
with many frames
to send)
r Throughput: long-
run fraction of
slots in which there
was a transmission
(not necessarily
successful)
5: DataLink Layer
5-51
Slotted Aloha Goodput
Goodput: long-run
fraction of successful slots
(many nodes, all with many
frames to send)
r suppose: N nodes with
many frames to send,
each transmits in each
slot with probability p
r prob that given node
succeeds in a slot =
p(1-p)N-1
r prob that some node
succeeds = Np(1-p)N-1
r max efficiency: find
p* that maximizes
Np(1-p)N-1
r for many nodes, taking
limit of Np*(1-p*)N-1
as N goes to infinity,
gives:
Max efficiency = 1/e = .37
At best: channel
used for useful
transmissions 37%
of time!
5: DataLink Layer
!
5-52
Pure (unslotted) ALOHA
r Unslotted Aloha: simpler, no synchronization
r Each host behaves as if it runs slotted ALOHA
m Sends a frame with probability p on each of its own slots.
r collision probability increases:
m
frame sent at a slot the starts in time t0 collides with other
frames sent in slots that start in [t0-1,t0+1]
5: DataLink Layer
5-53
Pure Aloha Goodput
P(success by given node) = Pr(node transmits) .
Pr(no other node transmits in [t0-1,t0+1]
= Pr(node transmits) .
Pr(no other node transmits in [t0-1,t0] .
Pr(no other node transmits in [t0,t0+1]
= p . (1-p)N-1 . (1-p)N-1
= p . (1-p)2(N-1)
… choosing optimum p and then letting n -> infty ...
even worse
than= .18
slotted Aloha!
= 1/(2e)
5: DataLink Layer
5-54
CSMA (Carrier Sense Multiple Access)
CSMA: listen before transmit:
If channel sensed idle: transmit entire frame
r If channel sensed busy, defer transmission
r human analogy: don’t interrupt others!
5: DataLink Layer
5-55
CSMA collisions
spatial layout of nodes
collisions can still occur:
propagation delay means
two nodes may not hear
each other’s transmission
collision:
entire packet transmission
time wasted
note:
role of distance & propagation
delay in determining collision
probability
5: DataLink Layer
5-56
CSMA/CD (Collision Detection)
CSMA/CD: carrier sensing, deferral as in CSMA
m
m
collisions detected within short time
colliding transmissions aborted, reducing channel
wastage
r collision detection:
m easy in wired LANs: measure signal strengths,
compare transmitted, received signals
m difficult in wireless LANs: received signal strength
overwhelmed by local transmission strength
r human analogy: the polite conversationalist
5: DataLink Layer
5-57
CSMA/CD collision detection
5: DataLink Layer
5-58
“Taking Turns” MAC protocols
channel partitioning MAC protocols:
m share channel efficiently and fairly at high load
m inefficient at low load: delay in channel access,
1/N bandwidth allocated even if only 1 active
node!
Random access MAC protocols
m efficient at low load: single node can fully
utilize channel
m high load: collision overhead
“taking turns” protocols
look for best of both worlds!
5: DataLink Layer
5-59
“Taking Turns” MAC protocols
Polling:
r master node
“invites” slave
nodes to transmit in
turn
r typically used with
“dumb” slave
devices
r concerns:
m
m
m
data
poll
master
data
slaves
polling overhead
latency
single point of
failure (master)
5: DataLink Layer
5-60
“Taking Turns” MAC protocols
Token passing:
r control token passed
from one node to next
sequentially.
r token message
r concerns:
m
m
m
token overhead
latency
single point of failure
(token)
T
(nothing
to send)
T
data
5: DataLink Layer
5-61
Summary of MAC protocols
r channel partitioning, by time, frequency or code
m Time Division, Frequency Division
r random access (dynamic),
m ALOHA, S-ALOHA, CSMA, CSMA/CD
m carrier sensing: easy in some technologies (wire), hard in
others (wireless)
m CSMA/CD used in Ethernet
m CSMA/CA used in 802.11
r taking turns
m polling from central site, token passing
m Bluetooth, FDDI, IBM Token Ring
5: DataLink Layer
5-62