Interesting Links on EE&CS

Terms:
- Sidetone: a problem introduced by letting the transmission and reception circuits share the same pair of wires, so that the electrical signal from the voice-to-signal transducer flows directly to the reception circuit. As a result, whatever is spoken is heard loudly on the receiver. This is called sidetone.
- TDM: Time Division Multiplexing is used to combine 64-Kbps digital voice streams to form higher bandwidth streams.
- Switching System: A special-purpose computer used to interprets tones or pulses from the dialers to place a call or activate special features such as call forwarding. The switching system also sends a ring signal to the instrument when a call arrives. This rings a bell or activates a beeper.
Basics:
- Concepts: The telephone network offers a single basic service to its users: two-way, switched voice service with small end-to-end delays and a guarantee that a call, once accepted, will run to completion. It achieves its quality of service by setting up a circuit between the two endpoints. The network guarantees enough resourcess to each circuit to ensure that the service quality is met.
- Components: end-systems, transmission, switching, and signaling
Cellular Communications:
- Basics: Cellular Communication replaces the two wires between an end-system and the central office with radio transmission over a pair of frequencies. Because of the scarce resource of radio frequencies, they must be carefully allocated.
- Problems:
  location tracking; moving end-systems; bandwidth scarcity; higher prices

II. Internet:

Terms:
- Router: a special-purpose computer that transfers data to and from the next higher level of the network hierarchy. It is a host that has an interface on more than one network.
- Packets: The Internet carries all information using packets. A packet has two parts: payload, is the information content; meta-data or header, is information about the payload.
- Bridges: small switches allowing several Ethernets to talk to each other.
Basics:
- Concept: The Internet is the set of computers that are reachable with IP. It is a loose collection of networks organized into multilevel hierarchy using a wide variety of interconnection technologies.
- Components:
  - addressing: addressing provides a uniform way of identifying a destination in the network.
  - routing: routing allows information to traverse the network from end to end.
  - IP (Internet Protocol): IP allows data to be interpreted consistently as they travel across the network.
- Protocols:
  - ARP: Address Resolution Protocol. It helps a source system deliver data directly to a destination system when the two systems are on the same physical network. It allows the source system to determine the destination system's physical hardware address given the destination system's internet address.
  - ICMP: Internet Control Message Protocol. It employs the services of IP to allow systems to report on error conditions and to provide information about unexpected circumstances.
  - IP: Internet Protocol. It is the core protocol of the TCP/IP protocol suite. It provides a connectionless, best-efforts data delivery service used in moving packets from one system to another through the network.
  - RARP: Reverse Address Resolution Protocol. It allows a system that does not yet have its internet address to obtain it. PARP is typically used to support workstations and intelligent terminals that do not have their own disk storage.
  - RP: Routing Protocol. The basic routing function in a TCP/IP network is performed by IP. However, in networks that employ routers, the routers often run additonal routing protocols that allow them to exchange rouging information with each other.
  - TCP: Transmission Control Protocol. It is a connection-oriented data transport protocol that provides for reliable, sequenced, stream data delivery.
  - UDP: User Datagram Protocol. It is a simpler transport protocol compared to TCP. It is a best-effort, connectionless data transport protocol that adds little to the underlying IP datagram delivery service.
- Address Classes: We partition the address space into Class A, B, C, and D addresses with different numbers of bits assigned to the network and interface numbers in each class. The network distinguishes among the four classes of addresses depending on the first few bits of network number.
  - Class A: It has 8 bits for the network number and 24 bits for the interface number. The first bit of Class A address is 0.
  - Class B: It has 16 bits for the network number and 16 bits for the interface number. The first few bits of Class B address is 10.
  - Class C: It has 24 bits for the network number and 8 bits for the interface number. The first few bits of Class C address is 110.
  - Class D: It has 32 bits for the network number. The first few bits of Class D address is 1110.

III. ATM Networks:

Terms:
- Analog Telecommunications Circuits: They are often used for data transmission by employing modems (short for modulator/demodulator) on each end of the circuit. A number of protocols for the data link subnetwork have been developed to control transmission over long-distance telecommunications circuits.
- ANSI: American National Standards Institute
- ARCnet: ARCnet is the name of a family of LAN products that implement a relatively low-speed form of LAN data link technology in which all systems are attached to a common coaxial cable. Like the Token Bus form of LAN, a system transmits when it has the token.
- ATM: Asynchronous Transfer Mode
- Big-Endian: high-order bytes first
- Carrier Sensing: the process of checking to see if the medium is already in use before placing a frame
  on the link.
- Checksum: the function performed on each byte of the file the sender sends to check if it matches.
- Congestion Control: Congestion Control mechanisms come into play if the network, or some portion of it, becomes overloaded in spite of the flow control mechanisms. The principal technique used to control congestion in an ATM network is to discard cells.
- Connectionless Service: Also called datagram service . With a conncetionless data transmission service, no connection is established between the source system and the destination system. Each packet transmitted is sent from a source system to a destination system independently of all others without requiring that a connection first be established between them.
- Connection-Oriented Service: With a connection-oriented service, a logical association, called a connection, must be established between the source and destination systems before data can be exchanged between them. A connection in a computer network may be a permanent connection that is always available for transmission, like a conventional leased line. Alternatively, a connection can be a temporary connection that is established when there is data to be transmitted, usd for transmission and then terminated.
- Contention Resolution: With every switch design, there are places where contention for resources may occur. Buffers can be used to hold cells until the contention is resolved, but at times buffer capacity can be exceeded. Switch design includes the size and placement of buffers, as well as methods for avoiding or handling congestion.
- CSMA/CD: carrier-sense multiple access/collision detect
- Datagram: each packet containing the full destination address is called a datagram.
- Digital Telecommunications Circuits: Computer data can be carried directly in digital form in Digital transmission facilities.
- Data Link Connection Establishment and Release: Dynamically establishes, for a connection-mode service, a logical data link connection between two users of the data link service and releases the connection when it is no longer required. These functions are not provided for a conncetionless service, in which connections are not established or released.
- Data Transfer Transfers frames over a physical circuit, extracts the data unit from each frame and passes it up to the user of the Data Link layer service in the receiving device.
- Error Detection: Detects transmission errors, frame format errors, and procedural errors on the data link connection using a value carried in the frame control information.
- Error Recovery: Recovers from errors detected on data links using frame sequencing and acknowledgments (does not apply to a connectionless service).
- Ethernet: Ethernet is a LAN data technology in which systems are attached to a common transmission facility, such as a coaxial cable or twisted-pair cable, to form a bus- or tree-structured configuration. A system typically attempts to transmit whenever it has data to send. Ethernet is the most widely used form of LAN data link technology.
- FDDI: Fiber Distributed Data Interface. FDDI is a high-speed LAN data link technology in which systems are connected to one another using point-to-point fiber-optic cable segments to form a ring structure. A system is allowd to transmit only when it has the token.
- Flow Control: Controls the rate at which a user of a connection-mode Data Link layer service receives frames to prevent a user from being overloaded (does not apply to a connectionless service). The primary method used for flow control is input rate cotrol, in which a limit is placed on the rate at which each switch accepts cells arriving over input transmission paths.
- Frame Relay: It is a data transmission technology developed to provide a high-speed alternative to the packet-switching techniques used in conventional computer networks. By eliminating much of the overhead associated with error handling, Frame Relay networks offer data rates that are often higher than those offered by conventional computer networks.
- Frame Sequencing: Uses sequence numbers to ensure that frames are delivered in the same order in which they were transmitted (does not apply to a connectionless service).
- Frame Synchronization: Establishes and maintains synchronization between the sending device and the receiving device. This means the receiving device must be capable of determining where each frame begins and ends.
- Framing: It is the informal name most often used for the DLPDU (Data Link protocol-data-unit) exchanged between peer Data Link layer entities. Creates a single Data Link protocol-data-unit from the data unit passed from a user of the Data Link layer service, marks the beginning and the end of the DLPDU when sending, and determines the beginning and ending of the DLPDU when receiving. (It is insertion of bit patterns (markers) in the bit stream to tell "data bits" from "idle bits"
- HDLC: high-level data link control
- IP: Internet Protocol
  It presents an end-system with the abstraction of an end-to-end logical link by providing routing, packet-forwarding, and segmentation and reassembly.
- Integrated Service: To carry telephone voice, digital video, and computer data, using a universal networking technology, and allow users to access one or more media simultaneously.
- ISDN: Integrated Services Digital Network. ISDN is a public telecommunications network --- typically administered by a telecommunications provider --- that supplies digital end-to-end data transmission services that can be used for any purpose, including voice, data, graphics, image, and facsimile.
- ISO: International Organization for Standardization
- Isochronous Service: It is uniform with respect to time and in which the information being transmitted is guaranteed to arrive at regular intervals. Isochronous service is required for voice and video transmissions, where delays or interruptions in the transmission flow would be noticeable to the receiver.
- ITU: International Telecommunication Union
- Jitter: In a typical ATM network, the transit delay for cells transmitted over a particular connection may not always be exactly the same. The variation in transit delay is called jitter.
- LAN: local Area Network
- Latency: Also called transit delay . It is the time required for each individual cell to traverse the network from a source endpoint to its destination endpoint.
- Little-Endian: high-order bytes last
- LocalTalk: LocalTalk is a low-speed LAN data link technology --- part of Apple computer's AppleTalk networking scheme --- in which systems are attached to a common cable. LocalTalk technology has been built into most of the computing devices that Apple Computer has manufactured for many years, although Apple has been building Ethernet technology as well into much of its more recent equipment.
- MAC: Medium Access Control. a sublayer of datalink layer. It supports point-to-point communication with 6-byte globally unique datalink addresses to identify end-systems.
- Named Pipes: The Named Pipes facility is a system for interprocess communication codeveloped by IBM and Microsoft for their respective network operating system products. The Named Pipes mechanism has become a de facto standard that has been implemented in other network operating systems, including Novell NetWare software. The Named Pipes application programming interface allows a program to send data to or receive data from another program in a way similar to issuing ordinary file management commands.
- Open Protocol: a protocol whose details are publicly available.
- Open System: a system that implements open protocol standards.
- OSI: Open Systems Interconnection
- Peer Entities: the two (or more) parties at the same level communicating with each other.
- Physical Layer Services: Uses the services of the Physical layer to transmit and receive data and to control the operation of the physical communication link.
- Pixel: For the purposes of carrying video information in digital form, each frame may be considered to consist of some number of fixed points, often called pixels. Each pixel has certain intensity and color values associated with it.
- Protocol: a set of rules and formats that govern the communication between communicating peers.
- Protocol-Data-Units: Data units sent from a layer entity in one system to a peer layer entity in another system are called protocol-data-units.
- PDU: Protocol Data Units. the packets exchanged between peer entities in the same layer.
- RPC: A Remote Procedure Call (RPC) facility can make it possible for programmers to implement a client/server application without needing to explicitly issue requests for communication services. The idea behind a remote procedure call facility is that procedure calls are a well-understood mechanism for transferring control and data from one prodecure to another in a computing application. It is of great utility to extend the procedure call mechanism from a set of procedures in a single-computer environment to a set of procedures in a distributed, client/server environment.
- SAP: Service Access Point. the interface between the lower and the upper layer.
- Scalability: Some switch designs work well with lower numbers of transmission paths but may not retain acceptable levels of performance or cost when the number of transmission paths becomes large.
- SDU: Service Data Units. the packets handed to a layer by an upper layer.
- Skew: Any difference in time of presentation between the related audio and video informatin is called skew. In order for voice information to appear synchronized with lip movements, skew values of up to 120 milliseconds are not ordinarily apparant to the viewer.
- Sockets: The Sockets application programming interface was initially developed for use with TCP/IP networking software in the BSD UNIX operating system environment. By using the Socket API, two application programs, one running in the local system and another running in the remote system, can communicate with one another in a standardized manner. The Sockets API is typically used to implement a client/server relationship between two application programs running in different computing systems. The client and server programs each invoke functions that set up an association between them. The client and server applications then invoke functions to send and receive information over the network in a similar manner to the Named Pipes API.
- SONET: synchronous optical network. SONET data rates begin at 51.84 Mbps and theoretically could reach 13,219.2 Mbps.<
- Statistical Multiplexing Gain: peak serving rate over output line speed
- Statistical Multiplexing Gain: peak serving rate over output line speed
- STM: Synchronous Transfer Mode
- TCP: transmission control protocol. It provides error detection and correction, as well as multiplexing and flow control.
- Throughput: Throughput measures the rate at which data passes through the switch. It reflects the switch's internal processing and transmission rates, as well as the degree of parallelism within the switch. ATM switches may be designed to process cells at state-of-the-art speed, at the time of writing in the neighborhood of 2 Gbps. It is interesting to note that at a data rate of 2 Gbps, the switch needs to be capable of receiving and transmitting cells at the rate of 4 million cells per second over each active transmission path.
- Token Bus: Token Bus is a LAN data link technology in which systems are connected to a common transmission medium in a similar manner as an Ethernet LAN. A system is allowed to transmit only when it has a special data unit, called the token, passed from one system to another. Token Bus LANs are sometimes used in factory automation environments.
- Token Ring: Token Ring is a LAN data link technology in which systems are connected to one another using point-to-point twisted-pair cable segments to form a ring structure. A system is allowed to transmit only when it has the token, which is passed from one system to another around the ring.
- UDP: user datagram protocol. It provides multiplexing, but not error recovery or flow control.
- VCI: Virtual Circuit Identifier
- Virtual Circuit: A virtual circuit is established between each pair of users that need to communicate in a computer network. A virtual circuit makes it appear to two users as if they have an actual point-to-point link between them, just as if they were using a circuit in a telephone network for communication. Hardware and software in each user machine and in the devices that make up the network cooperate to implement the required virtual circuits.
- Virtual Circuit Switching: It is one of two ways to build a packet-switched network. Packet headers carry identifiers instead of addresses, and each switch maintains a translation from the identifier to a destination. This saves header space, but the mapping from an identifier to a destination must be set up at each switch along the path before data transmissin begins.
- Windows Sockets: It's a variation of the Sockets application programming interface now being widely used in networks of systems using Microsoft operating systems software, such as Windows. The Windows Sockets, or WinSock, API has been developed in an effort to provide a standardized API for netowrk communication in the Microsoft operating system environment. The WinSock API is based on the BSD UNIX Sockets API and supports both a connectionless data transport service and a connection-oriented data transport service.
- Wireless LAN Technology: Local area network infrastructure vendors have developed various technologies for implementing LAN communication over wireless transmission media, such as radio and infrared signals.
- X.25 Packet-Switched Data Network (PSDN): X.25 is the name of an ITU-T Recommendation that describes standards for packet-switched data networks(PSDN). A PSDN carries data in data units called packets that can be routed individually through the PSDN. An X.25 PSDN implements virtual circuits that appear to end users as simple point-to-point data links.
Basics:
- Basic Concepts: virtual circuits; cells (fixed-size packets); small packet size; statistical multiplexing; integrated services
- Advantages OF ATM: ATM allows us to build networks that can carry multiple classses of traffic with quality-of-service guarantees provided to individual streams. They also enable large, parallel switches and provide a uniform framework of network management.
- Services Offered: ATM provides different types of transmission services, including connection-oriented, connectionless data transfer services, and isochronous data delivery services.
  There are four classes of services provided by ATM:
  - Class A. Class A service is connection-oriented and isochronous. Class A service maintains a constant bit rate and timing relationship between the source and the destination. Class A service can be used to replace a conventional circuit-switched telecommunications link for voice or video transmission.
  - Class B. Class B service is connection-oriented, has a variable bit rate, and maintains a timing relationship between the source and the destination. Class B service is intended for compressed audio and video transmission, such as that used in teleconferencing, where a variable bit rate can be tolerated as long as delays are within guaranteed boundaries.
  - Class C. Class C service is connection-oriented with a variable bit rate and no timing relationship. This service is intended to provide data transfer similar to that provided by a virtual circuit in a conventional computer network.
  - Class D. Class D service is connectionless with a variable bit rate and no timing relationship. This service is intended to provide the connectionless data transfer service commonly used with local area network.
- ATM Components:
  - ATM switches: ATM switches perform functions related to the routing of information. An ATM switch is sometimes called an intermediate system(IS).
  - ATM endpoints: Also called end system. A device playing the role of an ATM endpoint services as the source or destination of user data. An ATM endpoint might be implemented in an ordinary computing system by installing an ATM network interface card (NIC) in the computer and running appropriate communication software.
    An ATM endpoint might also be implemented in a special-purpose network device to which one or more ordinary computing systems can be attached via conventional LAN NICS.
  - transmission paths: Transmission Paths are physical communication links interconnecting ATM switches and ATM endpoints. They can be implemented using different types of physical communication circuits using fiber optic or electrical transmission media.
- ATM Cell Format:
  - Virtual Path Identifier (VPI): The VPI is 12 bits in length and is used to group virtual channels into paths for routing purposes.
  - Virtual Channel Identifier (VCI): The VCI is a 16-bit field that identifies a particular virtual channel within a virtual path.
  - Payload Type (PT): The PT field is a 2-bit field that identifies the type of information contained in the Payload field.
  - Cell Loss Priority (CLP): The CLP field is a 1-bit field that ATM equipment uses to determine which cells to begin discarding first when congestion occurs.
  - Header Error Check (HEC): The HEC field contains an error detection and correction code value used to detect and sometimes correct errors in the five header octets of the cell.
  - Payload: The cell's payload contains 48 octets (384 bits) of user data and/or additional control information.
- Importance of Layering:
  1. Layering allows us to decompose complicated tasks into simpler ones that we can solve independently and in parallel.
  2. Implementation can be seperated from specification by layering, so that implementations of lower layers can be replaced at will without risking the investment made at higher layers in the protocol stack.
  3. Functionality can be reused, that is, functionality of a lower layer can be shared by many upper layers.
- The Seven Layers: physical layer, datalink layer, network layer, transport layer, session layer, presentation layer, and application layer.
- Characteristics of Seven Layers:
  1.Physical Layer: It is responsible for moving information between two systems connected by a single physical link; it is independent of link technology. Physical-layer protocols describe such things as the coding scheme used to represent a bit on a communication link, connector shape and size, and bit-level synchronization.
  2.Datalink Layer: It provides framing and point-to-point communication. Some datalink layers have the ability to retransmit packets corrupted on the link (error control), and to pace the rate at which packets are placed on a link (flow control). The datalink is dependent on the nature of the physical medium. In most systems, a single host-adaptor card provides both the datalink layer and the physical layer. A Ethernet card provides not only the transmitter and receiver for placing and receiving bits on a coaxial cable, but also the ability to frame packets and the medium access control to arbitrate acccess to the medium.
  3.Network Layer: Its function is to logically concatenate a set of links to form the abstraction of an end-to-end link. It hides idiosyncracies of the datalink layer from higher layers.
  At an end-system, it may also provide some error detection.
  At intermediate systems, and at end-systems connected to multiple routers, it participates in a routing protocol to discover the next hop for every possible destination in the network.
  In a datagram network, it is responsible for forwarding packets, scheduling their transmission order, and, if necessary, dropping rate guarantees.
  4. Transport Layer: It creates the abstraction of an error-controlled, and flow-controlled end-to-end link. Transport-layer protocols deal with loss by retransmitting lost packets; with corruption by detecting, discarding, and retransmitting corrupted packets; and with duplication by detecting and discarding duplicate packets.
  5. Session Layer: It adds the abstraction of full-duplex service, expedited data delivery, and session synchronization.
  6. Presentation Layer: It deals with data, instead of meta-data in headers. It converts the representations to a network standard so that the difference between applications is hidden from the applications. It may also encrypt data, both to authenticate it to the receiving application and to prevent unauthorized parties from accessing it.
  7. Application Layer: It is the set of applications running on an end-system. Any program that uses the presentation layer's services can be considered a part of the application layer.

IV. Others

Frame Relay Technology: It's a high-speed alternative to the packet-switching techniques used in conventional computer networks. In a Frame Relay network, messages are routed from one Frame Relay node to the next based on an identifier associated with the virtual circuit being used.
major difference between Frame Relay network and a conventional computer network: error correction.
In a conventional computer network: each time a packet is sent acorss a circuit between routers in the network, the router makes a check to ensure that the packets are not corrupted during transmission. The reciving router sends the sending router an acknowledgement for each correctly received packet. If a sending router does not receive an acknowledgment, it retransmits the packet in question.
In a Frame Relay network: it is assumed that a transmission medium is being used that supports a very low error rate, and acknowledgments are not used as packets move through the network. It is assumed that error detection and retransmission will be handled by sofeware running in the user machines that are communicating.
Reference:
1. An Engineering Approach to Computer Networking
by S. Keshav (AT&T Labs-Research)
ISBN: 0-201-63442-2
2. Asynchronous Transfer Mode --- ATM Architecture and Implementation
by James Martin, Kathleen Kavanagh Chapman, Joe Leben
ISBN: 0-13-567918-4
Back to Home

URL: http://geocities.datacellar.net/redoakland/homepage/research.html