
Home >Administration >Information Systems >Tutorials >Networking Basics
Networking Basics
Computer Networks Introduction
The OSI
Model
Network Topologies
Extension/Segmentation
Modules
Four Eras of
Networking
![]()
Excerpt from Computer Intelligent Agents ManualThe OSI Model
The Open Systems Interconnection Reference Model (OSI) for networks was developed in 1978.
The OSI defines the following seven network layers:
7.
6.
5.
4.
3.
2.
Logical Link Control
Media Access Control1.
The three lower layers constitute the subnet portion of the network model and are implemented at the hardware level. The three upper layers are usually implemented by networking software on the node. The middle transport layer is also implemented on the node.
Data packets are assembled as they pass down through the protocol layers of the transmitting workstation and disassembled as they pass up through the protocol layers of the receiving workstation.
Physical Layer
Transmits bits over a communication channel in digital or analog form. It is oblivious to what those bits represent. It deals with the electrical, mechanical and procedural interfaces over the physical transmission medium.
Physical layer interfaces include RS-232 and RS-449.
Some Local Area Network (LAN) protocols are:
Ethernet II, Ethernet 802.3, Token Ring, ARCnet, Fiber Distributed Data Interface (FDDI), Integrated Systems Data Network (ISDN).
Data Link Layer
Defines the rules for sending and receiving information across the physical connection between two systems.
Groups the transmitted bits into logical units called frames.
Marks the beginning and ending of a frame by special bit patterns.
Divides a frame into the following four fields:
Address-Control-Data-Error ControlIs the first layer in which checksum error control occurs.
Is divided into two sub-layers; the lower Media Access Control (MAC), and the higher Logical Link Control (LLC).
Bridges operate at this layer level.
Some common protocols for the data link layer are:
- LAN drivers and access methods, such as Ethernet and Token Ring
- Microsoft's Network Driver Interface Specification (NDIS)
- Novell's Open Data-link Interface (ODI)
- Wide Area Network (WAN) protocols such as,
- Packet-switched:
Asynchronous Transfer Mode (ATM), Frame Relay.
- Serial lines:
Point-to-Point (PPP), Serial Line Internet Protocol (SLIP)
Network Layer
Extends the node to node data link connection across the network.
Organizes frames into packets and decides how to route them.
Provides a uniform addressing mechanism so more than one network can be interconnected. For example a connection between an Ethernet and a Token Ring network occurs at the network layer.
Routers operate at this layer level.
Some protocols for the network layer are:
- Internet Packet Exchange (IPX)
- Internet Protocol (IP)
- Address Resolution Protocol (ARP broadcast)
- X.25 protocol
Protocols at the network layer provide connectionless datagram services.
Connectionless implies that prior to data transmission, no control packets are sent to establish a connection (session). After transmission, no teardown of the connection is necessary. No guarantees are made about the successful arrival of a packet.A datagram is a packet for a connectionless protocol.
Transport Layer
Ensures reliable data delivery and end-to end data integrity by providing connection-oriented services between two end systems.
Provides the last chance for error recovery.
Creates several logical connections over the same network connection (multiplexing)
Creates transport addresses by combining the network layer address and a transport layer Service Access Point (SAP) number.
Transport addresses are also called sockets or port numbers.
Some protocols of the transport layer are:
- Sequenced Packet Exchange (SPX)
- Transfer Control Protocol (TCP)
- User Datagram protocol (UDP connectionless)
- Microsoft NetBEUI
Protocols at the transport layer provide connection oriented services.
Connection oriented protocols are used when a reliable virtual-circuit connection (session) is needed between two nodes.
Session Layer
Provides general activity management.
Permits full duplex or half-duplex communications across a connection.
Some Session layer protocols are:
- File Transfer Protocol (FTP)
- Simple Network Management Protocol (SNMP)
- Telnet
- Simple Mail Transfer Protocol (SMTP)
- Microsoft NetBIOS
Presentation Layer
Manages the way data is represented (ASCII, EBCDIC or other character sets).
Protocols at this layer are part of the operating system and application the user runs in a workstation. Information is formatted for display or printing in this layer.
Date encryption and the translation of other character sets are also handled in this layer.
Application Layer
Contains the protocols and functions user applications require to perform communication tasks. These services are called Application Programming Interfaces (APIs).
Example of an API for NetWare is the NetWare Core Protocol (NCP).
Some application layer protocols are:
- Services Advertising Protocol (SAP) for NetWare
- Routing Information Protocol (RIP)
- Domain Name Services (DNS)
- Message Handling System (X.400)
- Directory Services (X.500)
- Windows Sockets
- Multipurpose Internet Mail Extension (MIME)
OSI
NetWare
UNIX
Apple
LAN Manager
Application
Netware Core Protocol (NCP)
Network Filing System (NFS)
AppleShare Server message blocks
Presentation
AppleTalk Filing Protocol Session
Named Pipes, NetBIOS SNMP, FTP, SMTP, Telnet ASP, ADSP, ZIP, PAP NetBIOS, Named Pipes Transport
SPX TCP/UDP ATP, NBP, AEP, RTMP NetBEUI
Network
IPX IP Datagram Delivery Protocol Data Link
LAN drivers, ODI, NDIS LAN drivers, MAC LAN drivers, LocalTalk, EtherTalk, TokenTalk
LAN drivers, NDIS P h y s i c a l Network Topologies
The geometrical arrangement of the wiring scheme of a network is called the topology.
Topology examples include star, bus, and ring.
The physical network topology describes the way the cables are routed. You can see the physical network topology.
The logical network topology describes the way the network behaves. You cannot see the logical topology; it is the network from the perspective of data transmission. A logical network has its own unique network address and each node on this network is identified by its unique MAC address.
For example an Ethernet network can have either bus (thin/thick net) or star (10Base-T) as its physical topology, but its logical topology is always bus.
In a NetWare environment network numbers are given by the network administrator. In a TCP/IP network environment connected to the Internet you need to obtain a unique network ID portion from the Internet Network Information Center (InterNIC).
Networks connected by routers need their own network numbers.
A network node is defined as any intelligent device that can be addressed by another intelligent device on the network.
All nodes on the same physical network must have the same network number.
Ethernet
Ethernet was approved by IEEE as the 802.3 standard on February 1980. Be aware that the Ethernet-II standard (DIX), and the IEEE 802.3 standard are not exactly the same.
The Ethernet media access mechanism is called a Carrier Sense with Multiple Access /Collision Detection (CSMA/CD).
Ethernet is described as a "listen before talking" protocol. An Ethernet node listens for activity on the transmission channel (carrier) before it transmits.
When a node detects a busy channel, it refrains from transmission. After the last bit of the passing frame, the Ethernet data link layer continues to wait a random time interval for a minimum of 9.6 microseconds. After this random inter-frame spacing a transmission can be initiated.
If a node tries to transmit when the channel is busy, a garbled transmission (collision) results.
Ethernet's minimum frame size is 64 bytes and maximum is 1,518 bytes. Three octets of a frame always represent a manufacturer's code followed by the next three octets that are assigned by the manufacturer a unique MAC address. This way each Ethernet network card has a unique six-octet MAC address.
Ethernet speed is 10 Mbs (10BASE-T)
Fast Ethernet speed is 100 Mbs (100BASE-X)
Gigabit Ethernet speed is 1000 Mbs
100VG-AnyLAN speed is 100 Mbs with demand priority access methodThe maximum span of an Ethernet LAN is 2,500 meters using multiple segments and a max of four repeaters (hubs). By using bridges and/or routers you can go beyond the 2,500 meter limit. After a packet crosses a bridge/router it is on another logical LAN, the repeater count is reset and it is subjected to the restrictions of that LAN only. A single Ethernet LAN can have 8,000 attached workstations.
Ethernet wiring topologies
- 10BASE-5Coaxial cable with maximum segment length of 500 meters or 1,640 feet (bus topology).
- 10BASE-2
Coaxial cable (RG-58) with maximum segment length of 185 meters or 607 feet (bus topology).- 10BASE-T
Twisted-pair cable with maximum segment length of 100 meters or 328 feet (star topology).- 10BASE-F
Fiber-optic cable backbone only (cannot attach workstations) of up to 4 kilometers.10BASE-T Wiring Rules
The 10 stands for the 10Mbs transmission speed, BASE stands for the baseband (square waves) operation, and the T stands for twisted pair wiring.
It is in a star physical network topology.
Unshielded Twisted Pair (UTP) Category 5 is the most common wiring used for 10BASE-T.
UTP Category 5 is the only cable that will allow speeds up to 100 Mbs.
Maximum segment length for UTP is 100 meters or 328 feet.
Up to 1,024 workstations are possible on a 10BASE-T network without using bridges.
It uses RJ-45 type of connectors officially called Media-Dependent Interface (MDI).
Two twisted pairs (four wires) are between each NIC and the concentrator (hub). Each two-wire path forms a simplex link segment. One simplex segment is used for transmitting and the other for receiving. Only two pairs -- one for transmission (TD) and another for receiving (RD) -- are used.
A crossover function is implemented in every twisted pair link so that the transmitter at one end will be connected to the receiver at the other.
Pin Assignments for the MDI Connector
Pin # MDI signal
- TD+
- TD-
- RD+
- Not used
- Not used
- RD-
- Not used
- Not used
Ethernet Frame Formats
- Ethernet_II
Used on TCP/IP , AppleTalk Phase I and DEC networks.- Ethernet_802.3
Used on Novell NetWare 3.x networks.- Ethernet_802.2
Used on Novell NetWare 4.x networks.- Ethernet_SNAP
Used on AppleTalk Phase II networks.
Network Extension/Segmentation Modules
Concentrators/Repeaters (hubs)
Operate at the physical layer of the OSI model.
Perform the following functions:
- Data packet signal retiming and amplification.
- Link Integrity Test on each port.
- Port autopartitioning, which disconnects the port in the event of 30 consecutive collisions or jabber input.
- Send packets through all its ports except the one that received it (even collisions!).
Do NOT provide network segmentation.
Maximum of four concentrators (the four-repeater rule) in the data path between any two nodes applies.
Bridges
Connect two separate network segments to form a single logical network.
Can connect similar or different networks (e.g. Ethernet to Ethernet or Ethernet to Token Ring)
Operate at the data link layer of the OSI model and rely on MAC addresses for their operation.
Provide network segmentation.
Have storage capacity to store frames and act as a store-and-forward device.
Each bridge port has a unique MAC address itself.
Do not propagate Ethernet collisions.
Help solve traffic bottleneck problems.
Cannot make decisions about packet routes.
Cannot prevent broadcast storms. They use every 60 seconds broadcasts to communicate with other bridges.
Routers (also known as Gateways)
Connect two separate physical networks to form a single logical network.
Can connect similar or different networks using similar or different network protocols (e.g. IPX and IP).
Operate at the network layer of the OSI model and they rely on network and node addresses for their operation.
Provide network segmentation.
Are aware of many possible paths to get to a destination and are aware also of which path is optimal by using cost metrics.
Cost metrics are usually based on hops.
A hop is a path between two store-and-forward devices such as a router.
They store routing tables and are slower than bridges.
They can use dynamic or static routing tables.
Dynamic routing tables are set and updated automatically via the Routing Information Protocol (RIP), Open Shortest Path First (OSPF), XNS or other routing protocols. Static routing tables are entered manually into the router.
Not all network protocols are routable. For example the NetBIOS protocol is not routable and it cannot cross a router, but it can cross a bridge.
They can filter out certain type of network traffic and prevent broadcast storms.
In an IP network a router decrements the Time To Live (TTL) by at least 1 or more. If the TTL reaches zero, the packet is discarded.
Switches
Operate at the data link layer of the OSI model and rely on MAC addresses for their operation.
Can be described as multiport bridges.
Allow dedicate maximum bandwidth availability on every switched port.
Provide network segmentation.
Use modern Application Specific Integrated Circuits (ASIC) technology and RISC processors.
They can contain network-layer routing services as well as MAC-layer switching services.
Require very little configuration.
Provide more aggregate bandwidth to the network.
Allow the creation of Virtual LANs (VLAN).
The Four Eras of Networking
Era 1: Mainframe Networks (1965 - 1975)
Era 2: Minicomputer Networks (1975 - 1985)
Era 3: Shared-Bandwidth LAN's (1985 - 1995)
Era 4: Switching (1995 - )