ECEN 5032 Data Networks Introduction Peter Mathys - - PowerPoint PPT Presentation

ECEN 5032 Data Networks

Introduction

Peter Mathys

mathys@colorado.edu

University of Colorado, Boulder

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.1/63

ARPANET 1969

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.2/63

Creation of ARPA

1957 October 4. The USSR launches Sputnik, the first artificial earth satellite. 1958 February 7. In response to Sputnik launch, the US Department of Defense issues directive 5105.15, establishing the Advanced Research Projects Agency (ARPA). 1962 J.C.R. Licklider is chosen to head ARPA’s research in improving the military’s use of computer technology. To quickly expand technology, Licklider moved ARPA’s contracts from the private sector to universities. This laid the foundations for what would become the ARPANET.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.3/63

Computer/Communications History

1948 Transistor developed at Bell Laboratories. 1951 UNIVAC I machine installed at US Census Bureau. 1959 Jack St. Clair Kilby of TI invents integrated circuit. 1962 Telstar, first communication satellite, launched. 1962 First PCM transmitting system used by Bell Telephone. 1965 DEC PDP-8, first successful minicomputer, $18,500. 1969 Creation of ARPANET computer network. 1970 Norman Abramson implements ALOHANET in Hawaii. 1971 Engineers at Intel invent microprocessor. 1974 First personal computers introduced. 1976 Ethernet invented by Bob Metcalfe. 1981 IBM introduces personal computer. 1983 ARPANET split into Internet and Milnet. 1984 Court ruling ends AT&T telephone monopoly. 1991 World Wide Web begins.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.4/63

Computer History Outline

Mainframes (1950’s-1970’s). Centralized computers (filling a whole room) for large businesses. Users bring their work (punched cards) to the computer. Companies: Univac, IBM, CDC. Minicomputers (1960’s-1990’s). Smaller and less expensive, but still exclusively business computers. Users connect with terminals via telephone lines. Companies: Digital (DEC), Data General, HP , Sun. Microprocessors (1970’s-now). Started as chipset for programmable calculator. First “home computer”: Altair 8800 ($397), based on 8080 processor. Companies: Intel, Zilog, Motorola, TI. Personal Computers (1980’s-now). On August 12, 1981, the first IBM home computer goes on sale (8088 processor, PC-DOS operating system). Companies: IBM, Apple, Compaq and many more.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.5/63

Networks: Business Applications

Resource Sharing. Make all programs, equipment, and especially data available to anyone on the network without regard to the physical location of the resource and the user.

• Communication. Exchange of messages between

employees, collaboration between users at different locations, videoconferencing, etc.

• B2B. Business-to-business: Doing business

electronically with other companies, especially suppliers and customers.

• B2C. Business-to-consumer or e-commerce: Doing

business over the Internet directly with consumers.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.6/63

A Fatal Mistake

Ken Olsen, 1977 (President of Digital Equipment Corporation). When asked why Digital is not going after personal computer market in a big way, he responded: “There is no reason for any individual to have a computer in his home.” Today Digital, which was once the world’s computer maker # 2 (behind IBM), no longer exists!

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.7/63

Networks: Home Applications

Word Processing, Games. These were the initial uses in the early days of personal computers. Internet Access. Access to remote information on the World Wide Web. Includes government, health, hobbies, newspapers, sports, travel, etc. Person-to-Person Communication. E-mail, instant messaging, chat rooms, newsgroups, etc. P2P (Peer-to-Peer). File exchange directly between end-users. A well-known example is (mostly illegal) swapping of music files. Other applications are sharing

• f family photos, or playing multiperson on-line games.
• Entertainment. Downloading music files is already a
• reality. Video on demand might become the killer

application a couple of years down the road.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.8/63

Transmission Technology

Classification of networks by transmission technology:

Point-to-point: One sender, one receiver. Broadcast: One sender, many receivers. Multi-access: Many senders, one receiver. Broadcast network: Many senders, many receivers.

Multicasting: Transmission of data packets to a subset

• f machines.

Unicasting: Another term for point-to-point transmission.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.9/63

Network Scale

Classification of networks by distance or scale.

Interprocessor distance Processors located in same 10,000 km 1,000 km 100 km 10 km 1 km 100 m 10 m 1 m Planet Continent Country City Campus Building Room Cubicle

• ✁✄✂

The Internet Wide area Metropolitan area Local area Personal area Example network

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.10/63

Local Area Networks (LAN)

Privately-owned networks within single building or

• campus. Up to a few kilometers in size.

LANs are distinguished from other networks by

Size

• Worst-case transmission time is short, bounded,

and known in advance. Transmission technology. Many senders, many receivers all (logically) connected to the same wire

• Broadcast

networks.

• Topology. Examples: IEEE 802.3 (Ethernet) bus

topology, IEEE 802.5 (IBM token ring) ring topology.

Broadcast networks can be static (e.g., time slots assigned to users in advance) or dynamic. Dynamic allocation methods can be either centralized (e.g., bus arbitration unit) or decentralized (each machine decides for itself).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.11/63

Metropolitan Area Networks (MAN)

A MAN covers roughly the area of a city. Cable TV is an example of a MAN. Until the late 1990’s cable TV was intended for (re)broadcasting TV only. As the Internet and the WWW became more popular, cable TV operators realized that they had a faster pipe into homes than the telephone company. Recent development: IEEE 802.16 Wireless MAN or Broadband Wireless Access (BWA). Intended for fast deployment of high-speed networks between fixed locations (buildings). Frequency band: 10-66 GHz.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.12/63

Wide Area Networks (WAN)

Spans large geographical area, e.g., country or continent. Hosts (user machines) are connected by communication subnet. Subnet is owned and/or operated by a telephone company and/or an Internet service provider. Pure communication aspects (subnet) areseparated from application aspects, which simplifies complete network design. Subnet consists of two distinct components: Transmission lines and Routers. Transmission lines: Point-to-point links using wire, coax, optical fiber, wireless, etc. Routers (switching elements): Specialized computers that connect three or more transmission lines.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.13/63

Packet Switching

The telephone network uses circuit switching to make a connection between two users. Data networks use packet switching to share links. Packets are moved, one hop at a time, through the subnet from the sender to the receiver via routers. At each intermediate router packet is stored and waits for its turn to be forwarded along required output link. This is called store-and-forward or packet-switching. Each router has to decide (locally) along which link packets are forwarded to the next router. Different routing algorithms (e.g., distance vector routing and link state routing) exist to make these decisions. Most WANs use packet-switching. Exception: Satellite networks which are inherently broadcast.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.14/63

Wireless Networks

Wireless networks can be subdivided into three categories:

System interconnection. Short-range radio, e.g., for wireless mouse or printer connection

• Bluetooth

(IEEE 802.15). Wireless LANs. Similar to Ethernet but without wires. A popular and widespread standard is IEEE 802.11 (802.11a, 802.11b, 802.11g). Wireless MANs. Low-bandwidth example is third generation cellular telephone network. High-bandwidth example is wireless MAN for high-speed Internet access from fixed locations (IEEE 802.16).

Most wireless networks hook up to wired networks for Internet access.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.15/63

Home Networks

In future every device in the home will be capable of communicating with every other device and all of them will be accessible over the Internet. Home networking presents quite a few challenges, especially to software companies. Downloading and installing newest operating system before Internet refrigerator can be used is not attractive. Home networks and devices must be foolproof in

• peration. People will not read a 100-page manual for
• perating their Internet coffee/espresso/latte maker.

Main application may be multimedia

• network needs

sufficient capacity. Other requirements: Low price, start small and expand network gradually, high security and reliability.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.16/63

Internetworks

Users connected to networks with different hardware and software want to exchange data. A collection of interconnected networks is called an internetwork or internet (the global Internet is a specific case of such an internet). The terms subnet, network, and internetwork are often confused.

Subnet: Collection of routers and communication links

• wned by a network operator, e.g., backbone or WAN.

Network: Combination of a subnet and its hosts. LAN: Cable and hosts form the network (no subnet). Internetwork: Network obtained by interconnecting distinct networks, e.g., through gateways and/or firewalls.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.17/63

Network Software

In the early days of computer networks most emphasis was placed on hardware. Software was added on an as-needed basis, once the hardware was in place. Nowadays hardware is (almost) universally programmable and thus the software defines how a network operates. This requires a highly structured approach to software design and implementation to reduce complexity and cost and to improve maintainability and reliability.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.18/63

Protocol Hierarchies

To manage complexity, most networks are organized as a stack of layers, each one relying on the one below and supporting the one above. Each layer offers specific services to the layers above, without burdening them about details of the actual implementation. Similar in programming: Abstract data types, data encapsulation, and object-oriented programming. Communication between two hosts at level

• uses

“layer

• protocol”. A protocol is a series of steps,

involving two or more parties, designed to accomplish a

• task. The entities which carry out the protocol at the

same level are called peers. A 5-layer stack is shown in the figure on the next slide.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.19/63

Layers, Protocols, Interfaces

Physical Medium Layer 1 Layer 1 Layer 2 Layer 2 Layer 3 Layer 4 Layer 5 Layer 3 Layer 4 Layer 5

Layer 1 protocol Layer 2 protocol Layer 3 protocol Layer 4 protocol Layer 5 protocol Layer 4/5 interface Layer 3/4 interface Layer 2/3 interface Layer 1/2 interface

Host 1 Host 2

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.20/63

Network Architecture

No data is directly transferred from layer

• at host 1 to

layer

• at host 2. Each layer passes data and control

information to the layer below. Below layer 1 is the physical medium through which actual communication occurs. Interface between layers defines services lower layer provides to upper layer. Clean and concise interface is important design criterion. Network architecture: Set of layers and protocols. Details of implementation and specification of interfaces are not part of architecture. Protocol stack: List of protocols (one per layer) used by a specific system. Book is about network architectures, protocol stacks, protocols.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.21/63

Philosopher Analogy

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.22/63

5-Layer Technical Example

Message

• is produced by application at layer 5 and

given to layer 4 for transmission. Layer 4 puts header for identification in front of message and passes result on to layer 3. Layer 3 breaks message up into packets, labels each with

header, and decides through which line at layer 2 to output. Layer 2 adds another header and a trailer (for error control) and then hands the result over to layer 1 for physical transmission. At the receiving end the sequence is reversed. Headers are stripped off when moving from lower to higher layers. None of the headers (and trailers) for layers below

• are seen by layer
• .

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.23/63

5-Layer Example Figure

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.24/63

Virtual/Actual Communication

Virtual communication: Peer processes at layer

• think of their communication as being “horizontal,”

using layer

• protocol. Each has procedure similar to

SendToOtherSide and GetFromOtherSide. This peer process abstraction is crucial to network

• design. It enables breaking the complex task of

network design into several smaller design problems,

• ne for each layer.

Actual communication: Takes place across layer

• ✁

• interface.

Note: Lower layers of protocol hierarchy are often implemented in hardware or firmware.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.25/63

Design Issues

Adressing: Every layer needs mechanism for identifying senders and receivers. Data transfer rules: Simplex or duplex, transmission priorities (e.g., normal and urgent packets). Error control: Common scheme for error detection and/or correction is needed. Retransmission protocols: Need to keep track of which packets are acknowledged by receiver and retransmit missing ones. Flow control: Need mechanism that prevents fast transmitter from swamping slow receiver.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.26/63

Design Issues (contd.)

Fragmentation: Some processes impose limit on length of packets. Mechanism is needed to split up long messages for transmission and then reassembling them again at receiver. Multiplexing: Combining several low rate communication processes into one high rate process can be less expensive. Routing: If multiple paths exist between sender and receiver, a route must be chosen according to some criterion (speed, cost, laws, etc).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.27/63

Connection-Oriented Services

Connection-oriented: Similar to telephone service. A connection is established first, then data transfer(s) take place, followed by releasing the connection. During connection phase sender and receiver may negotiate parameters to be used, e.g., transmission speed, message size, QoS, etc. Quality of service (QoS): Reliability (no data loss) is important for file transfers, speed (or upper limit on delay) is crucial for voice traffic.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.28/63

Connectionless Services

Connectionless: Similar to mail service. Each message carries full source and destination address and is transmitted through the system independently. Datagram service: Unreliable (meaning not acknowledged) connectionless service. Similar to sending a telegram. Acknowledged datagram service: Similar to sending registered mail and requesting return receipt. Request-reply service: Sender transmits single datagram containing request. The reply contains the answer and serves as an implicit acknowledgement. Commonly used under client-server model.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.29/63

Service Primitives Example

Service primitives for simple connection-oriented service. Client-server interaction for simple connection-oriented service.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.30/63

Services vs. Protocols

Services and protocols are distinct concepts. Service: Set of primitives or operations that a layer provides to the layer above it. Protocol: Set of rules governing format and meaning

• f packets that are exchanged by peer entities within a
• layer. A protocol relates to the implementation of a

service and is not visible to the user of a service.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.31/63

Reference Models

OSI Reference Model (Day and Zimmerman, 1983). OSI stands for Open Systems Interconnection. Proposed by ISO as first step toward international standardization of protocols used in various layers. Protocols associated with OSI models are rarely used in practice. But model itself is quite general and still valid. TCP/IP Model (Cerf and Kahn, 1974). Major design goal from beginning was to connect multiple networks in seamless way. The TCP/IP model is not very general and of little theoretical use, but both the TCP and the IP protocols are very widely used.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.32/63

OSI Reference Model

The OSI model has seven layers. Principles used to arrive at the seven layers:

Create new layer when different level of abstraction is needed. Each layer should perform a well-defined function. Keep international standardization in mind when defining layer functions. Choose layer boundaries to minimize information flow across interfaces. Make number of layers large enough so that distinct functions need not be thrown together. Make number of layers small enough so that architecture does not become unwieldy.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.33/63

OSI Reference Model

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.34/63

OSI Layers

Physical Layer (1): Transmits raw bits over a communication channel. Converts 0’s and 1’s to waveforms and back. Design issues deal with mechanical, electrical, timing interfaces, and physical transmission medium below layer 1. Data Link Layer (2): Transform raw bit transmission pipe of layer 1 into line that transmits frames free of undetected errors. For reliable services ARQ and acknowledgements are provided. Media Access Control (MAC) Sublayer: Controls access to shared channel for broadcast networks. Network Layer (3): Controls operation of subnet. Adressing, routing, congestion control, and quality of service (QoS) are design issues for the network layer.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.35/63

OSI Layers (contd.)

Transport Layer (4): Main function is to break up messages into packets for transmission and ensure that all pieces arrive and are assembled correctly at the

• ther end. Transport layer is true end-to-end layer that

can provide connection-oriented or connectionless service. Session Layer (5): Allows users on different machines to establish common session with services such as dialog control (whose turn to talk), token management (only one user can do critial operation at one time), and synchronization (e.g., to establish recovery point after crash).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.36/63

OSI Layers (contd.)

Presentation Layer (6): Lower layers move bits

• around. This layer is concerned with syntax and

semantics of transmitted information. Abstract data structures are defined in this layer and then converted to machine/platform specific representations. Application Layer (7): Typical application protocols are HTTP , FTP , SMTP .

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.37/63

TCP/IP Reference Model

When the ARPAnet started operating in 1969, an early version of NCP (network control protocol) was used to

• perate the network.

The TCP (Transmission Control Protocol) was outlined by Kahn and Cerf in 1974 and introduced in 1977 for cross-network connections. TCP was faster, easier to use, and less expensive than NCP . The IP (Internet Protocol) was added to TCP in 1978 to take over the routing of messages. By spring of 1983 all sites connected to the ARPAnet converted to TCP/IP . DoD wanted connections to remain intact as long as source and destination machines were functioning, even if some machines and links in between were suddenly blown to pieces.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.38/63

TCP/IP vs OSI Reference Model

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.39/63

Initial TCP/IP Model

TCP: Transmission Control Protocol. UDP: User Datagram Protocol. IP: Internet Protocol.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.40/63

TCP/IP Layers

Internet layer. Holds the whole architecture together. Job of IP protocol is to deliver packets where they are supposed to go, across possibly heterogeneous networks. Transport layer. Designed to allow peer entities at source and destination hosts to communicate. TCP protocol is reliable connection-oriented protocol. UDP is unreliable (i.e., no ACK), connectionless protocol for applications that do not need TCP .

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.41/63

TCP/IP Layers

TCP/IP model does not have session or presentation

• layers. No need was perceived, so they were not
• included. Experience with OSI model proved this view

correct. Application layer. Contains all higher-level protocols such as TELNET, FTP , SMTP , NNTP , HTTP , etc.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.42/63

OSI vs. TCP/IP

OSI and TCP/IP have much in common. Both are based on concept of stack of independent protocols. Three concepts are central to the OSI model:

Services. Interfaces. Protocols.

TCP/IP did not originally clearly distinguish between those three concepts. OSI reference model was devised before the corresponding protocols were invented. Unfortunately designers did not have much experience with networks. With TCP/IP the reverse was true: Protocols came first and the model was really just a description of the existing protocols.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.43/63

Critique of ISO Model

Neither OSI model and its protocols nor TCP/IP model and its protocols are perfect. Around 1990 it looked like the OSI model was going to take over the world and push everything else out of their way. This did not happen, mainly for the following reasons:

Bad timing. Bad technology. Bad implementation. Bad politics.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.44/63

Apocalypse of 2 Elephants

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.45/63

Critique of TCP/IP Model

TCP/IP model problems:

Does not clearly distinguish concepts of service, interface, and protocol. Not at all general and poorly suited to describe anything else than TCP/IP (e.g., Bluetooth). Host-to-network layer is more an interface to the network than a layer. TCP/IP does not distinguish physical and data link layers. Many protocols on top of TCP/IP are ad hoc (e.g., TELNET).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.46/63

Example Networks

The subject of computer networking covers many different kinds of networks, large, small, different goals, scales, and technologies. Examples:

Internet: Collection of different networks. Evolved from ARPANET and later NSFNET. Main glue: TCP/IP X.25, Frame Relay, ATM: Connection-oriented networks designed by telephone companies. Ethernet: Most widely used local are network standard. Origin: ALOHANET. Wireless LANs: Also known as WiFi. Most commonly used standard is IEEE 802.11.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.47/63

Architecture of Internet

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.48/63

Architecture of Internet

Client connects to ISP’s (Internet service provider) POP (point of presence) via dial-up telephone line. At POP signal is transferred from telephone system into ISP’s regional network. If data is destined for host served by same regional ISP it is delivered directly, otherwise it is handed over to ISP’s backbone operator. Major backbone operators like AT&T and Sprint

• perate large international networks with thousands of

routers interconnected by high-bandwidth optical fiber. NAPs (network access point) are interconnection points between competing backbones where packets can hop between backbones. Large corporations and hosting services run server farms, often directly connected to a backbone.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.49/63

X.25, Frame Relay, ATM

Two camps in networking world: People who support connectionless subnets and people who support connection-oriented subnets. Proponents of connectionless subnets come from ARPANET/Internet community. Main goal is fault-tolerance and dynamic network reconfiguration. Proponents of connection-oriented subnets are the telephone companies. Main goals are billing and quality of service. X.25, frame relay and ATM are connection-oriented wide area subnets that make use of the global network infrastructure set up by the telephone companies.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.50/63

X.25

X.25 (ITU-T standard) was the first public data network. It operates at level 3 of the OSI model and includes data-link and physical layer protocols (LAP-B and X.21). To use X.25, a computer first establishes a connection to a remote computer. The connection is then given a connection number (12 bits) to be used in data transfer packets. Data packets are simple, consisting of 3 byte header (connection number + other overhead) and up to 128 bytes of data. X.25 provides data transfer rates from 9.6 kbps to 256 kbps, depending on connection method.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.51/63

Frame Relay

In the 1980s X.25 networks were largely replaced by a network called frame relay. Frame relay is approved by ANSI and ITU-T and works at OSI level 2 (data link layer). It is connection-oriented (

• packets are delivered in correct order), but provides

no error control and no flow control. Data transfer rates for frame relay range from 56 kbps to 1.544 Mbps (T1 line).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.52/63

ATM

ATM (Asynchronous Transfer Mode) is also a connection-oriented network. It can be used for both LANs and WANs. ATM was designed in the early 1990s and launched amid truly incredible hype. It was going to solve all the world’s networking and telecommunications problems, merging data, voice, video, etc, seamlessly. ATM is extremely scalable with data rates ranging from 25 Mbps to 2.4 Gbps. The higher data rates are achieved using SONET (optical fiber transmission standard at physical layer).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.53/63

ATM Virtual Circuits

ATM uses virtual circuits that are set up during the connection phase. Permanent virtual circuits are also possible.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.54/63

ATM Cells

ATM is a cell relay technology, which means that it uses standard-sized packets called cells. Each cell is 53 bytes long and carries 48 bytes of payload. Part of the 5-byte header is the connection identifier so that all routers along the way know how to forward the cell. Cell delivery is not guaranteed, but the order is since all cells use the same path.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.55/63

ATM Reference Model

ATM has its own reference model, different from OSI and TCP/IP .

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.56/63

ATM Layers

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.57/63

SONET

SONET (Synchronous Optical NETwork) is a physical layer standard for high-speed transmission of digital data over optical fibers. It uses synchronous TDM (time-division multiplexing) and was developed by Bellcore and ITU-T in the 1980s for long-distance telephone traffic. SONET defines a number of OC (Optical Carrier) levels, each defining and optical signal and a corresponding electrical signal. Rates vary 51.8 Mbps (OC-1), to 155.5 Mbps (OC-3), to 622.0 Mbps (OC-12), and up to 2.48 Gbps (OC-48).

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.58/63

Standards

Standards are needed to ensure interoperability between equipment manufactured by different companies. Worldwide standards are needed for networks that span the globe like the telephone network and the Internet. Standards also increase the market for products adhering to the standard. The resulting larger market leads to bigger mass production and and decreased price. De facto standards: Standards that “happened” without any formal plan, e.g., IBM PC, UNIX. De jure standards: Formal, legal standards adopted by some authorized standardization body, e.g., IEEE 802.3, ITU-T V.90 modem standard

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.59/63

Standards Organizations

ITU-T (International Telecommunication Union, Telecommunications Standardization Sector). Predecessors date back to 1865 (telegraphy days). ISO (International Standards Organization), founded in 1946). Members

US: ANSI (American National Standards Institute) GB: BSI (British Standards Institute) Germany: DIN (Deutsche Industrie Norm) and many others

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.60/63

Standards Organizations (contd)

NIST (National Institute of Standards and Technology) is part of US Dept. of Commerce. EIA (Electronic Industries Alliance). National trade

• rganization that includes the full spectrum of U.S.

manufacturers. IEEE (Institute of Electrical and Electronics Engineers). IEEE is largest professional organization in the world.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.61/63

IEEE 802.11 Standards

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.62/63

Internet Standards

IAB (Internet Architecture Board). Subsidiaries:

IRTF (Internet Research Task Force), concentrates on long-term research IETF (Internet Engineering Task Force), deals with short-term engineering issues.

RFC (Request for Comment). Technical reports that deal with issues related to design, operation and maintenance of the Internet. Stored on-line at www.ietf.org/rfc. RFCs can advance to proposed standards, then to draft standards and finally to Internet standards.

Data Networks, Introduction, c

• 1996–2005, P

. Mathys – p.63/63