Tutorial 1: Introduction to Wireless and Mobile Networks

Infrastructure mode uses a base station (AP) as a central hub for all communication and connection to the wired network. Ad hoc mode has no base station; wireless hosts communicate directly in a peer‑to‑peer manner.

c) Collision detection is not straightforward in wireless; it is difficult to detect collisions while transmitting because the transmitter cannot listen simultaneously. Wired Ethernet uses CSMA/CD; wireless uses CSMA/CA.

Multipath propagation occurs when a transmitted signal travels along multiple paths to the receiver due to reflections, diffractions, and scattering. The multiple copies arrive at different times and with different phases, causing constructive/destructive interference (fast fading) and intersymbol interference (ISI) that corrupts the received symbols.

WPAN (~10 m), WLAN (~100 m), and WWAN (regional/global). Sometimes WMAN (metropolitan) is also included.

TCP’s congestion control assumes that packet loss is due to network congestion. On wireless links, loss is often caused by bit errors, interference, or handoff, not congestion. TCP responds by reducing its congestion window, causing unnecessary throughput reduction.

A base station (e.g., AP, eNodeB) is a fixed infrastructure node that bridges wireless hosts to the wired network. Its primary functions include medium access control, association/authentication, power management, and mobility support (handoff).

Two stations A and C are out of range of each other but both are in range of an access point B. A transmits to B; C cannot sense A's transmission (because it is hidden) and also transmits to B, causing a collision at B. RTS/CTS helps by reserving the medium for the duration of the transmission.

User mobility refers to a person moving and potentially using different devices at different locations (e.g., desktop at work, smartphone at home). Device mobility refers to a single device changing its point of network attachment while it is in use.

The path loss exponent α describes how quickly signal power decays with distance (P_r ∝ d^−α). Free space: α=2; urban with moderate obstructions: α=3–4; dense indoor: α=4–6.

Advantages: Free to use, rapid innovation, open for anyone. Disadvantages: No interference protection, crowded, power limitations, no QoS guarantees.

RTS/CTS is a four‑frame handshake (RTS, CTS, Data, ACK) used to reserve the medium for the duration of a transmission. It mitigates the hidden terminal problem by allowing stations that hear either RTS or CTS to set their NAV (Network Allocation Vector) and defer.

C = B log₂(1 + SNR) where C is channel capacity (bps), B is bandwidth (Hz), and SNR is the signal‑to‑noise ratio (linear). It shows that capacity increases linearly with bandwidth and logarithmically with SNR.

Key challenges: IP address changes (which break TCP sessions), handoff latency, session continuity, location management (tracking the node), and handling heterogeneous networks (vertical handoffs).

Slow fading (shadowing) is variation in received power over long distances due to obstructions; it changes slowly (over seconds or minutes). Fast fading is rapid fluctuation due to multipath interference; it changes over fractions of a second and is caused by movement of the receiver or surrounding objects.

The near‑far problem occurs when a transmitter close to the receiver causes high‑power interference, drowning out weaker signals from farther transmitters. It is addressed by power control (adjusting transmit power to equalise received power) and by CDMA systems using orthogonal codes.

Consumer (high bandwidth, low latency), IoT (low power, massive scalability), industrial URLLC (ultra‑reliable low latency), vehicular (high mobility, low latency).

Interference degrades SINR, leading to higher error rates, lower modulation rates, and reduced throughput. Co‑channel interference comes from other transmitters using the same frequency; adjacent‑channel interference comes from neighbouring frequency channels due to imperfect filtering or spectral leakage.

The link layer provides error detection (CRC), error correction (FEC), retransmission (ARQ), medium access control (CSMA/CA), and power management. It also handles association, authentication, and mobility (roaming).

In infrastructure mode, routing is simple: traffic flows between host and AP (or between APs via wired infrastructure). In ad hoc mode, routing is multi‑hop and dynamic; nodes must run distributed routing protocols (e.g., AODV, OLSR) to discover and maintain routes, which adds significant complexity and overhead.

Session continuity ensures that an ongoing communication session (e.g., VoIP call, video stream) is not interrupted when the device moves to a new network attachment point. It is critical for real‑time applications and user experience.

OFDM splits a wideband channel into many narrow orthogonal subcarriers. Each subcarrier experiences flat fading (since its bandwidth is less than the coherence bandwidth), making equalisation simpler (one‑tap per subcarrier) and reducing ISI via a cyclic prefix.

Lower frequencies provide better coverage and penetration (fewer base stations) but limited bandwidth and capacity. Higher frequencies offer more spectrum and capacity but suffer from higher path loss and poor penetration, requiring dense deployment and advanced techniques like beamforming.

NEMO refers to the case where an entire network (e.g., a vehicle, train, or aircraft) moves and all its attached hosts must maintain connectivity. The network as a whole changes its point of attachment, requiring mobility support for the entire subnet. Example: a Wi‑Fi network on a train that connects to the Internet via cellular backhaul.

Wired MAC (e.g., Ethernet) uses CSMA/CD (collision detection) because the medium is isolated and collisions can be detected. Wireless MAC uses CSMA/CA (collision avoidance) because collisions cannot be reliably detected (transmitter cannot listen while transmitting) and because of hidden/exposed terminal issues.

Spectral efficiency measures how much data can be transmitted per unit of bandwidth (bits per second per Hz). It is critical because spectrum is a limited and expensive resource. Higher spectral efficiency allows more capacity without needing more spectrum.

Solution: Deploy 802.11ac/ax APs (one per ~200-300 m² indoors, more in lecture halls). Use 5 GHz for capacity, 2.4 GHz for legacy/coverage. Implement WPA3‑Enterprise with 802.1X (RADIUS). Use a controller‑based architecture with seamless roaming (802.11r). Backhaul: 1 GbE per AP, aggregated to 10 GbE uplinks. Conduct site survey to optimise placement.

UDP is preferable because it has lower latency and does not retransmit lost packets, which is acceptable for live video (some packet loss is tolerable). TCP would add retransmission delay and congestion control, causing jitter and stalls. Use UDP with application‑level error resilience (e.g., FEC) to handle packet loss.

Using L_FS = 92.45 + 20 log₁₀(f) + 20 log₁₀(d) with f in GHz, d in km.

• 2.4 GHz: 10m (0.01 km) → 92.45+7.60-40=60.05 dB; 100m → 80.05 dB; 1000m → 100.05 dB
• 5 GHz: 10m → 92.45+13.98-40=66.43 dB; 100m → 86.43 dB; 1000m → 106.43 dB

Observation: 5 GHz suffers ~6.4 dB more loss than 2.4 GHz at the same distance. Also, loss increases by 20 dB per decade of distance.

Without ARQ: P_success = (0.9)⁵ = 0.59049 ≈ 59%.

With ARQ (up to 3 attempts per hop): P_hop = 1 - (1-0.9)³ = 1 - 0.001 = 0.999. End‑to‑end = (0.999)⁵ ≈ 0.995 ≈ 99.5%.

ARQ significantly improves reliability at the cost of increased latency and overhead.

Scenario: Node A transmits to B, and C can hear A's transmission but wants to transmit to D (which is outside A's range). C defers because it senses the medium busy, even though its transmission to D would not interfere with A↔B. This reduces throughput.

Solution: Use RTS/CTS; C can hear the RTS from A but not the CTS from B, so it knows it is not in the interference range of the receiver and may transmit (if it hears CTS, it defers).

Measurement: Device measures signal strength from neighbouring cells (RSRP/RSRQ). Decision: Network (or device) decides when to hand off based on thresholds, hysteresis, and time‑to‑trigger. Execution: Device disconnects from source cell and connects to target (or make‑before‑break).

Factors affecting latency: measurement interval, decision algorithm, authentication/security procedures, backhaul delays, and whether the handoff is hard or soft.

In infrastructure mode, the AP can buffer packets and the device can use power‑save modes (e.g., 802.11 PS‑Poll) to sleep between beacon intervals. In ad hoc mode, devices must relay packets for others, which consumes additional energy. Therefore, infrastructure is generally more energy‑efficient for sensors, provided APs are available. However, ad hoc may be necessary when no infrastructure exists.

SNR = 10^(25/10) = 316.23. C = 40e6 × log₂(317.23) ≈ 40e6 × 8.31 ≈ 332.4 Mbps.

Actual 802.11ac with 40 MHz, 256‑QAM, 5/6 coding, and short GI gives ~200 Mbps (MCS 9). The difference is due to MAC overhead, pilot subcarriers, guard intervals, and other protocol overhead.

LoRa is suitable due to its long range (up to 15 km) and low power. Use a star‑of‑stars topology with gateways on hills. Sensors sample every 10 minutes and send small packets. Use spreading factors to trade data rate for range. Battery life: >5 years with low duty cycle. Data aggregation at gateways (edge computing) before sending to cloud.

Vertical handoff is a handoff between different network technologies (e.g., Wi‑Fi to cellular). Example: a smartphone moves out of Wi‑Fi coverage and switches to LTE. Challenges: different IP subnets, authentication, QoS mapping, and maintaining session continuity. Requires cross‑layer coordination and protocols like Mobile IP or MPTCP.

Frame size = 1500 bytes = 12000 bits. P_{no error} = (1-10⁻⁵)¹²⁰⁰⁰ ≈ exp(-0.12) ≈ 0.887 (88.7%).

With up to 3 retransmissions (total 4 attempts): P_success = 1 - (0.113)⁴ ≈ 1 - 0.000163 ≈ 0.9998 (99.98%).

Ad hoc mode is generally more vulnerable because there is no central authentication server, encryption is often weaker (or absent), and routing can be attacked (e.g., blackhole, wormhole). Infrastructure mode supports enterprise‑grade security (WPA2‑Enterprise, 802.1X) and centralised monitoring. However, the infrastructure itself (APs, controllers) becomes a potential target.

Summary: 802.11 (2 Mbps) → 802.11b (11 Mbps) → 802.11a/g (54 Mbps) → 802.11n (600 Mbps, MIMO) → 802.11ac (Gbps, wider channels, MU‑MIMO) → 802.11ax (Wi‑Fi 6, OFDMA, target wake time) → 802.11be (Wi‑Fi 7, 320 MHz channels, 4096‑QAM, multi‑link operation).

Frequency reuse uses the same frequencies in cells separated by sufficient distance to avoid interference. Reuse factor N defines the cluster size; capacity per cell ≈ total bandwidth / N. Cell splitting reduces cell radius (R) and increases number of cells, thereby increasing capacity (since smaller cells reuse frequencies more densely). Capacity increases approximately as (R_old/R_new)².

DCF uses CSMA/CA. A station senses the channel; if idle for DIFS, it transmits. If busy, it waits until idle, then chooses a random backoff counter from [0, CW] where CW starts at CW_min and doubles with each collision (up to CW_max). The counter decrements when the channel is idle; when it reaches zero, the station transmits. This exponential backoff reduces collision probability after repeated collisions and provides fairness by randomising access.

Mobile IP (v4 and v6) uses a Home Agent (HA) and Foreign Agent (FA) or the mobile node itself to maintain a Care‑of Address (CoA) that reflects its current location. Packets are tunnelled from HA to CoA. MIPv6 uses route optimisation to avoid triangular routing. Proxy MIPv6 moves the mobility function to the network (access routers) so that the mobile node does not need to participate, simplifying the client.

For 802.11ac, MCS 8 at 80 MHz gives 780 Mbps (with short GI). After MAC overhead (acknowledgments, contention, overhead), typical throughput is about 50‑60% of PHY rate, so ~390‑468 Mbps, but actual may be lower due to collisions and retransmissions.

Doppler shift causes frequency offset (f_d = v f_c / c). At high speeds, this can cause phase rotation and loss of orthogonality in OFDM, leading to ICI. Techniques: pilot subcarriers for frequency tracking, advanced channel estimation, and using wider subcarrier spacing (e.g., 5G NR supports 15, 30, 60 kHz spacing).

Wi‑Fi is cost‑effective, supports high throughput, and works with existing client devices. However, it may not provide deterministic QoS or seamless mobility. 5G small cells offer better mobility, lower latency, and QoS guarantees but are more expensive and require new infrastructure. Hybrid approaches are common.

Network slicing creates multiple virtual networks on a common physical infrastructure. Each slice is optimised for a specific service type: eMBB for high throughput, URLLC for low latency/reliability, mMTC for massive connectivity. Slices are isolated in terms of resources (compute, storage, bandwidth) and management, enabling simultaneous support of diverse requirements.

Massive MIMO uses a large number of antennas (e.g., 64, 128) at the base station to simultaneously serve multiple users on the same time‑frequency resource via spatial multiplexing. Beamforming focuses the transmitted energy towards specific users, increasing SINR and reducing interference. This improves both coverage (by directing power) and capacity (by enabling multi‑user MIMO).

v = 60 km/h = 16.67 m/s. f_d = v f_c / c = 16.67 × 5e9 / 3e8 ≈ 277.8 Hz. Coherence time ≈ 1/(4×277.8) ≈ 0.9 ms. Therefore, channel estimation should be updated at least every millisecond.

LTE uses coordinated approaches like Inter‑Cell Interference Coordination (ICIC) to allocate different frequency or power resources to edge users, and enhanced ICIC (eICIC) with time‑domain muting. Wi‑Fi uses distributed mechanisms: Clear Channel Assessment (CCA) to detect energy, dynamic channel selection (DFS) to avoid radar, and rate adaptation to adjust modulation when interference is high. LTE is centrally managed; Wi‑Fi is distributed.

While wireless is increasingly prevalent, wired networks (fibre) offer higher capacity, lower latency, absolute reliability, and security. Wireless is subject to interference, spectrum scarcity, and environmental factors. However, for many consumer and enterprise access networks, wireless is the primary means. Backbone and data‑centre networks will remain wired due to capacity demands. A hybrid approach is likely to persist.