This module provides an overview of the concepts behind data telecommunications networks.
Data Networks can be classified into three major categories: Local area networks (LANs), access networks, and backbone networks. Although part of the difference is in the speeds that data can be transferred, the main attribute that determines how a network is classified is its position in the interconnection between end users. The details explaining the above elements are contained in other modules.
We will first look at Local Area Networks as the closest link to the end user. Before we do this, however, we need to understand how users on a network are connected to each other. This is called the network topology.
Local Area Networks are constructed using one of three topologies: bus, star, or ring. Within any given LAN, there are physical, electrical, and logical topologies. Physical refers to the way cables are physically routed between devices on the LAN. Electrical is the way the wires are connected. Logical is the way data units are routed from one device to another. Most of the time, electrical and logical topologies are the same. Many times, however, the physical and electrical topologies of a LAN will differ.
Bus topologies are common in the data communications networks. Ethernet is an example of a bus topology, which was originally developed as a proprietary technology by Xerox Corporation. The IEEE standardized most of the original Ethernet technology in its 802.3 standard for Carrier Sense Multiple Access/Collision Detection. Cable modems use the physical part and some of the higher layers of the Ethernet protocol.
Rings are another common LAN topology. IBM introduced the Token Ring as its technology for Local Area Networks. The corresponding IEEE standard is 802.5.
This chart shows how local area network protocols correspond to layers of the OSI model. Notice that the layer above the physical layer is called the Medium Access Control, or MAC, sublayer. One of the most important functions of the MAC layer is to regulate when a device may communicate on the network. This may be done either centrally in the network by a single controller, or on a decentralized basis, by stations on the network dynamically determining the order in which transmissions may occur. While centralized control is easier to implement, it creates a single point of failure, which can bring the entire network down, and may become a bottleneck during high network traffic periods. For these reasons, the majority of MAC protocols specify a decentralized operation. How a device is given permission to communicate is also a function of the MAC layer. The technique used in LANs is called asynchronous access control. This type of control allows the network to allocate its physical capacity dynamically, in response to immediate demand by the attached devices.
Access networks are intermediate networks in an end to end connection. They provide a path from a local area network to a high speed backbone network. Typically, these networks are entered at layer 3 of the OSI protocol, via a router on a LAN. Two common technologies for access networks are: X.25 and Frame Relay
The X.25 protocol specifies data flow in the first three layers of the OSI model. As data moves from switch to switch through an X.25 network, each stage must process all three lower layers of the OSI protocol. While this is very reliable information transfer, it also is a slow way to move data.
As networks became more reliable at the physical layer, protocol designers modified X.25, creating Frame Relay. Frame Relay processes layers 1 to 3 only at endpoints of the network. All other data movement through the network is at layers 1 and 2. The error correction that was done at layer 3 for in the intermediate points of the network for X.25 is now the responsibility of the endpoints and the lower layers of the protocol. Removing the need to go up one layer increases the speed at which data can move through the network.
Backbone networks have the highest bandwidth or speed of data transfer. They may be used to interconnect LANs into Wide Area Networks, or WANs, or simply to provide a one-time high speed connection between lower speed networks. Backbone technology is constantly evolving. At one time, X.25 was considered a backbone technology, but it has been replaced by the higher speed ATM transported on SONET, or a related cell technology, SMDS. These are discussed in greater detail in other modules.
One of the reasons for the high speed of ATM is the small size of the ATM protocol data unit, which is called a cell. In an ATM cell, information is contained in 48 bytes. A 5 byte header contains the address and cell priority information used in routing and delivery.
The Python Snake skin is a good metaphor for the flexibility needed from data networks. ATM can support applications of different bandwidths. Everything going into the network is encoded onto the 53 byte cell structure, and then multiplexed onto the network. What this means is that ATM equipment can be gracefully added to a network. An ATM multiplexer can be used to bring a combination of high speed data and lower speed voice or Ethernet LAN traffic onto a fiber operating at SONET OC-x rates. The information part of the cell can contain frame relay headers/trailers, T1, x.25, etc. So several technologies can be carried over ATM in the network, and sorted out at convenient nodes by ATM Multiplexers, Hubs, and Routers. The ATM Module provides detailed information.
Rather than allocating network times to equipment ports even when no data is being sent, ATM uses network capacity only when a device has information to send. The result is “Bandwidth on Demand”, which means that each device on an ATM network can have the share of the ATM transport mechanism that it needs.
Data Network concepts and equipment in telecommunications networks have been merged with the world of Computer Data Networking.The design and operation of data Networks can be understood in the context of the OSI Reference Model. Layering protocols provides structure to data communications systems. What may not be so apparent, however, is that there can be several ways to layer the protocols. Without a standard, systems created from multiple vendor sources would still find it difficult to communicate. In 1977, the International Organization for Standardization (ISO) recognized this problem, and established a committee to develop such a standard. The result of that committee’s work is now known as the Seven Layer Open Systems Interconnection (OSI) reference model. This model is generally regarded as the framework for any layered protocol architecture. It is important to note that the OSI model is not a particular architecture. It is a reference model. This means that no one layered architecture will have all the elements of the model. Most layered architectures, however, can be directly related to the OSI model, as can individual standard specifications for types of networks.
It is useful to notice some general characteristics before looking at each of the layers. First, the lowest level deals with hardware. This is where system designers spent most of their time prior to the advent of computer systems. Second, the seven layers can be grouped into two broad categories. Layers 1 through 3 specify the way that a computer will interact with a network. These layers also govern data communications between chains of computers or networks, and are implemented within those individual computers or networks. They are concerned with routing the individual protocol data units, rather than the total message. Layers 4 through 7 look beyond the individual computers or networks in a chain, and specify the rules for end to end communications. These layers are implemented within host computers. These are the layers that ensure the complete message is transmitted
The OSI Seven Layer Model can be understood in the context of the above 7 questions.
Layer 1 is the hardware layer. It is where the physical and electrical characteristics of the transmission path are defined. The Protocol Data Unit is the bit. This layer allows basic data to pass over an electrical connection. Layer 2 groups bits into data transmission units (Frames). Error checks are made on both the sending and receiving sides of a transmission, but not end to end. Layer 2 looks at intermediate points in the network . Signals for error checks are defined. The protocol asks for retransmissions between two points in the network if errors occur, but doesn’t monitor end to end. Layer 3 defines codes within the frames. The Protocol Data Unit is commonly known as the packet . This layer determines routing sequences and timing across a network, preventing misroute. Layer 4 manages end to end flows and error controls . It ensures end to end delivery of data across a network. Layer 5 specifies when it is time for either side to talk. It negotiates and reaches agreement on the nature and duration of communications. It also checks for security violations or terminal incompatibilities. Layer 6 translates codes (e.g.: ASCII to EBCDIC) when necessary. It makes sure data is intelligible from one side to another. Layer 7 interfaces to end user applications.
The Physical Layer deals with the mechanical, electrical, configuration and functional characteristics. It defines how 1s and 0s are to changed into signals (encoding) and sent either electrically or optically over the transmission medium.
Layer 2, the Data Link Layer, pertains to line protocols . These are the rules that govern the flow of data between two, and only two, entities. The objective of this layer is to provide a high throughput, fast response time, and minimal logic to account for data integrity. Functions performed at Layer 2 are: Flow control Synchronization, accomplished by software timers set by the protocol designer. Synchronization and flow control work together to regulate the rate of data transmission Error detection of data sent from another immediately connected entity Identification of the entities involved in the communication Identification of whether data or control information is being sent.
Layer 3 is the Network Layer. This is the layer that regulates how groups (networks) of individual entities communicate with each other. The following key issues are resolved by Layer 3 protocols: Different sizes and formats for Protocol Data Units in each network Different timers and timeouts Different quality of service Different addressing schemes Different levels of performance Different routing methods Different user interfaces (Connectionless versus Connection-oriented) Different levels of security Different troubleshooting, diagnostics, and network maintenance issues Cost and billing allocations
The Transport Layer defines the rules for information exchange and manages end-to-end delivery of information within and between networks. Includes error recovery and flow control.
Messages in a telecommunications Common Channel Signaling System or an Integrated Services Digital Network (ISDN) are a type of PDU that begins and ends with a flag , and contains information that spans multiple OSI layers. We will give an example in the applications section of this module as Common Channel Signaling. The key idea is that routing information is contained in a header field and customer information is in a separate field.
The general pattern of routing is common to all packet switched networks and is store and forward oriented, using a HEADER FIELD, USER DATA FIELD, and frequently a TRAILER FIELD.
The above graphic shows both a packet as defined by Layer 3 of the OSI model, and a message unit which may be sent the same as a packet, in a packet switched network. Notice that the message unit contains information from multiple layers of the OSI model. This is because certain types of packet switched networks are built upon protocols that were defined before the OSI model. In many ways, the message illustrated conforms to a hierarchical, rather than layered, protocol architecture, because it includes information from lower, as well as higher, OSI-defined layers. However, it may be possible that some of the layers are not hierarchical, in that they cannot directly access information from other layers without going through a lower level. In this sense, the protocol is itself a type of hybrid, containing elements of both layered and hierarchical architectures. In order to reconcile both views, the layered equivalent has been overlaid onto the hierarchical message. The key point, however, is that the network is still processing the message as a unit of information, using a store and forward methodology.
While discussing the rules that govern the communication of information, we need to be aware of how that information is structured as it moves through various systems and networks. This has been standardized as a basic block that is transferred between layers and across networks. This block is called the Protocol Data Unit, or PDU . A PDU has a data part , a header , and sometimes a trailer . The data part is contains information from a higher layer that is being operated upon by the current layer of the protocol standard. The header is information that provides the layer with direction concerning what to do with the information in the header. While the header is not a set of detailed instructions, quite often software will use header contents to structure those instructions. The trailer is supplemental information to that contained in the header. Trailers are typically only found in the second layer of the OSI protocol reference model.
Routers simply read destination information and forward the packet to the next node. Examine the above graphic carefully and routing will no longer seem mysterious. A packet enters the network at Node A with customer data. Its’ destination is Node C. There is no direct path from Node A to Node C. The packet is routed by Node A to Node B based on an internal routing table containing Node B’s connections stored at Node A. Note the multiple addresses in the routing PDU, C for final destination and B for the next node in the path to Node C. Upon arrival at Node B the packet is stored, analyzed and re-addressed to its final destination at Node C. Node B uses its internal routing table of direct connections to determine path and address.
We can now understand packet routing tables in the context of Ethernet Switching.
Each frame received by a bridge is stored, checked for errors and then re-sent as follows: If the Destination address is broadcast (FF:FF:FF:FF:FF:FF), the frame is sent out all ports except the one it arrived on. If the Source and Destination are both reachable from the same port of the bridge, the frame is dropped If the Source and Destination are reachable from different ports of the bridge, then the frame is re-sent out the destination port Each bridge keeps an internal Forwarding Table that associates addresses with ports. Learning : For each arriving data frame, switch examines source address and adds/updates entry in Forwarding Table containing Source Address (6-byte format) Port that this frame arrived on Current Time
Packet networks store and forward blocks of information. The current view of packet switching includes all the layers of the OSI model, so PDUs, rather than packets, are being processed by nodes. The PDU may simply contain a header and an information part, or it may be a message that spans multiple layers of the OSI model. The header (in the packet PDU) or some part of the message (in the message PDU) contains routing information. In the simplest case, this is the address of the intended receiver. Somewhere in the network, this routing information must be stored and processed. The table is a summary of the ways this can occur.
Several common data network elements will be roughly mapped into their OSI Reference Level. Terms and equipment definitions can be moving targets in the data arena, but the reference model is fixed.
Repeaters are Layer One devices. A better term for repeater is actually regenerator. Regeneration is a digital signal processing technique. Repeaters use analog amplification.
Bridges are Layer Two devices. A bridge can be used to connect two similar LANs. Bridges connect to hubs. Each bridge examines the destination address in a frame and either forwards this frame onto the next LAN or drops the frame.
A Hub broadcasts every packet to every device port at same data rate. Hubs are generally associated with bridges. Switches perform similar functions, but have significant processing power, speed and features at greater cost.
Switches have hybrid or multi-layer applications. They can be viewed as Layer Two devices with expanded capabilities. Switches function like bridges connecting two or more network segments,and take the place of a hub. A switch can be used to connect two similar LANs. Switches can connect to PCs, hubs, or other switches. Each switch examines the destination address in a frame and either forwards this frame onto the next LAN or drops the frame.
Routers are Layer Three Devices. They operate at the third layer, or OSI network layer, of the packet. Routers modify layer 2 frame headers & trailers so packet can travel end-to-end over many links. These devices connect a LAN to a WAN or a WAN to a WAN. The router modifies an outgoing packet by removing any LAN headers and trailers, and encapsulating the necessary WAN headers and trailers. They they strip off the “layer 2” header (such as Ethernet), and then create a new “layer 2” header for the next hop to the next router or final destination. Routers often incorporate firewall functions to protect end users. .
Gateways are Layer Three Plus Devices. The term gateway was originally synonymous with router, but currently is used as a generic term for devices that connect two networks. Gateways connect dissimilar networks and manipulate or translate unlike protocols and timing. Telecommunications Gateways examples follow.
We will look at several Data Network Applications that are merging circuit and packet switching, and mixing data and telecommunications devices.
Typical data network application.
The above network merges data network concepts and devices with PSTN concepts and devices.
This is a common data application using the PSTN.
IP Telephony mixes traditional circuit switched elements with LAN style packet switching elements. V oice packets are the data units that carry a voice conversation through the Internet. Like all packets, they are routed independently of each other. Because of this, individual packets may experience different delays as they move across a network, or may be dropped entirely. Dropped or delayed packets can adversely affect the quality of a voice call. The public Internet is more susceptible to delays or dropped packets than a managed, private network.
IP Telephony traffic is similar to the packet traffic that is generated by 3rd Generation Mobiles. If you replace the land line telephone with a voice coded (VOCODER) mobile, you will see a future architecture for Wireless. The CODEC like a VOCODER is a two way device. .
Protocol Errors Certain data errors are due to violation of various rules for data transmission. Two examples are: Exceeding a Committed Information Rate (CIR) , causing loss of bit content in excess of that rate. The CIR is a rate of data transmission that the carrier guarantees will be sent between points on the network. Although a network user may transmit at rates exceeding the CIR, there is no guarantee that data over that rate will be sent uncorrupted over the network, especially during periods of high traffic. Data collisions resulting when two terminals illegally transmit at the same time. Ethernet, for example, only allows one terminal at a time to transmit data. A certain level of collisions is normal, typically in the 1% range. When these occur, the data communications system stops all transmission of data, and reinstates transmission one terminal at a time. When the level of collisions exceeds the “normal” threshold, detailed protocol tests are necessary to identify the offending terminals. These data errors are observable using test equipment called Protocol Analyzers , which allow the technician to directly observe data messages as they travel through all or part of a data network. Many protocol analyzers also feature statistical analysis of the condition of the network and its ports, providing a way to find error patterns.
Block Parity Check Block or parallel parity is one of the simpler forms of error detection. It operates on blocks of data words. During a data transmission of several words, a predefined number of these words are grouped together into a block, which will consist of rows and columns of data, one row of parity bits, and one column of parity bits. (See Figure 4) A parity bit is appended to the end of each row and also to the end of each column in the block. The parity bits are based on Boolean Algebra functions called exclusive OR (XOR) and exclusive NOR (XNOR) which have been performed on the data bits in each row and column. Note that in Figure 4 the result is that the sum of the 1’s in each row and column is always an odd number. In other cases, it could set to always be an even number. On the receiving end, the parity bits at the end of the rows and columns are verified as being correct for the data bits which are received. If they are not, an error has occurred somewhere in the transmission process, and the data must be either reconstructed with a more sophisticated algorithm, or resent.
The block parity character is sometimes called the Longitudinal Redundancy Check (LRC) . It is also known as the checksum , because it is formed by performing a binary addition without the carry of each successive character. The limitations of block error checking are that error correction is not possible by this method alone, and it cannot detect more than one error in the block of data. In many cases, the combination of additional cost for more sophisticated methods, the low probability of more than one error per block, and the degree that errors will be critical justify the use of this simple scheme over more complex methods. XMODEM Checksum is another type of checksum used with the XMODEM protocol typically used between personal computers. Under the XMODEM protocol, data is sent in 128 character blocks. The blocks are preceded by a Start of Header (SOH) character and block number information. At the end of the block, the XMODEM checksum is appended. This error detection device is calculated by adding all the 128 bytes in the character block, dividing the result by 255, and taking the remainder as the checksum. In addition to being sent with the block of data, the checksum is computed at the receiving end. The result of the receiving end computation is compared with the value that was sent. If the two agree, the transmission is acknowledged by the receiver as error-free.
Cyclic Redundancy Check (CRC) Cyclic Redundancy Check is the method used to detect line errors in T-1 Extended Superframe transmission. Because it is very accurate, it has also become a standard for many other forms of block data transmission. The methodology of CRC is that the data communications system treats the bits in a data block as the coefficients of a polynomial (eg: x 15 + x 4 + 1). . Internally, the system creates other polynomials which it uses in a mathematical function performed on the original polynomial. (An example is the division of the data block polynomial by the polynomial created by the system.) This function results in a remainder. The same function is performed on the data block at both ends of the data transmission, and remainders are compared. When the transmission is error free, the remainders are identical. In order to make the comparison of the remainders, both the data block and the remainder must be sent over the transmission facility. The exact format for that transmission varies by the application. For T-1, for example, designated framing bits of the Extended SuperFrame are reserved for the CRC result. In other systems, such as local area networks, the CRC result may be appended to the end of the data block being sent. See Figure AA There are several CRC standards, such as CRC-12, CRC-16, and CRC-CCITT. CRC 12 specifies a polynomial of degree 12 (i.e.: x 12 + …..), and CRC-16 and CRC-CCITT specify polynomials of degree 16. With the proper selection of polynomials, undetected errors may be as low as 1 in 10 9.
Automatic Repeat Request (ARQ) Automatic Repeat Request is an error correction method typically used in Local Area Networks. Typically, it works together with an error detection method. After a block of data has been sent, the receiving equipment must signal the transmitting equipment that it has either accepted the data or rejected it. If it has rejected the data, it requests a retransmission. Acceptance is generally indicated by the acronym ACK, and rejection by the acronym NAK. The examination of the data blocks may occur in one of two ways: stop and wait, or continuous. With stop and wait, the transmitter must wait until each block of data is examined and either accepted or rejected before sending another block. This will decrease the transmission rate of the system. With continuous examination, transmission of data blocks continues until a NAK is sent back to the transmitter. At that point, one of two corrective measures may be taken. With selective ARQ, the transmitter will resend only the faulty block of data (the one which generated the NAK on the receiving end). In continuous transmission, ACK and NAK signals are sent on different channels than the actual data, and the system will allow a predetermined number of blocks of data to be transmitted without receiving an acknowledgment or rejection. Transmission of several valid data blocks may therefore occur prior to the retransmission of the faulty block. For selective ARQ to work, the receiver must therefore manage the data with buffers (storage) so that a retransmitted block is correctly reinserted into the data stream where the faulty block would have been. Go back N ARQ simplifies the management of the retransmitted data by having the transmitting end resend all blocks of data from the faulty block to the present block. Because several blocks of data may need to be resent, the transmitter must buffer (store) as many blocks as the receiver will allow to go unacknowledged. The need to retransmit multiple blocks of valid data when a faulty block is detected makes Go back N ARQ inefficient in noisy environments.
Forward Error Correction (FEC) The theory behind Forward Error Correction is to mathematically create a pattern of parity bits from a block of data and send those parity bits along with the data to the receiving end. On the receiving end, another mathematical operation is performed on the parity bits, which should result in a specific bit pattern. If the bit pattern which resulted from the receiving end’s mathematical operation is not as expected, individual bits within the erroneous bit pattern point to where the data bits are in error, and a correction can be applied. Notice that although this process works similarly to the Cyclic Redundancy Check, it goes one step further. Errors are not only detected. They are also corrected. There are several types of Forward Error Correction, and in general, the more complex the mathematical operation, the better the error correction. The mathematics of these methods is beyond the scope of this text. For example, Hamming Code, one of the more common methods of Forward Error Correction, uses matrix algebra to generate the bit pattern which verifies the data. If an error has occurred, the result of the receiving end’s matrix multiplication will point out the particular data bit that was in error. The limitation of Hamming Code is that only one error per word can be corrected.
Bit Error Rate is probably the best known and most widely used indicator of the performance of a digital circuit. The theory of this measurement is that when a known pattern of binary digits is sent on a line, it should be received, detected, and decoded without error on the receiving side of the transmission. To make this measurement valid, however, it is necessary to perform this test over a predetermined time period, and then use the statistical average of the errors as the measurement. This methodology compensates for temporary conditions that may cause a short term high level of errors. There are three common bit pattern lengths: 63, 511, and 2047. The patterns can all 1s (all mark), alternate mark-space patterns, or complex patterns that may not repeat for millions of bits. Bit Error Rate Testers used on high speed digital lines can use patterns as long as 2 23 - 1 bits. In general, the more digits in the pattern, the more stringent the test for error. Manufacturers of digital equipment are free to set their own proprietary standards to measure Bit Error Rate. For example, a digital set top converter manufacturer may use a proprietary string of digits which are detected during the diagnostic phase of converter setup. In this scenario, the indication to the technician performing the setup may be minimal, perhaps only a pass-fail LED readout. More sophisticated test equipment will actually track and record Bit Error Rates over time, and provide that information on printed or graphically displayed outputs.
Block Error Rate is similar to Bit Error Rate, in that a predetermined bit pattern is sent over the line. BLER provides more information in that it tracks data on a block basis, rather than just individual data bits. This type of tracking is helpful in detecting recurring patterns of errors. To be most effective, the test block size should be the same as the typical block of actual data. A measurement called Effective Information Throughput is closely related to the BLER. If the BLER is 1 percent, then one block in every 100 will need to be retransmitted, resulting in an Effective Information Throughput of 99 percent. Errored Seconds and Severely Errored Seconds Any second containing one or more errors is called an error second. Any period of 1 second with a BER exceeding 10 -3 is called a severely errored second.
BERT (Bit Error Rate Tester) is a beginning step in the troubleshooting process. It may be likened to the “idiot lights” that the automotive industry places on dashboards, in that BERT will tell you that you have a problem, but will not provide much more than an indication of how severe the problem is. At a critical BER, the subscriber will encounter severe impairments such as freeze frame on the received video. Bit errors may be due to line or data errors, as we have discussed earlier. To find out more details when the BERT reveals errors above specifications, other tests must be performed, such as the jitter measurements, eye and constellation diagrams, and power level measurements. BERT by itself can be useful as a field measurement, however, when another unit is available. A simple substitution of a second unit followed by another BERT may narrow the problem to the initial piece of equipment. This is probably the most practical course of troubleshooting for a field technician, since most of the other tests require specialized test equipment and the associated training. As mentioned earlier, BERT is the most commonly used measurement of digital errors. To perform a BERT, the system must be equipped with a standard signal source and a detector. Figure 7 illustrates such a setup. In the illustration, a pseudo random bit sequence is generated by a feedback shift register driven by a very stable clock source. The data from the shift register is passed through an interface circuit to generate the correct code format and output level. At the receive end, the same type of interface circuit strips off the code and recovers a clock. This clock drives a reference pseudo random bit sequence generator which has an output that is compared to the received data. When the system is properly synchronized, all errors in the received signal are recorded in the error counter. This combination of signal source may be self contained in the system, or be part of portable test equipment. In portable test equipment, BERT testing is often part of a larger set of tests that can be done on a system.
Data Networks are the future of Wireless and General Telecommunications. They draw upon concepts, devices and terminology from Telecom and Information Technology, LANs and the PSTN, and Circuit and Packet Switching. This merging can be very confusing. The good news is, if you understand either the PSTN, Data or Wireless you have a head start at understanding the integrated world of data networks. The OSI Reference Model provides a unifying structure upon which the new integrated voice and data networks will be based.
Transcript of "Special Material 3"
Copyright 2001 Global Wireless Education Consortium BASIC DATA NETWORKS
Bandwidth Narrowband 0 - 64 kbps Wideband 64 Kbps - 45 Mbps Broadband 45 Mbps and Beyond X.25 Frame Relay ATM SMDS Token Ring FDDI x x T-1, PRI T- 3 SONET LAN Backbone Network Hierarchy and Protocols IP Ethernet Network Access
Definition: Topology The Physical and Logical Way the Network Elements are Connected
Local Area Network (LAN) Topologies Legend Device on LAN = Bus Star Ring
LAN Topologies Example: Ethernet CSM/CD: Carrier Sense Multiple Access/ Collision Detection Send Data Units out unless you detect another data unit. Legend Device on LAN = Bus
LAN Topologies Example: Token Ring Send data units when you possess the token Token: Special Type of Data Unit Ring
LAN Standards IEEE 802.1 INTERNETWORKING lOGICAL LINK CONTRO (LLC) OSI LAYERS 3-7 OSI LAYER 2 OSI LAYER 1 PHYSICAL MEDIUM ACCESS CONTROL (MAC) IEEE 802.2 TYPE 1 - UNACKNOWLEDGED CONNECTIONLESS SERVICE TYPE 2 - CONNECTION MODE SERVICE TYPE 3 - ACKNOWLEDGED CONNECTIONLESS SERVICE 802.3 CSMA/CD (ETHERNET) 802.4 TOKEN BUS 802.5 TOKEN RING 802.6 DQDB (MAN) BASEBAND COAXIAL AND UNSHIELDED TWISTED PAIR BROADBAND COAXIAL BROADBAND COAXIAL OPTICAL FIBER SHIELDED OR UNSHIELDED TWISTED PAIR OPTICAL FIBER OR COAXIAL CABLE
Access Networks: Entry to Wide Area Network (WAN) Enter via port on LAN router Customer Premises Bridge-Router Access Network
Data Flow in an X.25 Network Protocol Layers Involved in Data Transfer 3 2 1 Switch A Switch C X.25 Network Switch B
Data Flow in a Frame Relay Network Frame Relay Network Protocol Layers Involved in Data Transfer 3 2 1 Switch A Switch C Switch B
Backbone Networks High Speed Interconnection of Networks
ATM Cell Structure Header Payload 48 Octets 5 Octets 1 2 3 4 5 Octet 8 7 6 5 4 3 2 1 Bit Position GFC VPI VPI VCI VCI VCI PTI HEC VCI = Virtual Channel ID VPI = Virtual Path ID PTI = Payload Type Identifier GFC = Generic Flow Control CLP = Cell Loss Priority HEC = Header Error Control C L P Backbone Example: Asynchronous Transfer Mode (ATM)
ATM Users Receive as Much Bandwidth as They Require PYTHON WITH EXPANDABLE SKIN (BANDWIDTH)
User 1 User 2 User 3 125 sec 250 sec 375 sec 500 sec Asynchronous means data units (cells) are placed on the network as the application requires User 1 User 1 User 1 User 2 User 2 User 2 User 3
OPEN SYSTEMS INTERFACE (OSI) MODEL Application Presentation Session Transport Network Data Link Physical Application Presentation Session Transport Network Data Link Physical Network Data Link Data Link Network Node 7 6 5 4 3 2 1 7 6 5 4 3 2 1
Wireless OSI Model MESSAGES USE BOTTOM 3 LAYERS Application Presentation Session Transport Network Data Link Physical Application Presentation Session Transport Network Data Link Physical Network Data Link Data Link Network Node 7 6 5 4 3 2 1 7 6 5 4 3 2 1
A Set of Rules The OSI Model Application Presentation Session Transport Network Datalink Physical Header User Info 1 7 6 5 4 3 2
Layer Details – Layer 1 Network Data Link Physical Application Presentation Session Transport Transmits bits received from the Data Link layer across the transmission medium
Layer Details – Layer 2 Data Link Physical Sequences messages and checks for errors between adjacent link stations Transmits bits received from the Data Link layer across the transmission medium Network Application Presentation Session Transport
Layer Details – Layer 3 Network Data Link Physical Fragments or “packetizes” messages, and routes them to the proper destination Sequences messages and checks for errors between adjacent link stations Transmits bits received from the Data Link layer across the transmission medium Application Presentation Session Transport
Layer Details – Layer 4 Network Data Link Physical Transport Fragments or “packetizes” messages, and routes them to the proper destination Sequences messages and checks for errors between adjacent link stations Transmits bits received from the Data Link layer across the transmission medium Provides multiplexing, network connection management, quality of service, etc. Application Presentation Session
<ul><li>Header </li></ul><ul><li>User Data </li></ul><ul><li>Perhaps a Trailer </li></ul><ul><li>Store and Forward </li></ul>Packet Switched Routing
Protocol Data Units Header Information Packet Format Layer 2 Info Layer 3-5 Info Layer 2 Info Flag Flag Message Format Types of Protocol Data Unit Being Sent Through a Packet Switched Network
Protocol Data Unit Routing Fields Header Trailer User Data
Destination Node Won’t Change Address of Next Node on Route Router Stores Packet, Reads Destination Information And Forwards To The Next Node Routers Routing Packets Final Path ROUTER NODE B DATA C C DATA B C NODE A NODE C
Data Switch Routing Example <ul><li>Ethernet Switch </li></ul><ul><ul><li>Receives Ethernet frame </li></ul></ul><ul><ul><li>Looks up 6-byte Destination Address in a Forwarding Table </li></ul></ul><ul><ul><li>Sends frame out only the port associated with the Destination Address </li></ul></ul>
<ul><li>Data Network Devices </li></ul><ul><li>& the OSI Reference Model </li></ul><ul><li>Gateways </li></ul><ul><li>Routers </li></ul><ul><li>Switches </li></ul><ul><li>Hubs </li></ul><ul><li>Bridges </li></ul><ul><li>Repeaters </li></ul>
REPEATER <ul><li>LAYER 1 PHYSICAL </li></ul><ul><li>I PHYSICALLY REPEAT AND REGENERATE BITS FOR MY OWN LAN </li></ul>
BRIDGE <ul><li>LAYER 2 DATA LINK LAYER </li></ul><ul><li>I LINK DATA FRAMES BY BRIDGING BETWEEN IS LAN D S. </li></ul>LAN 1 LAN 2
HUB <ul><li>LAYER 2 DATA LINK LAYER </li></ul><ul><li>I ROUTE PACKETS BETWEEN IS LAN D S WITH COMMON PROTOCOLS </li></ul>USER USER USER USER
DATA SWITCH <ul><li>LAYER 2 DATA LINK LAYER </li></ul><ul><li>I ROUTE PACKETS BETWEEN PORTS </li></ul><ul><li>PACKETS ARE NOT SENT TO EVERY PORT </li></ul><ul><li>(ONLY TO THE DESTINATION PORT) </li></ul>USER USER USER USER
ROUTER <ul><li>LAYER 3 NETWORK LAYER </li></ul><ul><li>I ROUTE PACKETS BETWEEN IS LAN D S WITH COMMON PROTOCOLS </li></ul>
GATEWAY <ul><li>LAYER 3+ </li></ul><ul><li>I CONVERT PROTOCOLS BETWEEN IS LAN D S </li></ul><ul><li>SYNCHRONOUS TO ASYNCHRONOUS </li></ul>SYNC ASYNC
Data Network Devices in Data Network Applications
Typical Packet Data Services Router Router Router Carrier Network Router Chicago Router LAN Hub Client Detroit Router LAN Hub Server Customer Network Customer Network Access Line Access Line
<ul><li>Routers modify layer 2 frame headers & trailers so packet can travel end-to-end over many links. They provide Gateways into and out of the PSTN. </li></ul>IP packets / TR frames IP packets / PPP frames Router or Gateway modifies frame IP packets / TR frames Packets, Routers and the PSTN
Combined Network CMTS MSC TELCO NIU NIU Host Digital Terminal SS7 SS7 Internet Intranet or Public Data Network Average Residence Power Residence Corporate A Corporate B Trunks to Telco Switch Gateway
IP Telephony Using a Computer as the Terminal PSTN Internet CODEC Gateway Router Cable Modem Microphone
PSTN Internet CODEC Gateway Router IP Telephony Generates Mobile Like Traffic Cable Modem Telephone Adapter
<ul><li>Committed Information Rate (CIR) </li></ul><ul><li>Data Collisions </li></ul><ul><li>Protocol Analyzers </li></ul>Detecting and Fixing Errors
Summary <ul><li>Types of Data Networks </li></ul><ul><li>OSI Reference Model </li></ul><ul><li>Packet Structures </li></ul><ul><li>Routing </li></ul><ul><li>Devices </li></ul><ul><li>Applications </li></ul><ul><li>Detecting And Fixing Errors </li></ul>
A particular slide catching your eye?
Clipping is a handy way to collect important slides you want to go back to later.