An Introduction to Ethernet Networking

Ethernet networking is a media access technology used on LANs that allows network hosts to share a link's bandwidth. Not only is Ethernet easy to implement, it is also scalable. Integrating newer Ethernet technologies, such as Fast and Gigabit Ethernet, into an existing network is also easy as is troubleshooting.

The Physical Layer and Ethernet

The standards body responsible for the Physical layer specifications for Ethernet is known as the Electronic Industries Association and the newer Telecommunications Industry Alliance, or EIA/TIA. According to EIA/TIA specifications, Ethernet uses a registered jack (RJ) connector with a 4 5 wiring sequence on unshielded twisted-pair cabling. Known as RJ-45, this specification is also called an 8-pin modular connector.

Ethernet cables lose signal strength as the signal travels through the cable. This signal loss is known as attenuation. Cables are measured in categories with higher quality cables having higher category numbers and lower attenuation.

DIX (Digital, Intel, and Xerox) created and implemented the first Ethernet LAN specification which was later used by the IEEE 802.3 Committee. The 802.3 Committee later evolved, forming different committed to cover the changing technologies: Fast Ethernet (802.3u), Gigabit Ethernet on category 5 (802.3ab) and 10Gbps over fiber and coax (802.3ae). IEEE.org offers free PDF downloads of select IEEE standards including Ethernet standards on its website.

Ethernet networking typically uses CAT5 cables featuring 8-pin modular connectors and the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol. This protocol allows for even bandwidth sharing and solves an earlier problem involving collisions due to simultaneous packet transmissions.

Before a host transmits data, it looks for a digital signal. If no other host is transmitting, then it will begin its transmission, constantly checking for other transmissions. If it encounters another signal, it will issue a "jam" signal which interrupts the other host's transmission. The other host will wait before resuming the transmission. Retransmission depends on "backoff" algorithms and timeouts occur after 15 unsuccessful attempts.

What is Half-Duplex Ethernet?

The original 802.3 Ethernet is half-duplex Ethernet. In a nutshell, half-duplex Ethernet consists of a single wire pair that accommodates a bidirectional digital signal. Half-duplex Ethernet uses the CSMA/CD protocol to prevent and manage collisions and retransmissions. When hubs are connected to switches, they must use half-duplex mode so that the network devices can detect collisions. Typical 10BaseT networks offer from 3 to 4 Mbps, making half-duplex Ethernet about 30 to 40 percent efficient.

What is Full-Duplex Ethernet?

Full-duplex Ethernet uses two pairs of wires and point-to-point connections between sending and receiving devices, making for a faster transmission of data than half-duplex Ethernet. Collisions are avoided because the transmitted and received data travels on different sets of wires like a divided, two way highway. Efficiency improves dramatically with up to an aggregate of 100 percent efficiency both ways. For example, a 10 Mpbs Ethernet in full-duplex mode is capable of 20 Mbps.

Full-duplex Ethernet is typically used to connect: switches to hosts, switches to switches, and hosts to hosts)

The Data Link Layer and Ethernet Addressing

The Data Link layer provides for Ethernet addressing, which is also known as MAC (Media Access Control) addressing. Ethernet receives framing packets from the Network layer and prepares them to be transmitted with the Ethernet contention media access method. Current Ethernet frames included: SNAP, Ethernet II, IEEE 802.2, and IEEE 802.3.

What is Media Access Control? All Ethernet Network Interface Cards (NICs) feature a hexadecimal hardware address known as the MAC address consisting of 48 bits. The first 24 bits are assigned by the IEEE to an organization such as a hardware vendor. This part of the address is known as the Organizationally Unique Identifier (OUI).

Within the OUI, the first bit is the "Individual/Group (I/G)" bit. Depending on its value of either a 0 or 1, it could represent:

• The address of a device (0)
• A broadcast address in Ethernet (1)
• A multicast address in Ethernet (1)
• A broadcast or functional address in TR and FDDI (1)

Following the I/G bit is the G/L bit. The G/L bit is also called the U/L, or Universal, bit. Depending on its value of either a 0 or 1, it could represent:

• A globally administered address such as by the IEEE (0)
• A locally governed and administered address (1)

Beyond the OUI portion of the MAC address, is the vendor-assigned section. The vendor assigns its own unique 24-bit number to the second part of the MAC address. This vendor-assigned number is unique for every single manufactured NIC device. Manufacturers typically assign a series of 24 zeros for the first NIC card manufactured and sequences the following devices in order until the last card features 24 ones. Because of the hexadecimal n umber format, over 16.7 million unique numbers can be generated.

What are Ethernet Frames?

The Data Link layer takes data in bit form and combines them into bytes which are combined into frames. These frames encapsulate data packets from the Network layer for transmission on one of three types of media access: Ethernet / contention, token passing (Token Ring and FDDI), and polling. Using a MAC frame format, Ethernet devices pass data frames and perform a cyclic redundancy check (CRC) for error detection. Below is a table detailing the Ethernet frame types:

Ethernet II	Preamble 8 bytes	DA 6 bytes	SA 6 bytes	Type 2 bytes	Data	FCS 4 bytes
802.3	Preamble 8 bytes	DA 6 bytes	SA 6 bytes	Length 2 bytes	Data	FCS
802.2 (SNAP)	Destination SAP AA (1)	Source SAP AA (2)	Ctrl 03 (1 or 2)	OUI ID (3)	Type (2)	Data (variable)
802.2 (SAP)	Destination SAP (1	Source SAP (1)	Ctrl (1 or 2)	Data (variable)	Data (variable)	Data (variable)

About CBT Planet:

Want to learn more about Ethernet, MAC addressing, or frames? Enroll in CBT Planet's CCNA training program and learn important Cisco networking topics from a certified instructor. CCNA classes provide you with networking knowledge and prepare you for the Cisco Certified Network Associate certification exams. Whether you want to get certified as a CCNA or simply want to expand your internetworking knowledge, these instructor-led classes and videos provide you with comprehensive coverage of Cisco networking. CBT Planet offers the training in several convenient formats including live classes, certification boot camps, a self-study CCNA CBT, virtual classrooms, and CCNA online training.