Local Area Networks (LAN) notes
Principles of Telecommunications: Local Area Networks (LAN)
6.1 Introduction
- The term "local" in LAN refers to networks that are small and managed by a single organization.
- LANs provide benefits such as:
- Sharing of peripheral devices.
- Business applications.
- Organizational databases.
- File storage.
- Enabling efficient organizational workflows.
- Electronic communication for employees.
6.2 Local Area Network Architecture
- Key questions addressed include:
- Difference between physical and logical network topology.
- Hub vs. switch devices.
- Centralized vs. decentralized control (Peer-to-peer vs. Client/Server).
- How CSMA/CD works with Ethernet LANs.
- Difference between shared and switched full-duplex Ethernet.
- LANs are characterized by their data link layer and physical layer topologies.
- Physical topology: Describes how nodes are connected using guided mediums like UTP, coaxial, or fiber optic cables.
- Logical topology: Describes the communication mechanism for data exchange.
- Physical and logical topologies may be the same or different, influenced by interface devices.
- Figure 6.1(a): Star topology with a HUB.
- A hub is a central node that broadcasts any transmission from a single node to all connected nodes.
- The logical topology is similar to a bus configuration (Figure 6.2), where all nodes receive communications from any node.
- Figure 6.1(b): Star topology with a switch.
- A switch replaces the hub and forwards data only to the intended destination based on source and destination addresses.
- The physical star topology can act as a logical star, with connections going through the central switch.
- The physical star can also act as a logical ring using centralized access control (e.g., tokens), as shown in Figure 6.3.
- LANs are built on peer-to-peer (P2P) or client-server models.
- Peer-to-Peer (P2P) Networks:
- All computers participate equally.
- Lack centralized authentication, data storage, and may have multiple data copies on several computers.
- Generally a poor choice for organizations.
- Client-Server Model:
- Dedicated servers provide centralized services and data storage accessed by client computers.
- Servers offer centralized services:
- Sharing peripherals (printers, scanners, etc.).
- Applications and databases.
- Email.
- Security functions (authentication, access control).
- Gateway to external networks.
6.2.1 Centralized and Decentralized Access Control
- Centralized Access Control:
- Involves passing a token or polling by a master controller.
- Token Ring: Passing a token (a special data frame without payload data); only the node with the token can transmit.
- Centralized control leads to deterministic access, guaranteeing sole access to the medium for the token holder.
- Polling: A master controller polls each node to see if it has data to transmit; if so, the polled node can send data.
- Decentralized Access Control:
- No centralized process; each node determines access to the shared medium.
- Nodes listen to the medium to detect traffic before transmitting.
- Potential data collisions can occur due to simultaneous transmissions or propagation delays.
- An algorithm is needed to recover from data collisions (e.g., CSMA/CD).
- Figure 6.4: Data collisions due to propagation delay.
- Node A transmits at t=0 sec.
- Node D, the furthest node, detects the transmission at Δt sec (propagation delay).
- If D transmits between t and Δt believing the medium is free, a collision occurs.
- Successful transmission without interference is non-deterministic (not guaranteed).
6.3 IEEE 802.3 Ethernet LAN
- With shared Ethernet, all nodes share a common medium using decentralized access control.
- Example: Hubs in a physical star or shared physical bus.
- Multiple hubs can be connected in a partial mesh network (Figure 6.5).
- Shared Ethernet LANs require CSMA/CD.
- With switched Ethernet, each node is connected directly to the switch via dedicated full-duplex links.
- Dedicated links with data buffers prevent data collisions.
- CSMA/CD is not needed in switched Ethernet.
- Table 6.1: IEEE 802.3 specification examples (year, standard name).
- IEEE 802.3 (1983): 10Base5 (Thicknet).
- IEEE 802.3a (1985): 10Base2 (Thinnet, Cheapernet).
- IEEE 802.3b (1985): 10Broad36 (coaxial cable).
- IEEE 802.3i (1990): 10Base-T (twisted pair).
- IEEE 802.3j (1993): 10Base-F (optical fiber).
- IEEE 802.3u (1995): 100Base-TX, 100Base-T4, 100Base-FX (Fast Ethernet).
- IEEE 802.3y (1998): 100Base-T2 (twisted pair).
- IEEE 802.3z (1998): 1000Base-X (optical fiber).
- IEEE 802.3ab (1999): 1000Base-T (Gigabit Ethernet).
- IEEE 802.3ae (2002): 10GBase-SR, 10GBase-LR, 10GBase-ER, 10GBase-SW, 10GBase-LW, 10GBase-EW (optical fiber).
- IEEE 802.af (2003): Power over Ethernet (PoE).
- IEEE 802.3an (2006): 10GBase-T (twisted pair).
- IEEE 802.3aq (2006): 10Gbase-LRM (MMF cable).
- IEEE 802.3ax (2008): Link Aggregation
- IEEE 802.3az (2010): Energy Efficient Ethernet.
- IEEE 802.3ba (2010): 40Gbps, 100Gbps.
6.3.1 Physical and Data Link Layers
- The Ethernet protocol consists of a physical layer and data link layer
- The data link layer divides into the logical link control (LLC) and the medium access control (MAC) sub-layers. (Figure 6.6) compares Layers 1 and 2 with the OSI RM.
- The physical layer's mechanical and electrical specifications include:
- The type of medium and mechanical interfaces.
- The type of line coding used (e.g., Manchester, NRZ, 4B5B, etc.), timing, and voltage levels.
- Attaching and removing preambles for timing and synchronization.
- The data link, MAC sub-layer, is responsible for:
- Generating the preamble (if used).
- Creating the Ethernet data frame (Figure 6.7), which encapsulates data from the LLC sub-layer, source (sender) and destination (receiver) 48-bit MAC addresses which resides in the network interface card (NIC) firmware.
- Generating CRC (cyclic redundancy check) for error detection (frame check sequence, FCS).
- Executing the CSMA/CD protocol.
- The LLC sub-layer:
- Manages the protocol interaction between the data link layer 2 and the network layer 3.
- Is responsible for the interface between the MAC sub-layer and network layer protocols (e.g., IP).
- Prepares packets for MAC sub-layer framing by multiplexing network layer 3 packets.
- Demultiplexes data frames upon receipt from the MAC sub-layer for transfer to the network layer.
- Provides connectionless or connection-oriented services.
6.3.2 Ethernet 802.3 Selected Standards
- Carrier Sense Multiple Access Collision Detection (CSMA/CD) is a method for handling access and data collisions.
- "Carrier Sense" means all nodes listen to the medium continuously for data traffic or data collisions.
- "Multiple Access" means that all nodes on the LAN access and share the same medium.
- If a data collision occurs, the "Collision Detection" algorithm is initiated by the closest node detecting the collision.
- The node detecting the collision sends out a jamming signal informing all network nodes to cease transmit activity for a random period of time.
- Ethernet LANs require greater data capacity which is achieved through innovation and careful selection of guided mediums (TP, Coaxial cable, fiber optics).
- Factors to consider:
- Shannon Capacity Theorem, Nyquist Capacity, and Hartley’s Law.
- Channel bandwidth, SNR, and M’ary line coding techniques.
- Challenges such as signal attenuation, noise (e.g., crosstalk), and impedance (line resistance and reactance).
- The quality of twisted pair (TP) conductive medium is a determining factor for data capacity.
- TP wire bundles consist of wire pairs with different twist ratios to combat crosstalk
- UTP (unshielded twisted pair) does not have a shielding to prevent RFI.
- TP cables can be constructed with a conductive (i.e., grounded) shielding placed around each wire pair (shielded TP, STP), around the entire cable (shielding over UTP), or both (shielding over STP).
- Shielding reduces signal distortion and improves SNR, which in turn improves capacity.
- The Telecommunications Industry Association (TIA) and Electronic Industries Alliance (EIA) developed TP cable standards include Cat 5e, Cat 6 (Cat 6a), and Cat 8.
- Cat 5e is between 24 and 26 AWG, while Cat 6a can be between 22 and 26 AWG.
- Wire diameter is only one of several parameters that determine the capability of a cable. Others include the use of specific terminal devices, type of cable shielding, cable’s specified bandwidth, and specified operating distances.
- Note the smaller the AWG number, the larger the conductor diameter.
6.3.2.1 TP Cable Standards
- Table 6.2: Highlights physical differences between Cat 5e and Cat 6a cables.
- Includes frequency, bandwidth, capacity, shielding, physical connector standards.
6.3.2.2 100Base-T (Fast Ethernet)
- 100Base-TX operates over Cat 5e UTP cable in either HDX or FDX mode. Cat5e cabling can support a symbol rate up to 125 Mbaud.
- In FDX mode, two pairs are used, one for transmit and one for receive. The encoding method uses 4B5B, which codes 4 bits of data into a 5-bit word, meaning a 54=80% information rate and a 51=20% overhead.
- The line code used is NRZI (non-return to zero inverted), M=2, +v for a logical "0", -v for a logical "1". Per Hartley's Law, M=2 tells us that the baud rate equals the data rate.
- Each pair in a Cat5e cable can support 125 Mbaud, therefore each pair supports 125 Mbps. However, since 20% of our bit rate is overhead, the true information rate becomes (80%)×(125E6bps)=100Mbps. Therefore 100Base-TX can support 100 Mbps.
- C(bps)=baud×N, where N=log2(M)
- C(bps)=125E6baud×log2(2)=125E6bps=125Mbps
- Information rate: 54×125Mbps=100Mbps
6.3.2.3 1000Base-T (GbE)
- Cat 5e, 125 MBaud each, four pairs (two pairs for transmit, two pairs for receive); two pairs each direction gives us 250 MBaud in each direction
- PAM-5, M=5, four values per symbol, with one value supporting FEC
- C(bps)=baud×log<em>2(M)=250MBaud×log</em>2(4)=250MBaud×2=500Mbps each direction
- hybrid canceller transceiver, which essentially cancels interfering data signals and enables FDX operations on each wire pair. This doubles the capacity of our LAN to 2×500Mbps=1Gbps (see figure 6.8).
6.3.2.4 10GBase-T
- Cat 6 or Cat 7 enables higher signaling rates, lower signal resistance, higher bandwidth than Cat 5e
- support 600 MBaud and 750 MBaud, respectively
- hybrid canceller transceivers to enable FDX on each cable pair
- PAM-16 (pulse amplitude modulation, M=16), N=log2(16)=4 bits/symbol
- Cat 6:
- C(bps per FDX pair)=600MBaud×4 bits/symbol=2.4Gbps
- C(all four pairs, FDX)=4×2.4Gbps=9.6Gbps
- Cat 7:
- C(bps per FDX pair)=750MBaud×4 bits/symbol=3Gbps
- C(all four pairs, FDX)=4×3Gbps=12Gbps
6.3.2.5 Ethernet and Fiber Optic Cables
- Today, single mode fiber (SMF) and multimode fiber (MMF) optic cables are popular due to decreasing costs and advances in splicing fibers.
- Benefits:
- Security (difficult to tap).
- Immunity to RFI.
- Greater distances compared to TP.
- Greater bandwidth capability.
- Optical signals require conversion of electrical signals to optical signals and vice versa.
- Very long fiber runs may need repeaters and optical amplifiers, increasing complexity.
6.3.2.5.1 100Base-FX
- In fiber optic cabling (100Base-FX), "FX" indicates fiber optic cables at speeds up to 100 Mbps (fast Ethernet over fiber), baseband signaling in either half-duplex (HDX) or full-duplex (FDX) modes.
- Transceivers perform the optical-electrical-optical (OEO) conversions between electrical devices (e.g., computers, repeaters, hubs, or switches) and the optical fiber cable.
- 100BASE-FX is configured using MMF in FDX mode with a central switch enabling maximum distance between device and switch of 2 km.
- Replacing MMF with SMF cables extends this distance to 10 km.
- Similar to 100BASE-TX, it uses 4B5B encoding and NRZI line coding methods. An alternative version of 100BASE-FX is 100BASE-SX, which uses a lower cost multimode fiber operating at a wavelength of 850 nanometer (nm) which reduces distances to 300 m.
6.3.2.5.2 1000Base-SX and LX
- IEEE 802.3z describes several 1GbE (1 Gigabit Ethernet) standards including 1000BASE-SX, 1000BASE-LX, and 1000BASE-CX.
- 1000BASE-SX operates at the 770 to 860 nm wavelength in either HDX or FDX modes.
- Two thicknesses of MMF can be used, either 62.5 µm, which support distances up to 275 m, or 50 µm which support distances up to 316 m.
- The smaller wavelengths associated with SX makes this standard suitable for short distances (i.e., within a building) for high data rate requirements.
- The 1000BASE-LX standard also supports 62.5 and 50 µm MMF, as well as SMF. The use of SMF and longer wavelength signaling (i.e., less attenuation) enables LX to support distances up to 5 km. Like SX, LX can also operate in the HDX or FDX modes.
- 1000BASE-CX uses a 9-pin shielded copper cable that has a maximum distance of 25 m.
6.3.2.5.3 10GBase-SR, 10GBase-LR, 10GBase-ER
- Several 10Gbps standards exist over fiber optic cables.
- 10GBASE-SR (Short Reach) operates in the 850 nm wavelength over MMF with a maximum distance of 400 m.
- 10GBASE-LR (Long Reach) operates at the 1310 nm wavelength, which experiences less attenuation than the SR operating wavelength and can therefore travel greater distances up to 10km.
- 10GBASE-ER (Extended Reach) operating at the 1550 nm wavelength supports an even greater distance of 40 km over SMF.
6.3.2.6 Ethernet and Fiber Optic Examples
- Table 6.2: Examples of Ethernet standards using fiber optic cabling (Ethernet 802.3, Fiber Optic Cable, Throughput, Transmission Distances).
- 100BASE-LX: SMF, 100Mbps, 20km.
- 100BASE-FX: MMF, 100Mbps, 2km.
- 100BASE-EX: SMF, 100Mbps, 40km.
- 1000BASE-LX: SMF, 1Gbps, 5km.
- 1000BASE-SX: MMF, 1Gbps, 550m.
- 10GBASE-LR: SMF, 10Gbps, 25km.
- 10GBASE-ER: SMF, 10Gbps, 40km.
- 10GBASE-SR: MMF, 10Gbps, 300m.
- 10GBASE-SW: MMF, 10Gbps, 300m.
- 10GBASE-LW: SMF, 10Gbps, 25km.
- 10GBASE-EW: SMF, 10Gbps, 40km.
- 40GBASE-LR4: SMF, 40Gbps, 10km.
- 40GBASE-SR4: MMF, 40Gbps, 100m.
- 100GBASE-LR4: SMF, 100Gbps, 10km.
- 100GBASE-ER4: SMF, 100Gbps, 40km.
- 100GBASE-SR10: MMF, 100Gbps, 150m.