Ethernet and the IEEE 802.3 Standards

• Ethernet transport takes advantage of the bus and star topologies.
• Uses a control method known as Carrier Sense Multiple Access with Collision Detection (CSMA/CD)
• CSMA/CD is an algorithm that transmits and decodes formatted data frames.
• The algorithm is a block of computer code or logic designed to accomplish a certain task or to solve a problem, such as to control data communications or to calculate a payroll deduction.
• There are several characteristics of Ethernet communications over a single cable run:

• No central control governs data transmission
• The network can be accessed from many points, and all nodes have equal ability to transmit on the network
• Data units are transmitted as encapsulated frames
• Each transmitted frame contains addresses of the sending and receiving nodes
• Frame addressing includes the ability to specify one node, multiple nodes, or all nodes on the network.
• The packets reach every network node
• It is the responsibility of the receiving node to recognize and accept frames with its address as the destination
• Data transmission is relatively fast
• Efficient error detection keeps transmission delays to a minimum
• The communication specification discourages the implementing of special features that might lead to incompatible network variations

• When data is transmitted in Ethernet communications, it is encapsulated in frames:

Ethernet II

• An Ethernet frame-formatting method used on the internet and other modern networks.
• Slight variation from the IEEE 802.3 Standard, making transmissions more efficient.
• It’s frames look like this:

• To avoid communication problems, Ethernet II and standard IEEE 802.3 frames should not be used on the same network.

Signal Transmission

• Using CSMA/CD the Ethernet sending node encapsulates the frame to prepare it for transmission.
• All nodes that want to transmit a frame over the cable are in contention with one another.
• No single node has priority over another node.
• The nodes listen for any packet traffic on the cable.
• If a packet is detected, the non-sending nodes go into a defer mode.
• Ethernet protocol permits only one node to transmit at a time.
• Signal transmission is accomplished by carrier sense.
• When no traffic is detected on the communication media for a given amount of time, any node is eligible to transmit.
• Carrier sense is the process of checking a communication medium, such as cable, for a specific voltage or signal level that indicates the presence of a data-carrying signal.
• When more than one node transmits at a time, that is called a collision.
• A transmitting node uses the collision detection algorithm to recover from packet collisions.

Network Addressing

• Ethernet addressing is accomplished through the OSI MAC(media access control) sub-layer within the data link layer.
• The MAC sub-layer uses an address associated with the NIC to direct encapsulated data to the data link layer of the receiving node.
• Every network node has a unique physical MAC address that is burned into a programmable read-only memory chip (PROM) in that node’s NIC.

Ethernet Physical Layer Media

• Ethernet data encapsulation and de-capsulation are performed at the physical and data link layers by the NIC in the network node.
• The NIC is a transceiver (transmitter and receiver) and provides channel access to coaxial, twisted-pair, or fiber-optic cable.
• The NIC contains the algorithms to receive, de-capsulate, transmit, or defer activity, collision detection, and collision response.
• The software algorithms that perform those functions are compiled into programs and related files called network drivers.
• Ethernet communications can use any of several different types of cable media:

• Thick coaxial
• Thin coaxial
• STP (Shielded twisted-pair)
• UTP (Unshielded twisted-pair)
• Fiber-optic

Token Ring and the IEEE 802.5 Standards

• Developed in the 70s by IBM and remains a primary LAN technology.
• The token ring transport method uses a physical star topology but with the logic of a ring topology.
• Although each node is connected to a central hub, the packet travels from node to node as though there is no starting or ending point.
• Nodes are joined by use of a multi-station access unit (MAU).
• A specialized token is continuously transmitted on the ring to coordinate when a node can send a packet.
• The token is 24 bits in length, with 3 8-bit fields.
• These fields are:

• The starting delimiter (SD)
• Access Control (AC)
• The ending delimiter (ED)

• In most implementations, only one token is available on the ring, although the IEEE specifications allow two tokens for networks operating at 16 Mbps or faster.
• When a node wants to transmit, it must capture the token. No other node can capture the token until the active node is finished.
• The packet is sent around the ring until it is read by the target node. It continues around the ring until the original transmitting station picks it up and checks the token to determine if it was received.

Error Determination Through Beaconing

• Each token ring network designates one node as the active monitor, usually the first station to be recognized when the network is brought up.
• The monitor is responsible for packet timing on the network and for issuing new token frames if problems occur.
• Every few seconds the active monitor broadcasts a MAC sub-layer frame to show it is functioning properly.
• Periodically, the standby monitors broadcast frames called "standby monitor present" frames to show that each is working normally and is available to become the active monitor in case the current active monitor station malfunctions.
• If no broadcasts are detected from the active monitor or one of the standby monitors, the ring goes into a beaconing condition.
• Beaconing starts when a node sends a beacon frame to indicate that it has detected a problem.
• The ring tries to self correct the problem by assigning a new standby monitor if the original has gone out of action.
• When beaconing occurs, no data tokens are transmitted until the problem is resolved.

Token Ring Communications Media

• Token ring communications use shielded and unshielded twisted-pair and fiber-optic cabling.
• Seven IBM cable types are used with token ring communications:

• Token ring networks are extremely reliable and for that reason are used in some mission-critical situations.
• One advantage of token ring networks over Ethernet networks is that broadcast storms and workstation interference are rare.
• Broadcast storms occur on an Ethernet network when one or more nodes broadcast continuously, interrupting or delaying successful network transmissions.
• Broadcast storms are rare on token ring networks, because only one node can transmit at a time.

Microsoft Networking Services

NetBIOS and NetBEUI

• NetBIOS Extended User Interface (NetBEUI) is a communications protocol that is native to Microsoft network communications.
• Developed by IBM in 1985, it is an enhancement of Network Basic Input/Output System (NetBIOS)
• NetBIOS, which is not a protocol, is a method for interfacing with network services. It also provides a naming convention.

NetBIOS

• NetBIOS names are used to name objects on a network, such as a workstation, server, or printer.
• The NetBIOS Name Query services translates those names, which make it easy for humans to identify particular network resources, into addresses for network communications.
• Each name must be unique.
• Names must not be longer than 16 characters.

NetBEUI

• Was developed when computer networking primarily meant local area network with about 2-200 computers.
• NetBEUI is best suited for:

• Microsoft Windows 3.1 or 3.11
• Microsoft Windows 95
• Microsoft LAN Manager
• Microsoft LAN Manager for UNIX
• Microsoft Windows NT
• IBM PCLAN
• IBM LAN Server

• NetBEUI is a protocol that you can use to connect to Windows NT, Windows for Workgroups, or LAN manager servers.

NDIS

• A software driver specification that enables Microsoft network protocols to communicate with a NIC.
• Binding a protocol to a NIC is accomplished through the NDIS driver.

ODI

• Another driver that can be used to transport multiple protocols is the Open Data-Link Interface (ODI).
• This driver is used on Novell NetWare networks to support communications with NetWare file servers, mainframe computers, and minicomputers, and the Internet, similar to NDIS.
• Enhanced mode is when the computer processor is set to indirectly access computer memory, enabling it to run several programs at once.
• Real mode is when a computer processor is set to run programs so the programs have direct access to memory locations at 1024 K or below and only one program can be run at a time.

IPX and NWLink

• Novell adapter one of the early LAN protocols, the Xerox Network System (XNS) protocol, for use with its NetWare file server operating system. Novell’s adaptation is called the Internet Packet Exchange (IPX) protocol for use with Novell NetWare.
• One advantage IPX has over NetBEUI is that is has routing capabilities, so data is transported over multiple networks in an enterprise.
• SPX is a Novell protocol used for network transport when there is a particular need for data reliability.
• NWLink is a network protocol that simulates the IPX/SPX protocol for Microsoft Windows 95 and NT communications with Novell NetWare file servers and compatible devices.
• The best way to install NWLink is as part of Client Service for NetWare, which installs the following three elements:

• Client Service for NetWare
• NWLink IPX/SPX Compatible Transport
• NWLink NetBIOS

• NWLink offers several advantages, such as routing over enterprise networks.
• It is easy to install and provides more effective communications with NetWare file servers than the ODI driver.
• Its disadvantages are that it is not transported as fast as NetBEUI.
• The most common situations for using Microsoft’s NWLink to emulate IPX/SPX are:

• to enable a workstation running Windows 95 or Windows NT to communicate with one or more NetWare servers
• set up a Microsoft Windows NT server as a gateway to one or more NetWare servers

TCP/IP

• Many network users need to connect to a host computer, such as a mainframe running IBM’s Multiple Virtual Storage (MVS) operating system or a minicomputer with UNIX.
• Another common need is to connect to a host computer that provides access to the internet or to a web server.
• The protocol for those jobs is Transmission Control Protocol/Internet Protocol (TCP/IP)
• Used around the world.
• TCP/IP is many protocols wrapped into one, all working together to establish the most error-free communications possible.
• A terminal is a device consisting of a monitor and a keyboard to communicate with a host computer that runs the programs. The terminal does not have a processor to use for running programs locally.

IP Addressing

• Before setting up TCP/IP, you need to decide how to set up the IP addressing on the network.
• Static addressing is an IP addressing method that requires the network administrator to assign and set up manually a unique network address on each workstation connected to a network.
• Dynamic addressing is a method in which an IP address is assigned to a workstation without the need for the network administrator to hard-code it in the workstation’s network setup.
• Dynamic addressing automatically assigns an IP address to a computer each time it is logged on.
• An IP address is leased to a particular computer for a defined period of time.
• This addressing method used the Dynamic Host Configuration Protocol (DHCP), which is a convention supported by Microsoft for dynamic addressing.
• The protocol enables a server with DHCP services to detect the presence of a new workstation and to assign an IP address to that workstation.
• Windows Internet Naming Service (WINS) is able to translate a workstation name into an IP address for Internet communications.
• A subnet mask is a method to show which part of the IP address uniquely identifies the network and which part uniquely identifies the workstation.

TCP/IP Advantages and Disadvantages

• One situation in which you would use TCP/IP would be on a large enterprise network, such as on a college or business campus that extensively uses routers and connectivity to mainframe or UNIX computers.
• You would also use it on smaller network situations, where 100-200 Windows-based workstations access intranet or Internet services through an NT server offering Web services from Microsoft’s Internet Information Server.

Protocols Associated with TCP/IP

• Complementing the main protocols are four application services provided through TCP/IP:

• Telnet
• File Transfer Protocol (FTP)
• Simple Main Transfer Protocol (SMTP)
• Domain Name Service (DNS)

• Telnet is an application protocol within TCP/IP that provides support for terminal emulation.
• Telnet enables a user to connect to a host computer so that the host responds as though it were connected to a terminal.
• FTP is an algorithm that enables the transfer of data from one remote device to another, using TCP and Telnet protocols.
• FTP is designed to transfer entire files only in bulk. It does not provide the capability to transfer a portion of a file or records within a file.
• FTP is a single stream of data concluded by an end-of-file delimiter.
• Network File System (NFS) is a popular alternative to FTP. Used mostly in UNIX environments.
• SMTP is designed for the exchange of electronic mail between networked systems.
• Messages sent through SMTP have two parts:

• An address header
• The message

• The address header can be very long, because it contains the address of every SMTP node through which it has traveled and a date stamp for each transfer point.
• DNS is used to translate domain computer names to an IP address.
• The process of translating names to addresses is called resolution.

Connecting to IBM Mainframes with DLC

• DLC is the Data Link Control Protocol.
• DLC is needed to connect to IBM computers when TCP/IP is not available.
• It does that by providing connectivity to IBM’s communication system, called systems network architecture (SNA)
• DLC is not routable.

Connecting to Macintosh Computers with AppleTalk

• AppleTalk is a peer to peer protocol used on networks for communications between Macintosh computers
• On a Microsoft network, Macintosh computers are linked in by setting up the Windows NT server services for Macintosh.
• The NT server becomes a file server for Macintosh computers as well as for computers running Microsoft operating systems.

Setting Protocol Priority

• When setting up a workstation to use multiple protocols, you may experience slow response from the protocol you use most.
• In the properties of the network, you can select which protocol takes priority over another.

Resolving a NIC Resource Conflict

• When setting up the NIC, check for resource conflicts to make sure communications work properly.
• If there is a resource conflict, an asterisk will appear next to the conflict.

Selecting The Right Protocol

• The protocols you employ on a network depend on several factors:

• Do frames need to be routed?
• Is the network smaller or larger?
• Are there Microsoft NT servers?
• Are there mainframe host computers?
• Is there direct access to internet or web based intranet applications?
• Are there mission-critical applications?

• In many cases, it is necessary to use a combination of protocols for different types of network applications.
• Modern networks often combine the major protocols, TCP/IP, NetBEUI, and IPX/SPX.