Presentation on theme: "Phones OFF Please Network Introduction Brian Bramer Home:"— Presentation transcript:
Phones OFF Please Network Introduction Brian Bramer Home: www.cse.dmu.ac.uk/~bb Email: firstname.lastname@example.org
Topics: 1.Network 2. Transmission Types 3. Development of Networks 4. Message Transmission Basic 5. Network Types 6. Communication Problems 7. OSI Model 8. Network Standards
1.What is A Network? A Network links two or more nodes (Subnets) together to communicate and share resources. Sharing resources includes both data sharing and hardware. Example?
2. Reason for Networks: stand alone computers are fine for individual work but users need to share resources and access remote information, e.g. the internet (or world wide web) reason for networks to share information, e.g. files, databases, etc. to share resources, e.g. expensive printers, backup systems, etc
3. Concept of Subnet Nodes connected with series of links called subnet. Subnet is a network that is part of another network. Connection between subnets is made through Routers, Bridges or Gateway. Example of Subnet?
4. Node: computers, terminals, peripherals or communications devices (such ad routers, switches etc) Nodes communicate by transmitting data, requesting/delivering services etc. across the subnet. Transmissions must arrive at right destination and be interpreted correctly by the receiver. (concept of addressing) For reliable communications (e.g. file transfer) errors (e.g. due to electrical noise on the communications lines, etc.) must be corrected for. other essential components are; protocols, NIC, Connectors and cables.
5. Network Topologies Communication can be: Simplex Transmission only possible in one direction. Half-Duplex Transmission possible in both directions but not at the same time. Full-Duplex Transmission possible in both directions simultaneously. Topologies: Direct Point to Point Link.
Mesh topology: no single path exists between source and destination the message is routed through the network from node to node until it reaches the destination at each node the message is switched onto a line as determined by the routing algorithm, e.g. to give the shortest path from source to destination e.g. WANs (Wide Area Networks)
Bus Topology: Signal transmitted by any station propagates in both directions. until being absorbed by terminating resistors at each end. Need of Terminating Registers at each end?
Ring Topology A ring network consists of a single loop of cable. usually coaxial or twisted telephone pair. traffic flows in one direction from node to node. Signal transmitted by a node is repeated by the intermediate node until it returns to the transmitter (which then removes it).
Star Topology A star network consists of one central device (Hub or Switch). Better Performance, isolation of devices and simplicity. highly dependent on central node (Failure ??)
Advantages and Disadvantages of each topology? Physical and Logical topologies?
6. Transmission types: Broadcast; a sender transmits a message which is received by all notes Bus and ring topologies are broadcast networks: on a bus the signal propagates both ways along the bus being received by each node in turn on a ring the signal is transmitted from node to node around the ring Multicast; is communication between a single sender and multiple receivers. Unicast; is communication between a single sender and a single receiver over a network. Anycast; is communication between any sender and the nearest of a group of receivers in a network Role of Addressing in Transmission?
7. Development of Networks: Master-Slave Model A central computer system which was accessed by dumb terminals either locally or from remote sites via telephone lines. Dumb terminal has no processing power. Processing was carried out by the central computer. As users hit the keyboard the characters were transmitted to the central computer, processing carried out and results sent back to the display screen. Master-slave network as the central computer was in complete control.
Distributed networks and the peer-to-peer model Distribute processing power throughout the network. Dedicate it to individual users or particular tasks. It is also often cheaper in terms of support cost. Independent processors exchange information with each other, to share information and resources. Called a peer-to-peer network and the connection of these independent peer machines is at the heart of distributed networks both local and wide area. Station AStation B send message rec message rec message send message send message rec message rec message send message
Stations must agree on when to send and receive information, And the format of the messages i.e. must agreed on same communications protocol (rules of communication) e.g. otherwise it is possible for deadlock to occur if station A is waiting for a message from station B and station B is waiting for a message from station A. Advantages improved reliability and fault tolerance, sharing of data, sharing of special hardware resources. Disadvantages Vulnerable to network and/or server failure, Cost of interconnection, Cost of networking/communications software, Problems of access to shared file systems, Difficult to manage, etc.
The client-server model: System resource is managed by a server process running on a computer somewhere in the network. A 'client' process on any machine in the network operates on a resource by sending a request to the corresponding server and receiving a response. Client/Server exchange of information Client Server loop send message wait for message rec message send message end loop
server runs continuously in a loop waiting for client requests (threads) when a request arrives a thread is created which processes the request and then terminates client sends request, waits for response and when it arrives carries on. processing workload is shared between the client and server machines client-server mechanism is generally invisible to the user server machines are usually; powerful running a multi-tasking operating system or network software. can handle multiple requests at a time
8. Basics of message transmission Circuit Switching A route is set up through a number of exchanges to the destination. the communication channels used are dedicated for the duration of the call channels are unusable by others until the call is terminated tariff is based on duration of call (not amount of data transferred) Example: the old (non digital) telephone system, i.e. when A and D have set up a call B and C cannot use line.
Packet Switching used with digital connections to allow multiple calls to exist on same circuit. Data broken into small blocks ("packets") Packet includes extra information in Header (serial number, destination address...etc). Each packet is routed to its destination individually Packets re-assembled into original message when destination reached. Data tend to be in bursts and charging can be based on the amount of data transferred not duration of call.
ADVANTAGES OF PACKET SWITCHING Communication channels can support many calls simultaneously (interleaved). Short messages not delayed by long messages. More efficient than circuit switching. DISADVANTAGES Performance drops when many users share same network Packet header overhead. Today the majority of communication systems use packet switching.
9. Types of Network 9.1 Local Area Network (LAN) A general purpose network serving a variety of devices which is local to a room, building or site and is usually owned by the organisation concerned. An organisation may have a number of LANs ; mainly used for resource and local information sharing Well defined topology (bus, ring, star) Usually owned and managed by the organisation Does not normally use public communications facilities Usually quite fast The majority of LANs only carry data but technology moving towards also supporting speech, video, etc.
9.2 Wide Area Network (WAN) Extends beyond one site - covers a large area (national and international distances) Usually use facilities provided by the public telecommunications system or equivalent private systems No simple topology - usually pictured as a sparsely connected mesh Can be quite slow (depends upon cost) can carry any type of information (data, video, etc.)
9.3 Metropolitan Area Network (MAN) Has similar characteristics to a LAN but covers larger distances and operates are higher data rates Between LAN and WAN can be private or public Covers metropolitan area (town/city) typically ring or bus topology typically faster than LANs some MANs can carry any type of information
9.4 Example LAN network configuration (Ethernet backbone with bridges to subnetworks)
10. The Communications Problem Nodes must be agree on same protocol (rules of communication). Communication must be reliable and error-free, across a potentially unreliable subnet, i.e. data signals travelling along a cable being corrupted by electro/magnetic noise induced in cable by external equipment or radio signals
Achieving this is a complex problem, e.g. 1.The simplest subnet – single point-to-point line. What issues need to be considered to ensure reliable communication?
2. Any additional issues for a multipoint network e.g. a bus topology? addressing – who is the message for? Which station can use the network at any instant? (collision)
3. And for a switched system, e.g. a mesh topology? i.e. routing – how is a packet routed across the network
11.1 How can we achieve reliable communication? Use Communication Protocols (well defined rules) Systems adopted by both sides for the purpose of managing the communication The protocols will define sets of rules dictating how each side should behave in any given situation. The interchange of information must be in an agreed format (an agreed ‘language’) usually implemented by sending additional management information along with the message data The extra information imposes a Protocol Overhead in that it makes the messages larger.
11.2 What actually is a communications protocol? Protocol: a set of rules defining how computers are allowed to talk to each other. What sort of cable is needed? What plug is required at each end? How will we differentiate between 0 & 1(voltage, frequency, ?) How will receiver know that transmission has started? how will the receiving device interpret the bit stream 01011011? how can receiver be sure that 01011011 was what the sender transmitted? How errors are detected and what remedial action is taken; How the dialogue between sender and receiver is managed; How the receiver knows that the transmission has finished?
In addition, if there are more than 2 devices connected together How can we arrange for the message to be sent to the correct destination? How should the receiving device be addressed? and for a broadcast network such as a bus What if several devices try to all talk at once! Once physical communication is established and devices can exchange bits Syntax – can they understand each other’s ‘language’? i.e. character codes, number representation, etc. Semantics – can they understand the content of the message? The key to reliable communication is adherence to protocol STANDARDS
11.3 How are the protocols implemented? 1. By using additional lines. i.e. as well as data lines, there could be extra lines (control lines) to handle the protocols – very expensive! Example?
2. By adding extra bits to the message bits. Protocol BitsMessage Bits Transmitting software attaches the protocol bits to the message Receiving software interprets the protocol bits and acts accordingly. N.B. Frequently, more than one protocol is implemented. What must be done to ensure that the protocol bits are interpreted correctly?
11.4 Standards Transmitters and Receivers must be operating the same protocol Protocols must be implemented and interpreted in a strictly defined order Protocols must be standardised to ensure compatibility between different hardware, software, manufacturers, countries etc. Aim is to produce Open Systems International Standards Organisation (ISO) is attempting to produce Open Systems Interconnection (OSI) by suggesting a standard architecture based on a number of layers ISO - OSI Model
12. ISO – OSI 7-layer model The overall communication problem is split up into 7 levels or layers. Each layer deals with a particular part of the overall communications problem. Each layer provides a solution by implementing a particular protocol. Each layer uses a separate protocol. Standard protocols are developed for each layer. The protocol at each layer only communicates with a similar protocol at the other end of the link. i.e. peer-to-peer communication The result is a 7-layer stack of standard protocols (called a protocol stack)
a pair of stations communicating via intermediate nodes Actual communication takes place only at the physical layer
The diagram below shows user data passed down the layers until it is transmitted as bits, i.e. each layer adds its own header containing information for its protocol
12 Layering and protocol stacks Layering separates the application processing from the communications function. Why? Divides the overall communications problem into a number of distinct sub-categories Each category is represented by a layer in the model can develop standard protocols for each layer can prevent change/updates in one layer impacting upon other parts of the system facilitates simple peer-to-peer communication.
Any disadvantages of layering? may be cumbersome leading to extra processing delay new functionality - where should it go?
13. Protocol Stacks or Protocol Suites: Many different protocol systems have been implemented over the past thirty years and most adopt a layering approach. The set of layers is usually called a Protocol Stack or Protocol Suite, e.g. below shows the OSI, TCP/IP and IBM’s SMA protocol stacks. In practice a computer many run several protocol stacks concurrently with the appropriate messages being routed to correct stacks.
There are disadvantages: A standard takes a long time to emerge and tends to freeze the technology There are multiple conflicting standards for the same thing. ISO specified the 7 layer OSI Reference Model which provides a framework for networks. Some standards implemented while the OSI model was being developed do not conform exactly, e.g. TCP/IP In practice there is wide agreement on layers 1-3; these are usually implemented in hardware or a combination of hardware and software. The higher layers are much less well defined; there is also some dispute as to whether the OSI approach is the best way of supporting the functions required.