Transcript Introduction

Chapter 1
Foundation
1
Problems
 How to build a scalable network that will
support different applications?
 What is a computer network?
 How is a computer network different from
other types of networks?
 What is a computer network architecture?
Chapter Outline
1.1 Applications
1.2 Requirements
1.3 Network Architecture
1.4 Implementing Network Software
1.5 Performance
3
Chapter Goal
 Exploring the requirements that different
applications and different communities place on
the computer network
 Introducing the idea of network architecture
 Introducing some key elements in
implementing network software
 Define key metrics that will be used to evaluate
the performance of computer network
4
1.1 Applications
 Most people know about the Internet (a
computer network) through applications
 World Wide Web
 Email
 Online social network
 Streaming audio video
 File sharing
 Instant messaging
5
Example of an Application
A multimedia application including video-conferencing
Application Protocols
 URL
 Uniform Resource Locater
 http://www.cs.princeton.edu/~llp/index.html
 HTTP
 Hyper Text Transfer Protocol
 TCP
 Transmission Control Protocol
 17 messages for one URL request
 6 to find the IP (Internet Protocol) address
 3 for connection establishment of TCP
 4 for HTTP request and acknowledgement
 Request: I got your request and I will send the data
 Reply: Here is the data you requested; I got the data
 4 messages for tearing down TCP connection
1.2 Requirements
 Application programmer
 list the services that his application needs: delay
bounded delivery of data
 Network designer
 design a cost-effective network with sharable
resources
 Network provider
 list the characteristics of a system that is easy to
manage
8
Requirements





Building blocks
Switched networks
Addressing and routing
Multiplexing
Inter-process communication
9
Building Blocks
 Nodes: PC, special-purpose hardware…
 hosts
 switches
 Links: coax cable, optical fiber…
 point-to-point
(a)
 multiple access
(b)
10
 Terms
 node
 a computer or a more specialized piece of hardware
 network switch
 a small hardware device that joins multiple
computers together within one local area network
(LAN)
 technically, network switches operate at layer two
(data link layer) of the OSI model
11
 link
 physical medium
 point-to-point
 two nodes share a single physical link
 multiple-access
 more than two nodes share a single physical link
12
Connectivity
 Need to understand the following terminologies
 scale
 link
 nodes
 point-to-point
 multiple access
 switched network
 circuit switched
 packet switched
 packet, message
 store-and-forward
Switched Networks
A network can be defined recursively as...
 two or more nodes
connected by a link, or
Switched network
two or more networks
connected by two or
more nodes
Interconnection of networks
14
Strategies
 Circuit switching
 original telephone network
 carry bit streams
15
 Packet switching
 store-and-forward messages
 each node first receives a complete packet over some
link
 stores the packet in its internal memory
 forwards the complete packet to the next node
 multiplex multiple flows of data over a single
physical link
 example: Internet
16
Addressing and Routing
 Address
 byte-string that identifies a node
 usually unique (IP address, MAC address)
 Routing
 process of how to forward messages to the
destination node based on its address
17
 Types of address
 unicast: node-specific
 broadcast: all nodes on the network
 multicast: some subset of nodes on the network
18
 IP address (Internet Protocol address)
 a unique address that certain electronic devices use in
order to identify and communicate with each other on a
computer network utilizing the Internet Protocol
standard (IP)—in simpler terms, a computer address
 any participating network device—including routers,
computers, servers, printers, Internet fax machines, and
some telephones—can have their own unique address
 example: 140.119.164.54
19
 MAC address (Media Access Control address)
 a MAC address or EHA (Ethernet Hardware Address)
or hardware address or adapter address is a quasi-unique
identifier attached to most network adapters (NICs)
 a number that acts like a name for a particular network
adapter, so, e.g., the network cards (or built-in network
adapters) in two different computers will have different
names, or MAC addresses
20
Multiplexing
 Synchronous Time-Division Multiplexing
(STDM)
 divide time into equal-sized quanta, and in a roundrobin fashion, give each flow a chance to send its
data over the physical link
21
 Frequency-Division Multiplexing (FDM)
Multiplexing multiple logical flows over a single physical link
22
 three flows of data (L1 to R1 and so on)
 multiplexed onto a single physical link by switch 1
 demultiplexed back into separate flows by switch 2
 transmit each flow over the physical link at a
different frequency
 e.g. signals for different TV stations are transmitted
at a different frequency on a physical cable TV link
23
Statistical Multiplexing
 Time-division & interleaved
 the physical link is shared over time (time-division) first data from one flow is transmitted over the physical
link, then data from another flow is transmitted, and so
on (interleaved)
24
 On-demand
 data is transmitted from each flow on demand rather
than during a predetermined time slot
 if only one flow has data to send, it gets to transmit that
data without waiting for its quantum to come around
and thus without having to watch the quanta assigned to
the other flows go by unused
 this avoidance of idle time gives packet switching its
efficiency
25
 Schedule link on a per-packet basis
 once a flow begins sending data, we need some way to
limit the transmission, so that the other flows can have a
turn
 an upper bound on the size of the block of data (packet)
is defined that each flow is permitted to transmit at a
given time
 the source may need to fragment the message into
several packets, with the receiver reassembling the
packets back into the original message
26
 each flow sends a sequence of packets over the
physical link, with a decision made on a packet-bypacket basis as to which flow’s packet to send next
 if only one flow has data to send, then it can send a
sequence of packets back-to-back
 should more than one of the flows have data to
send, then their packets are interleaved on the link
27
 Scheduling methods
 FIFO (First-In-First-Out)
 a fair scheduling method
 RR (Round-Robin)
 transmit the packets from each of the different
flows that are currently sending data
 ensure that certain flows receive a particular
share of the link bandwidth or that they never
have their packets delayed in the switch for more
than a certain length of time
28
 QoS (Quality of Service)
 a network that attempts to allocate bandwidth to
particular flows according service priorities
29
■■■
 Congested
 in the following figure, the switch has to multiplex three
incoming packet streams onto one outgoing link
 it is possible that the switch will receive packets faster
than the shared link can accommodate
 in this case, the switch is forced to buffer these packets
in its memory
 should a switch receive packets faster than it can send
them for an extended period of time, then the switch
will eventually run out of buffer space, and some
packets will have to be dropped
30
 when a switch is operating in this state, it is said to be
congested
■■■
A switch multiplexing packets from multiple sources onto one shared link
31
Inter-Process Communication
 Turn host-to-host connectivity into process-toprocess communication
 Fill gap between what applications expect and what
the underlying technology provides
Processes communicating over an abstract channel
32
 Figure
 cloud: abstractly represent connectivity among a set of
computers
 channel: connect one process to another
 view the network as providing logical channels over
which application-level processes can communicate
with each other, each channel provides the set of
services required by that application
33
Types of Communication Channels
 Request/reply channel
 applications
 file transfer
 digital library
 delivery
 guarantee every message sent by one side is received
by the other side and that only one copy of each
message is delivered
34
 privacy and integrity
 might protect the privacy and integrity of the data
that flows over it
 unauthorized parties cannot read or modify the data
being exchanged between the client and server
processes
35
 Message stream channel
 applications
 video-on-demand
 videoconferencing
 delivery
 might not need to guarantee that all messages are
delivered, since a video application can operate
adequately even if some video frames are not
received
36
 sequence
 need to ensure the messages are delivered arrive in
the same order in which they were sent, to avoid
displaying frames out of sequence
 privacy and integrity
 might want to ensure the privacy and integrity of the
video data
 might need to support multicast, so that multiple
parties can participate in the teleconference or view
the video
37
What Goes Wrong in the Network?
(Reliability)
 Bit-level errors
 a “1” is turned into a “0” or vice versa
 bit errors
 single bit is corrupted
 burst errors
 consecutive bits are corrupted
 causes (outside forces of electrical interference)
 lightning strikes, power surges, and microwave
ovens, etc. interfere with the transmission of data
38
 bit error rate
 one out of every 106 to 107 bits on a typical copperbased cable
 one out of every 1012 to 1014 bits on a typical optical
fiber
 Packet-level errors (congestion)
 a complete packet is lost by the network
 the packet contains an uncorrectable bit error and
therefore has to be discarded
39
 causes
 one of the nodes that has to handle the packet, e.g., a
switch that is forwarding it from one link to another,
is so overloaded that it has no place to store the
packet, and therefore is forced to drop it
 Node and link level failures
 a physical link is cut or the computer it is connected to
crashes
40
 causes
 software crashes, power failure, misconfiguration of a
network device
 Others
 messages are delayed
 messages are deliver out-of-order
 third parties eavesdrop
41
1.3 Network Architecture
 Layering and protocols
 OSI architecture
 Internet architecture
42
Layering
 The services provided at the high layers
 implemented in terms of the services provided by
the lower layers
 Abstraction defines a unifying model that can
 capture some important aspect of the system
 encapsulate this model in an object that provides an
interface that can be manipulated by other
components of the system
 hide the details of how the object is implemented
from the users of the object
43
 Use abstractions to hide complexity of the network
from application writers
 Abstractions naturally lead to layering
 start with the services offered by the underlying
hardware
 add a sequence of layers, each providing a higher
(more abstract) level of service
 the services provided at the high layers are
implemented in terms of the services provided by
the low layers
44
 Host-to-host connectivity
 attracts away the fact that there may be an
arbitrarily complex network topology between any
two hosts
Example of a layer network system
45
 Process-to-process channels
 builds on the available host-to-host communication
service
 attracts away the fact that the network occasionally
loses messages
46
 One provides a request/reply service and one supports
a message stream service at the same process-toprocess channel
Layered system with alternative abstractions
available at a given layer
47
 Layering provides two nice features
 decomposes the problem of building a network into
more manageable components
 may implement several layers, each of which
solves one part of the problem
 provides a more modular design
 to add some new service, may only need to
modify the functionality at one layer, reusing the
functions provided at all the other layers
48
Protocols
 Protocol
 used to provide a communication service that
higher-level objects (e.g. application processes,
higher-level protocols) use to exchange messages
 e.g. request/reply protocol, message stream
protocol
 building blocks of a network architecture
49
 Each protocol object defines two different
interfaces
 service interface
 defines a service interface to the other objects on
the same computer that want to use its
communication services
 defines the operations that local objects can
perform on this protocol
50
 examples
 a request/reply protocol would support
operations by which an application can send
and receive messages
 an implementation of the HTTP protocol
could support an operation to fetch a page of
hypertext from a remote server
 an application such as a web browser would
invoke such an operation whenever the
browser needs to obtain a new page
51
 peer-to-peer interface
 defines a peer interface to its counterpart (peer)
on another machine
 defines the form and meaning of messages
exchanged between protocol peers
52
 examples, in the case of HTTP, the protocol
specification defines in detail
 how a "GET" command is formatted
 what arguments can be used with the
command
 how a web server should respond when it
receives such a command
53
 summary: a protocol defines
 a communication service that it exports locally
(the service interface)
 a set of rules governing the messages that the
protocol exchanges with its peer(s) to implement
this service (the peer interface)
54
Service and peer interfaces
55
Protocol Machinery
 Peer-to-peer is direct only at hardware level
 Most peer-to-peer communication is indirect
 each protocol communicates with its peer by
passing messages to some lower-level protocol,
which in turn delivers the message to its peer
 Protocol graph
 there are potentially multiple protocols at any given
level, each providing a different communication
service
 protocol graph represents the suite of protocols that
make up a network system
56
 Example of a protocol graph (nodes are the protocols,
edges are “depends on” relations)
Host 1
File
application
Digital
library
application
Video
application
Host 2
File
application
Digital
library
application
Video
application
57
 process-to-process channels
 RRP: Request Reply Protocol
 MSP: Message Stream Protocol
 host-to-host protocol (provides a host to host
connectivity service)
 HHP: Host-to-Host Protocol
58
 hardware level
 peers directly communicate with each other over
a link
 the applications are said to employ the services of
the protocol stack RRP/HHP or MSP/HHP
59
Encapsulation (header/body)
Host
Host
Application
Application
program
program
Application
Application
program
program
Data
Data
RRP
RRP
RRP
Data
RRP
HHP
Data
HHP
HHP
RRP
Data
High-level messages are encapsulated inside of low-level messages
60
 Operation flow
 host1
 application sends a message to its peer by passing
the message to protocol RRP (uninterpreted)
 RRP
 communicates control info to its peer, instructing it
how to handle the message when it is received
 attaches a “header” to the message
61
 header
 a small data structure - from a few bytes to a few
dozen bytes
 usually attached to the front of a message
 body (or payload)
 the rest of the message
 data
 application data is “encapsulated” in the new
message created by protocol RRP
62
 encapsulation
 high-level messages are encapsulated inside of lowlevel messages
 the process of encapsulation is repeated at each level
of the protocol graph
 inspection & process
 nodes in the network (e.g., switches and routers) may
inspect the HHP header at the front of the message
63
Multiplexing and Demultiplexing
 A fundamental idea of packet switching is to multiplex
multiple flows of data over a single physical link
 The same idea applies up and down the protocol graph
 The header that RRP attaches to its messages contains
an identifier that records the application to which the
message belongs
 We call this identifier RRP’s demultiplexing key, or
demux key
64
 Source host
 at the source host, RRP includes the appropriate demux key
in its header
 Destination host
 when the message is delivered to RRP on the destination
host, it
 strips its header
 examines the demux key
 demultiplexes the message to the correct application
65
ISO Architecture
 ISO / OSI (International Standard Organization /
Open Systems Interconnection)
 ISO
 the ISO, usually in conjunction with ITU
(International Telecommunications Union),
publishes a series of protocol specifications (X dot)
based on the OSI architecture
 X dot series: X.25, X.400, X.500
66
 OSI
 defines a partitioning of network functionality into
seven layers
 not a protocol graph, but rather a reference model
for a protocol graph
67
Description of OSI Layers
68
 Layer 7: Application layer
 interfaces directly to and performs common
application services for application processes
 issues requests to presentation layer
69
 Layer 6: Presentation layer
 transforms data to provide a standard interface for the
Application layer
 MIME encoding, data encryption and similar manipulation of
the presentation are done at this layer to present data as a
service or protocol that the developer sees fit
 MIME (Multipurpose Internet Mail Extensions，多用
途互聯網郵件擴展)
 一個互聯網標準，它擴展了電子郵件標準，使其能支
源非ASCII字符、二進制格式附件等多種格式的郵件
消息
70
 examples
 converts an EBCDIC-coded text file to an ASCII-coded
file
 EBCDIC (Extended Binary Coded Decimal
Interchange Code)
 a character encoding used to represent
Unicode characters


為IBM於1963-1964年間推出的字元編碼表，根據早
期打孔機式的二進化十進數(BCD， Bindary Coded
Decimal)排列而成
為IBM迷你級以上電腦的標準
71
 serialize objects and other data structures into
and out of XML
 serialization
 the process of saving an object onto a
storage medium (such as a file, or a
memory buffer) or to transmit it across a
network connection link in binary form
 XML (eXtensible Markup Language)
 an extensible language that allows users to
define their own tags
72
 Layer 5: Session layer
 controls the dialogues / connections (sessions)
between computers
 establishes, manages and terminates the
connections between the local and remote
application
 provides for either full-duplex or half-duplex
operation
 establishes checkpoint, adjournment, termination,
and restart procedures
73
 Layer 4: Transport layer
 provides reliable data transfer services to the upper
layers
 controls the reliability of a given link through flow
control, segmentation/ desegmentation, and error
control
 some protocols are state and connection oriented
 i.e. the transport layer can keep track of the
segments and retransmit those that fail
74
 the best known example: Transmission Control
Protocol (TCP)
 the transport layer is the layer that converts
messages into TCP segments or User Datagram
Protocol (UDP), Stream Control Transmission
Protocol (SCTP), etc.
75
 Layer 3: Network layer
 provides the functional and procedural means of
transferring variable length data sequences from a
source to a destination via one or more networks
while maintaining the quality of service requested
by the Transport layer
 performs network routing functions, and might also
perform fragmentation and reassembly, and report
delivery errors
76
 “routers” operate at this layer
 send data throughout the extended network and
make the Internet possible
 there is a logical hierarchical addressing scheme
 the best known example: Internet Protocol (IP)
77
 Layer 2: Data Link layer
 provides the functional and procedural means to
transfer data between network entities and to detect
and possibly correct errors that may occur in the
Physical layer
 the best known example: Ethernet
 this layer manages the interaction of devices with a
shared medium
78
 other examples
 HDLC and ADCCP for point-to-point or packetswitched networks
 HDLC (High-Level Data Link Control)
 a bit-oriented synchronous data link layer
protocol developed by ISO
 HDLC can be used for point to multipoint
connections, but is now used almost
exclusively to connect one device to
another
79
 ADCCP (Advanced Data Communication
Control Procedures (or Protocol))
 a bit-oriented data link layer protocol used
to provide point-to-point and point-tomultipoint transmission of data frames that
contain error control information
80
 Aloha for local area networks
 Aloha
 a pioneering computer networking system
developed at the University of Hawaii
 first deployed in 1970
 one of the core concepts in the network is the
basis for the widely used Ethernet
 the first version of the protocol was basic
 if you have data to send, send the data
 if the message collides with another
transmission, try resending "later”
81
 Slotted Aloha
 an improvement to the original Aloha
protocol
 introduces discrete timeslots and increased
the maximum throughput
 a station can send only at the beginning of
a timeslot, and thus collisions are reduced
82
83
 on IEEE 802 local area networks (LANs), and some
non-IEEE 802 networks such as FDDI, this layer
may be split into Media Access Control (MAC)
layer and Logical Link Control (LLC) layer
 FDDI (Fiber Distributed Data Interface)
 provides a standard for data transmission in a
LAN that can extend in range up to 200
kilometers (124 miles)
84
 MAC
 a layer 2 sub-layer that provides addressing and
channel access control mechanisms that makes it
possible for several terminals or network nodes to
communicate within a multipoint network,
typically a LAN or MAN
 acts as an interface between the Logical Link
Control (LLC) sub-layer and the network's
physical layer
85
 LLC
 a sub-layer primarily concerned with
 multiplexing protocols transmitted over the MAC
layer (when transmitting) and demultiplexing them
(when receiving)
 providing flow control and detection and
retransmission of dropped packets, if requested
 the protocol used for LLC in IEEE 802 networks
and in some non-IEEE 802 networks such as
FDDI is specified by the IEEE 802.2 standard
86
 arranges bits from the physical layer into logical
chunks of data, known as frames
 “bridges” and “switches” operate at this layer
 connectivity is provided only among locally
attached network nodes forming layer 2 domains
for unicast or broadcast forwarding
 other protocols may be imposed on the data frames
to create tunnels and logically separated layer 2
forwarding domain
87
 Layer 1: Physical layer
 defines all the electrical and physical specifications
for devices
 includes the layout of pins, voltages, and cable
specifications
88
 hubs, repeaters, network adapters and Host Bus
Adapters (HBAs used in Storage Area Networks
(SAN)) are physical-layer devices
 HBA
 connects a host system (the computer) to other
network and storage devices
 SAN
 an architecture to attach remote computer storage
devices (such as disk arrays, tape libraries and optical
jukeboxes) to servers in such a way that, to the
operating system, the devices appear as locally
attached
89
 major functions and services performed by the
physical layer
 establishment and termination of a connection to a
communications medium
90
switch
hub, repeater,
network adapter,
HBA
OSI Network Architecture
92
 Operations
 physical layer
 handles the transmission of raw bits over a
communications link
 data link layer
 collects a stream of bits into a larger aggregate called a
frame
 network adaptors, along with device drivers running
in the node’s OS, typically implement the data link
level
 this means that, frames, not raw bits, are actually
delivered to hosts
93
 network layer
 handles routing among nodes within a packet-switched
network
 at this layer, the unit of data exchanged among nodes is
typically called a packet rather than a frame
 [note]
 the lower three layers are implemented on all network
nodes, including switches within the network and hosts
connected along the exterior of the network
94
 transport layer
 implements a process-to-process channel
 the unit of data exchanged is commonly called a
message rather than a packet or a frame
 the transport layer and higher layers typically run only
on the end hosts and not on the intermediate switches
or routers
95
 session layer
 provides a name space that is used to tie together the
potentially different transport streams that are part of
a single application
 example
 it might manage an audio stream and a video
stream that are being combined in a
teleconferencing application
96
 presentation layer
 concerned with the format of data exchanged between
peers, for example,
 whether an integer is 16, 32, or 64 bits long
 whether the most significant byte is transmitted first
or last
 how a video stream is formatted
 application layer
 protocols include things like the File Transfer Protocol
(FTP), which defines a protocol by which file transfer
applications can interoperate
97
Internet Architecture (TCP/IP Architecture)
 The Internet architecture evolved out of experiences
with an earlier packet-switched network called the
ARPANET
 Both Internet and ARPANET were funded by the
Advanced Research Projects Agency (ARPA), one of
the R&D funding agencies of the U.S. Department of
Defense (DoD)
 Internet and ARPANET were around before the OSI
architecture, and the experience gained from building
them was a major influence on the OSI reference model
98
FTP
HTTP
NV
TFTP
UDP
TCP
IP
 Internet
 a four-layer model
 the lowest level
NET1
NET2
■■■
NETn
 a wide variety of network protocols: denoted NET1,
NET2, and so on
 these protocols are implemented by a combination of
hardware (e.g., a network adaptor) and software (e.g.,
network device driver)
 examples
 Ethernet or FDDI protocols
99
FTP
HTTP
NV
TFTP
UDP
TCP
IP
 the second layer
NET
1
NET
2
■■■
 consists of a single protocol: Internet Protocol (IP)
 the protocol that supports the interconnection of
multiple networking technologies into a single, logical
internetwork
 the third layer
 contains two main protocols
 Transmission Control Protocol (TCP) and User
Datagram Protocol (UDP)
 TCP and UDP provide alternative logical channels
to application programs
100
NET
n
 TCP provides a reliable byte-stream channel
 UDP provides an unreliable datagram delivery
channel (datagram may be thought of as a
synonym for message)
 in the language of the Internet, TCP and UDP are
sometimes called end-to-end protocols, although it
is equally correct to refer to them as transport
protocols
FTP
HTTP
NV
TFTP
UDP
TCP
IP
101
NET1
NET
2
■■■
NETn
 the top layer
 application protocols, such as FTP, TFTP (Trivial File
Transport Protocol), Telnet (remote login), and SMTP
(Simple Mail Transfer Protocol, or electronic mail),
that enable the interoperation of popular applications
FTP
HTTP
NV
TFTP
UDP
TCP
IP
NET1
NET2
■■■
NETn
102
 the difference between an application layer protocol
and an application
 all the available different World Wide Web browsers
(Firefox, Safari, Internet Explorer, Lynx, etc) – application
 a similarly large number of different implementations of
web servers – application
 we can use any one of these application programs to access
a particular site on the Web is because they all conform to
the same application layer protocol: HTTP (HyperText
Transport Protocol) – application protocol
 confusingly, the same word sometimes applies to both an
application and the application layer protocol that it uses (e
g., FTP)
103
FTP
HTTP
NV
TFTP
UDP
TCP
IP
NET 1
NET 2
■■■
Internet protocol graph
NET n
Alternative view of Internet architecture
104
1.4 Implementing Network Software
 Application Programming Interface (Sockets)
 Protocol Implementation Issues
105
Application Programming Interface
(Sockets)
 The place to start when implementing a network
application is the interface exported by the network
 network Application Programming Interface (API)
 when we refer to the interface “exported by the
network,” we are generally referring to the interface
that the OS provides to its networking subsystem
 Socket interface
 originally provided by the Berkeley distribution of Unix
 now supported in virtually all popular operating systems
106
 Protocol, API and implementation
 protocol
 provides a certain set of services
 API
 provides a syntax by which those services can be
invoked in this particular OS
 implementation
 responsible for mapping the tangible set of
operations and objects defined by the API onto the
abstract set of services defined by the protocol
107
 Socket
 the main abstraction of the socket interface
 the point where a local application process attaches
to the network
 an interface between an application and the network
 an application creates the socket
108
 Socket interface defines operations of




creating a socket
attaching a socket to the network
sending/receiving messages through the socket
closing the socket
109
Socket API (TCP)
 Create a socket
int socket(int domain, int type, int protocol)
 domain
 specify the socket family that is going to be used
 examples
 PF_INET = Internet family
 PF_UNIX = UNIX pipe facility
 PF_PACKET = direct access to the network interface (i.e.
bypass TCP/IP protocol stack)
110
 type
 indicate the semantics of the communication
 examples
 SOCK_STREAM = a byte stream
 SOCK_DGRAM =a message-oriented service, e.g.
UDP
 protocol
 identify the specific protocol that is going to be used
 example
 UNSPEC (Unspecified)
111
 handle
 the return value from newly created socket
 an identifier by which we can refer to the socket in the
future
 it is given as an argument to subsequent operations on this
socket
112
Creating a Socket
int sockfd = socket(address_family, type, protocol);
 The socket number returned is the socket descriptor for the
newly created socket
int sockfd = socket (PF_INET, SOCK_STREAM, 0);
int sockfd = socket (PF_INET, SOCK_DGRAM, 0);
The combination of PF_INET and SOCK_STREAM implies
TCP
Client-Serve Model with TCP
 Server
 passive open
 prepares to accept connection, does not actually
establish a connection
 Server invokes
int bind(int socket, struct sockaddr *addr, int addr_len)
int listen(int socket, int backlog)
int accept(int socket, struct sockaddr *addr, int addr_len)
 bind operation
 binds the newly created “socket” to the specified
“address” (the server address)
 when used with Internet Protocols, “address” is a data
structure that includes
 the IP address of the server
 a TCP port number
 used to indirectly identify a process
 usually some well-known number specific to the service
being offered; e.g., web servers commonly accept
connections on port 80
115
 listen operation
 defines how many connections can be pending on the
specified “socket”
 accept operation
 carries out the passive open
 it is a blocking operation that does not return until a
remote participant has established a connection,
 when it does complete, it returns a new socket that
corresponds to this new established connection
116
 the “address” argument contains the remote
participant’s address
 when accept returns, the original socket that was given
as an argument still exists and still corresponds to the
passive open; it is used in future invocations of accept
117
Client-Serve Model with TCP
 Client
 application performs active open
 it says who it wants to communicate with by invoking
“connect”
 Client invokes
int connect(int socket, struct sockaddr *addr, int addr_len)
 connect operation
 it does not return until TCP has successfully
established a connection, at which time the
application is free to begin sending data
 “address” contains the remote participant’s address
119
Client-Serve Model with TCP
 In practice
 the client usually specifies only remote
participant’s address and let the system fill in the
local information
 whereas a server usually listens for messages on a
well-known port
 a client does not care which port it uses for itself,
the OS simply selects an unused one
Client-Serve Model with TCP
 Once a connection is established, the application
process invokes the following two operations to send
and receive data
int send(int socket, char *msg, int mlen, int flags)
int recv(int socket, char *buf, int blen, int flags)
 send operation
 it sends the given message over the specified socket
 receive operation
 it receives a message from the specified “socket” into
the given “buffer”
 both “send” and “receive” take a set of “flags” that
control certain details of the operation
122
Example Application: Client
#include <stdio.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <netdb.h>
#define SERVER_PORT 5432
#define MAX_LINE 256
int main(int argc, char * argv[])
{
FILE *fp;
struct hostent *hp;
struct sockaddr_in sin;
char *host;
char buf[MAX_LINE];
int s;
int len;
if (argc==2) {
host = argv[1];
}
else {
fprintf(stderr, "usage: simplex-talk host\n");
exit(1);
}
Example Application: Client
/* translate host name into peer’s IP address */
hp = gethostbyname(host);
if (!hp) {
fprintf(stderr, "simplex-talk: unknown host: %s\n", host);
exit(1);
}
/* build address data structure */
bzero((char *)&sin, sizeof(sin));
sin.sin_family = AF_INET;
bcopy(hp->h_addr, (char *)&sin.sin_addr, hp->h_length);
sin.sin_port = htons(SERVER_PORT);
/* active open */
if ((s = socket(PF_INET, SOCK_STREAM, 0)) < 0) {
perror("simplex-talk: socket");
exit(1);
}
if (connect(s, (struct sockaddr *)&sin, sizeof(sin)) < 0) {
perror("simplex-talk: connect");
close(s);
exit(1);
}
/* main loop: get and send lines of text */
while (fgets(buf, sizeof(buf), stdin)) {
buf[MAX_LINE-1] = ’\0’;
len = strlen(buf) + 1;
send(s, buf, len, 0);
}
}
Example Application: Server
#include <stdio.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <netdb.h>
#define SERVER_PORT 5432
#define MAX_PENDING 5
#define MAX_LINE 256
int main()
{
struct sockaddr_in sin;
char buf[MAX_LINE];
int len;
int s, new_s;
/* build address data structure */
bzero((char *)&sin, sizeof(sin));
sin.sin_family = AF_INET;
sin.sin_addr.s_addr = INADDR_ANY;
sin.sin_port = htons(SERVER_PORT);
/* setup passive open */
if ((s = socket(PF_INET, SOCK_STREAM, 0)) < 0) {
perror("simplex-talk: socket");
exit(1);
}
Example Application: Server
if ((bind(s, (struct sockaddr *)&sin, sizeof(sin))) < 0) {
perror("simplex-talk: bind");
exit(1);
}
listen(s, MAX_PENDING);
/* wait for connection, then receive and print text */
while(1) {
if ((new_s = accept(s, (struct sockaddr *)&sin, &len)) < 0) {
perror("simplex-talk: accept");
exit(1);
}
while (len = recv(new_s, buf, sizeof(buf), 0))
fputs(buf, stdout);
close(new_s);
}
}
1.5 Performance





Performance metrics
Bandwidth versus latency
Delay ╳ bandwidth product
High-speed networks
Application performance needs
127
Performance Metrics
 Network performance is measured in
 bandwidth (also called throughput)
 latency (also called delay)
 Bandwidth
 literally a measure of the width of a frequency band
 example
 a voice-grade telephone line supports a frequency band
ranging from 300 to 3,300 Hz
(Hz = the number of complete cycles per second)
 it is said to have a bandwidth of 3,300Hz - 300Hz =
3,000Hz
128
 bandwidth
 the range of signals that can be accommodated
measured in hertz
 bandwidth of a communication link
 the number of bits per second that can be
transmitted over a link
 example
 the bandwidth of an Ethernet is 10 Mbps (10
million bits/second)
129
 bandwidth is sometimes thought in terms of how
long it takes to transmit each bit of data
 example
 on a 10-Mbps network, it takes 0.1
microsecond (μs) to transmit each bit
130
 we can think of




a second of time：a distance that we could measure
bandwidth：how many bits fit in that distance
each bit：a pulse of some width
example
 each bit on a 1-Mbps link is 1 μs wide
 each bit on a 2-Mbps link is 0.5 μs wide
131
Bits transmitted at a particular bandwidth can be regarded as having some width:
(a)bits transmitted at 1 Mbps (each bit 1 μs wide);
(b)bits transmitted at 2 Mbps (each bit 0.5 μs wide)
132
 Bandwidth requirements of an application
 the number of bits per second that it needs to
transmit over the network to perform acceptably
 Throughput
 the measured performance of a system
 because of various inefficiencies of implementation,
a pair of nodes connected by a link with a
bandwidth of 10 Mbps might achieve a throughput
of only 2Mbps
133
 Latency (delay)
 corresponds to how long it takes a message to
travel from one end of a network to the other
(one-way)
 measured strictly in terms of time
 example
 a transcontinental network might have a
latency of 24 milliseconds (ms)
 i.e., it takes a message 24 ms to travel from
one end of North America to the other
134
 Latency = Propagation delay + Transmit delay +
Queuing delay
 Propagation delay = Distance / SpeedOfLight
 light travels across different mediums at different
speeds, examples
 3.0 × 108 m/s in a vacuum
 2.3 × 108 m/s in a cable
 2.0 × 108 m/s in a fiber
 Transmit delay = Packet size / Bandwidth
 Queuing delay = the time the packet switches
takes to store packets for some time before
forwarding them on an outbound link
135
 Round-trip time (RTT)
 how long it takes to send a message from one end
of a network to the other and back
136
Bandwidth versus Latency
 Relative importance (depends on applications)
 latency dominates bandwidth (latency bound)
 example: a client sends a 1-byte message to a server and
receives a 1-byte message in return (latency bound)
 the application will perform much differently on a
transcontinental channel with a 100-ms RTT than it will
on an across-the-room channel with a 1-ms RTT
 whether the channel is 1 Mbps or 100 Mbps is relatively
insignificant, however, since the former implies that the
time to transmit a byte (Transmit) is 8 μs and the latter
implies Transmit = 0.08 μs
137
 bandwidth dominates latency (bandwidth bound)
 example: a digital library program that is being asked
to fetch a 25MB image
 suppose that the channel has a bandwidth of 10 Mbps
 it will take 20 seconds to transmit the image,
making it relatively unimportant if the image is on
the other side of a 1-ms channel or a 100-ms
channel
 the difference between a 20.001-second response
time and a 20.1-second response time is negligible
138
 Summary
 for large file transfer, bandwidth is critical
 for small messages (HTTP, NFS, etc.), latency is
critical
139
 The following graph shows how long it takes to move
objects of various sizes (1 byte, 2KB, 1MB) across
networks with
 RTTs ranging from 1 to 100 ms
 link speeds of either 1.5 or 10 Mbps
140
p. 48 of 5th ed.
pp. 43-44 of 4th ed.
141
142
Delay ╳ Bandwidth Product
 Channel
 between a pair of processes as a hollow pipe
 Latency (delay)
 the length of the pipe
 Bandwidth
 the diameter of the pipe
 Delay × bandwidth
 the volume of the pipe
 i.e. the maximum number of bits that could be in transit
through the pipe at any given instant
143
 Example
 a transcontinental channel with a one-way latency
of 50ms and a bandwidth of 45Mbps can hold
280KB (= 2.25 × 106 bits) of data
144
Sample Delay ╳ Bandwidth Products
Link type
Bandwidth
(Typical)
(Distance
(Typical)
Round-trip
Delay
Delay x BW
Dial-up
56Kbps
10km
87μs
5bits
Wireless
LAN
54Mbps
50m
0.33μs
18bits
Satellite
45Mbps
35,000 km
230ms
10MB
Crosscountry fiber
10Gbps
4,000km
40ms
400MB
145
High-Speed Networks
 Example
 transmit a 1-MB file over a 1-Mbps network vs. over a 1-Gbps
network, both of which have an RTT of 100 ms (“high speed”
does not mean that latency improves at the same time as
bandwidth)
 1-Mbps network
 delay × bandwidth = 0.1Mb
 it takes 80 [= (1/0.1)*8] RTTs to transmit the file
 during each RTT, 1.25% of the file is sent
 1-Gbps network
 delay × bandwidth = 12.5 [= 0.1 * (1000/8)] MB
 it takes < 1 [= (1/12.5)*8] RTT to transmit the file
146
Relationship between bandwidth and latency.
A 1-MB file would fill the 1-Mbps link 80 times,
but only fill the 1-Gbps link 1/12 of one time.
*More data can be transmitted during each RTT on a high-speed network
147
Effective End-to-End Throughput
 Throughput = TransferSize / TransferTime
 TransferTime = RTT + (1/Bandwidth) x TransferSize
 TransferTime = one-way latency plus any
additional time spent requesting or setting up the
transfer
 RTT = a request message being sent across the
network and the data being sent back
 in a high-speed network (infinite bandwidth), RTT
dominates TransferTime
148
 Example
 a user wants to fetch a 1-MB file across a 1-Gbps
with a round-trip time of 100ms
 TransferTime = 100-ms (RTT) + transmit time for
1MB (1/1Gbps × 1MB = 8ms) = 108ms
 effective throughput = 1MB/108ms = 74.1Mbps
(not 1Gbps)
149
 Discussions
 transferring a larger amount of data will help
improve the effective throughput
 where in the limit, an infinitely large transfer size
will cause the effective throughput to approach the
network bandwidth
150
Application Performance Needs
 Some applications are able to state an upper limit on how
much bandwidth they need
 example
 suppose one wants to stream a video image; that is
one-quarter the size of a standard TV image; i.e., it
has a resolution of 352 by 240 pixels
 if each pixel is represented by 24 bits of information
(24-bit color), then the size of each frame would be
(352 × 240 × 24)/8 = 247.5 KB
151
 if the application needs to support a frame rate of 30
frames per second, then it might request a throughput
rate of 75 Mbps
 because the difference between any two adjacent
frames in a video stream is often small, it is possible
to compress the video by transmitting only the
differences between adjacent frames
152
 this compressed video does not flow at a constant
rate, but varies with time according to factors such as
 the amount of action
 detail in the picture
 the compression algorithm
 it is possible to say what the average bandwidth
requirement will be, but the instantaneous rate may
be more or less
153
 Jitter
 the variation in latency
 example
 the source sends a packet once every 33 ms, as
would be the case for a video application
transmitting frames 30 times a second
 if the packets arrive at the destination spaced out
exactly 33 ms apart, then the delay experienced by
each packet in the network was exactly the same
154
 if the spacing between when packets arrive at the
destination (interpacket gap) is variable, however,
then the delay experienced by the sequence of
packets must have also been variable, and the
network is said to have introduced jitter into the
packet stream
 such variation is generally not introduced in a single
physical link, but it can happen when packets
experience different queuing delays in a multihop
packet-switched network
155
 this queuing delay corresponds to the Queue component of
latency, which varies with time
Network-induced jitter
156
 Relevance of jitter
 suppose that the packets being transmitted over the
network contain video frames, and in order to display
these frames on the screen the receiver needs to receive
a new one every 33 ms
 if a frame arrives early, then it can simply be saved by
the receiver until it is time to display it
 if a frame arrives late, then the receiver will not have the
frame it needs in time to update the screen, and the
video quality will suffer; it will not be smooth
157
 if the receiver knows the upper and lower bounds on the
latency that a packet can experience, it can delay the
time at which it starts playing back the video (i.e.,
displays the first frame) long enough to ensure that in
the future it will always have a frame to display when it
needs it
 the receiver delays the frame, effectively smoothing out
the jitter, by storing it in a buffer
158