Insem Examination Bank: Unit 2

Section A: Cellular System Design

Explain the concept of frequency reuse in cellular networks. If a total of 33MHz of bandwidth is allocated to a particular FDD cellular telephone system which uses two 25kHz simplex channels to provide full duplex voice and control channels, compute the number of channels available per cell if a system uses a 4-cell reuse pattern and a 7-cell reuse pattern.

Frequency reuse is the foundational architectural concept that makes massive, scalable cellular networks physically and economically possible. Unlike high-power television broadcasting, where a single massive antenna blasts a signal across an entire state on an exclusive frequency, cellular networks operate on the principle of extreme spatial limitation. Base stations deliberately transmit at very low power levels. Consequently, the RF signal attenuates rapidly over distance. This physical limitation is weaponized as an advantage: it allows the network operator to take the exact same frequency channel and reuse it in another cell, provided the two cells are geographically separated by a “reuse distance” sufficient to attenuate the signal and prevent intolerable Co-Channel Interference. This concept allows operators to serve millions of simultaneous subscribers using only a tiny, finite sliver of licensed spectrum.

To compute the capacity, we first determine the total number of channels the system can support. Given a total licensed bandwidth of 33 MHz ($33,000$ kHz), and knowing that full-duplex communication (Frequency Division Duplexing) requires two distinct 25 kHz simplex channels (one for uplink, one for downlink), the total bandwidth required per individual duplex channel is 50 kHz.

The total number of available channels ($S$) for the entire system is simply the total bandwidth divided by the channel bandwidth:

$$S = 33,000 \text{ kHz} / 50 \text{ kHz} = 660 \text{ total channels}.$$

In a cellular architecture, these $S$ channels must be divided equally among $N$ cells to form a “cluster” where no frequencies overlap.

For a tight 4-cell reuse pattern ($N=4$), the number of channels available per cell ($k$) is:

$$k = S / N = 660 / 4 = 165 \text{ channels per cell}.$$

For a more conservative 7-cell reuse pattern ($N=7$), which provides greater geographical separation between interfering cells at the cost of lower capacity, the channels available per cell ($k$) is:

$$k = S / N = 660 / 7 \approx 94 \text{ channels per cell}.$$

Define Co-Channel Interference. Derive the equation for the Signal-to-Interference Ratio (SIR) in a hexagonal cell layout. If the path loss exponent $n=4$, calculate the theoretical SIR for a cluster size of $N=7$.

Co-Channel Interference (CCI) is the primary, inescapable performance-limiting factor in any frequency-reuse system. It is the interference caused exclusively by signals originating from other, distant base stations that are legitimately operating on the exact same frequency channel as the serving base station. Unlike thermal noise, which can be overcome by simply increasing transmitter power, increasing power in a cellular network proportionally increases the interference blasted into neighboring co-channel cells, meaning CCI cannot be defeated by raw power. It must be managed by geometry.

In a standardized hexagonal grid layout, the worst-case interference scenario involves a mobile unit located at the absolute edge of its serving cell. In this topology, the first tier of interferers consists of 6 equidistant co-channel base stations surrounding the serving cell.

Let $S$ be the desired signal power from the serving base station at distance $R$ (the cell radius). Let $I_i$ be the interference power from the $i$-th co-channel base station at distance $D$. Because received power drops exponentially according to the path loss exponent $n$, the Signal-to-Interference Ratio (SIR) is mathematically derived as:

$$ \frac{S}{I} = \frac{S}{\sum_{i=1}^{6} I_i} \approx \frac{R^{-n}}{\sum_{i=1}^{6} D_i^{-n}} $$

Assuming all 6 interfering base stations are roughly at the same distance $D$ from the mobile unit, we can simplify the summation:

$$ \frac{S}{I} \approx \frac{R^{-n}}{6(D^{-n})} = \frac{1}{6} \left( \frac{D}{R} \right)^n $$

We know from hexagonal geometry that the co-channel reuse ratio $Q$, which dictates the distance between co-channel cells, is defined as $Q = \frac{D}{R} = \sqrt{3N}$, where $N$ is the cluster size. Substituting $Q$ into the equation yields the final SIR formula:

$$ \frac{S}{I} = \frac{(\sqrt{3N})^n}{6} $$

To calculate the theoretical SIR for a system operating in a typical urban environment with a path loss exponent $n=4$ and utilizing a 7-cell cluster size ($N=7$):

$$ \frac{S}{I} = \frac{(\sqrt{3(7)})^4}{6} = \frac{(\sqrt{21})^4}{6} = \frac{21^2}{6} = \frac{441}{6} = 73.5 $$

Converting this linear power ratio into decibels provides a more standardized engineering metric:

$$ \text{SIR (dB)} = 10 \log_{10}(73.5) \approx 18.66 \text{ dB} $$

This 18.66 dB SIR is generally considered the minimum acceptable threshold for high-quality analog voice or robust digital transmission without severe bit errors.

Compare Fixed Channel Assignment (FCA) and Dynamic Channel Assignment (DCA). Under what traffic conditions does DCA significantly outperform FCA, and what is its primary computational drawback?

In Fixed Channel Assignment (FCA), the available radio spectrum is statically partitioned. Each cell in the network is permanently, rigidly allocated a fixed subset of the total available channels. This allocation is programmed into the base station upon deployment. If a massive crowd gathers in one specific cell and consumes all of its allocated channels, any new call attempts in that cell are instantly blocked, even if the adjacent cell is completely empty and its channels are sitting idle. The primary advantage of FCA is its extreme computational simplicity; the network requires zero real-time logic to manage allocations.

Dynamic Channel Assignment (DCA) abandons static partitioning entirely. In a DCA architecture, base stations hold no channels permanently. Instead, all frequencies are kept in a centralized, global pool managed by the Mobile Switching Center (MSC). When a user attempts to make a call, the base station requests a channel from the MSC. The MSC executes a real-time algorithmic analysis, evaluating current Signal-to-Interference Ratio (SIR) measurements across the entire network. It dynamically loans a channel to the requesting base station, provided that allocating that specific frequency in that specific location will not violate the minimum SIR threshold for any existing users.

DCA significantly outperforms FCA in environments characterized by bursty, unpredictable, or highly localized, asymmetrical traffic spikes. A prime example is a massive stadium immediately following a sporting event, or a highway during a localized traffic jam. DCA allows the network to fluidly funnel idle channels from empty residential areas into the congested zone, effectively absorbing the spike without dropping calls.

However, DCA suffers from a massive computational drawback. It requires immense, continuous processing power at the MSC to calculate complex interference matrices across hundreds of cells in real-time, hundreds of times per second. Furthermore, it generates massive signaling overhead on the backhaul network, as base stations must constantly report real-time SIR measurements to the MSC to inform the allocation algorithms.

Describe the techniques of Cell Splitting and Sectorization. How do these techniques increase the capacity of a cellular network without requiring additional frequency spectrum?

Cell Splitting and Sectorization are critical network engineering techniques used to massively increase the total capacity of a cellular network when the operator has exhausted their allocated frequency spectrum and cannot simply buy more bandwidth. Both techniques achieve capacity gains by fundamentally altering the physical layout to improve frequency reuse.

Cell Splitting is deployed when user density in a specific geographical area (a macrocell) exceeds the channel capacity of the central tower. The network operator physically subdivides the large macrocell into several smaller microcells. Each new microcell requires the installation of its own low-power base station. Because the transmit power is reduced, the interference footprint shrinks, allowing the same frequencies to be reused much closer together. The mathematical mechanism for capacity increase is straightforward: the total System Capacity ($C$) is defined as $C = M \times S$, where $M$ is the total number of clusters in the area and $S$ is the total channels available. By shrinking the cells, the operator fits far more clusters ($M$) into the same physical city, scaling the total capacity linearly without requiring a single extra Hertz of spectrum.

Sectorization takes a different approach, focusing on reducing Co-Channel Interference rather than shrinking the cell radius. Instead of deploying a single omnidirectional antenna in the center of the cell that blasts energy in a 360-degree circle, the operator installs directional antennas. A common configuration uses three 120-degree antennas or six 60-degree antennas on the same tower, dividing the cell into isolated “sectors.”

This physical isolation drastically reduces Co-Channel Interference. In an omnidirectional hexagonal grid, a user is exposed to interference from 6 surrounding co-channel cells. In a 120-degree sectorized system, the directional antennas mean the user only “sees” interference from 2 of those cells; the other 4 are physically blocked by the antenna patterns. Because the interference drops so drastically, the Signal-to-Interference Ratio (SIR) improves massively. The network operator exploits this improved SIR by deploying a tighter cluster size (e.g., dropping from an $N=7$ pattern to an $N=4$ or $N=3$ pattern). A tighter cluster size allocates more channels to each individual cell, directly increasing capacity without requiring new spectrum.

Section B: GSM Architecture and Protocols

Diagram the Network and Switching Subsystem (NSS) of GSM. Detail the specific roles of the MSC, HLR, VLR, AuC, and EIR in processing a Mobile Originated Call (MOC).

The Network and Switching Subsystem (NSS) is the highly intelligent, digital routing core of the GSM architecture. It is responsible for all call processing, mobility management, authentication, and billing. It completely isolates the radio access network (the BSS) from the external public telephone networks.

The Mobile Switching Center (MSC) is the central router and the heart of the NSS. When a user initiates a Mobile Originated Call (MOC), the setup request travels from the base station to the MSC. The MSC is responsible for physically routing the audio circuits to the external Public Switched Telephone Network (PSTN), tracking the duration of the call, and generating the Call Detail Records (CDRs) required for billing the user.

The Visitor Location Register (VLR) is an incredibly fast, volatile database physically co-located with the MSC. It holds a temporary, cached copy of the service profile for every roaming user currently located within the MSC’s geographic jurisdiction. During the MOC, the MSC queries the VLR to verify that the user possesses the appropriate subscription rights to execute an outbound call (e.g., verifying the account is not suspended for non-payment, or checking if international dialing is allowed).

The Home Location Register (HLR) is the massive, permanent master database. It stores the definitive subscription profile and the current location pointer for every user belonging to that network operator. Notably, the HLR is not directly involved in processing an active, standard MOC. It is only queried during initial registration or if the VLR’s cache is corrupted.

The Authentication Center (AuC) is a highly secure, heavily guarded cryptographic server, usually co-located with the HLR. It houses the ultra-secret master keys ($K_i$) for every SIM card. During call setup, the AuC generates the cryptographic security triplets (RAND, SRES, $K_c$) and pushes them to the VLR. The VLR uses these triplets to mathematically authenticate the user before the MSC is allowed to connect the audio circuit.

The Equipment Identity Register (EIR) is an optional security database. It contains lists of all valid and invalid mobile hardware identifiers (IMEIs). During call setup, the MSC can query the EIR with the phone’s physical IMEI. If the IMEI is listed on the central blacklist (indicating the phone was reported stolen), the MSC will drop the call immediately, rendering the stolen hardware useless, regardless of whether the user has inserted a valid, paying SIM card.

Describe the structure of the GSM physical layer. Explain how GSM utilizes a hybrid FDMA/TDMA access scheme. What is the duration of a standard GSM TDMA frame, and why are guard periods necessary?

The physical layer of GSM is a masterclass in maximizing the efficiency of limited radio spectrum through complex, hybrid multiplexing. The architecture does not rely on a single access method, but rather combines two orthogonal techniques to slice the spectrum into hundreds of thousands of discrete, usable communication channels.

The first layer of multiplexing is Frequency Division Multiple Access (FDMA). GSM takes the entire allocated 25 MHz spectrum block and slices it into 124 discrete, narrow carrier frequencies. Each carrier frequency is exactly 200 kHz wide.

The second layer of multiplexing is Time Division Multiple Access (TDMA). Rather than giving one user a dedicated 200 kHz carrier (which would limit the network to only 124 simultaneous users), GSM takes each 200 kHz carrier and divides it in the time domain. The time axis is structured into an infinite series of repeating TDMA frames.

A standard GSM TDMA frame has a highly specific duration of 4.615 milliseconds. This frame is further subdivided into 8 discrete time slots, each lasting exactly 577 microseconds. A single user is assigned one specific time slot within the frame. The phone powers on its transmitter, blasts a massive burst of digital data for half a millisecond, and then shuts off for the remaining seven time slots, allowing seven other users to share the exact same 200 kHz frequency channel sequentially.

Guard periods are an absolute physical necessity in any TDMA system due to the speed of light. Consider a user standing at the extreme edge of a large 35-kilometer macrocell, and a second user standing 10 meters away from the base station tower. When both phones transmit their bursts according to the synchronized clock, the edge user’s burst will take significantly longer to travel through the air. By the time it arrives at the base station, it will be delayed. Without mitigation, the tail end of the edge user’s burst would physically bleed into the beginning of the adjacent user’s time slot, causing a catastrophic collision and data corruption. Guard periods are tiny, blank slivers of time (typically 30 microseconds) appended to the end of every transmitted burst. They act as temporal shock absorbers, absorbing the propagation delay and guaranteeing that bursts arrive cleanly separated at the receiver.

Outline the challenge-response authentication mechanism in GSM. Explain the specific roles of the A3, A5, and A8 algorithms.

The GSM authentication framework relies on a highly secure Challenge-Response protocol designed to mathematically verify the identity of the user without ever transmitting their secret password over the vulnerable air interface, effectively nullifying replay attacks and radio sniffers.

The process begins when the network (specifically the VLR, acting on behalf of the AuC) transmits a “Challenge” to the mobile station. This challenge is a 128-bit completely random number, known as the RAND.

The intelligence of the system resides entirely within the physical SIM card, which houses a highly classified, permanently embedded 128-bit secret key ($K_i$). When the SIM receives the RAND from the network, it executes the A3 Algorithm. The A3 algorithm takes the network’s public RAND and the secret $K_i$ as inputs, processes them through a one-way cryptographic hash function, and outputs a 32-bit Signed Response (SRES). The phone transmits this SRES back to the network. The network independently calculates the expected SRES using its own copy of $K_i$. If the phone’s SRES perfectly matches the network’s expected SRES, the user is authenticated. Crucially, the secret $K_i$ never leaves the SIM card.

Simultaneously, the SIM executes the A8 Algorithm. The A8 algorithm also takes the RAND and the secret $K_i$ as inputs, but processes them to generate a completely different output: a 64-bit temporary session key known as the Ciphering Key ($K_c$).

Once authentication is successful, the phone and the Base Station transition to encrypting the actual voice and data traffic. They both utilize the A5 Algorithm. The A5 algorithm is a stream cipher. It takes the newly generated 64-bit $K_c$ and the current TDMA frame number as inputs, and generates a massive pseudorandom bitstream. This continuous bitstream is XORed directly against the digitized voice data. Because the TDMA frame number increments every 4.6 milliseconds, the encryption key effectively changes continuously, making the air interface incredibly difficult to decrypt in real-time.

Explain why the IMSI is rarely transmitted over the air interface in GSM. Describe the mechanism by which the VLR assigns and manages the Temporary Mobile Subscriber Identity (TMSI).

The International Mobile Subscriber Identity (IMSI) is the permanent, globally unique 15-digit identifier permanently burned into the SIM card. It is the ultimate digital identity of the user. If the IMSI were transmitted in plaintext over the air interface every time the phone received a call or moved to a new cell, the system would possess a fatal privacy flaw. Malicious actors, intelligence agencies, or stalkers could deploy cheap software-defined radios across a city, passively sniff the unencrypted paging channels, and permanently track the physical location and movement patterns of specific individuals in real-time.

To preserve anonymity and thwart tracking, GSM was engineered so that the IMSI is transmitted over the air interface only under extreme circumstances, such as the very first time a phone powers on in a completely new country, or if a database catastrophic failure occurs.

Under all normal operating conditions, identity is masked using the Temporary Mobile Subscriber Identity (TMSI). Upon successful cryptographic authentication, the Visitor Location Register (VLR) generates the TMSI—a completely random, 32-bit mathematical string. The VLR transmits this TMSI to the phone over an encrypted channel. The VLR securely stores the mapping between the temporary TMSI and the permanent IMSI deep within its protected internal RAM.

For all subsequent network interactions—including paging the phone for an incoming call, setting up data sessions, or performing periodic location updates—the network and the phone communicate using only this random TMSI. To external observers sniffing the radio waves, the traffic appears to belong to a random, anonymous string of numbers. Furthermore, to prevent correlation attacks, the TMSI is highly volatile. Every time the user moves across a Location Area boundary, or sometimes even after a single phone call, the VLR generates an entirely new TMSI, invalidates the old one, and forces the phone to update. This continuous rotation makes sustained, passive tracking mathematically impossible.

Section C: 2.5G and 3G Evolution

How does GPRS alter the fundamental circuit-switched paradigm of GSM? Detail the roles of the SGSN and GGSN. Why is statistical multiplexing superior to HSCSD for web browsing traffic?

General Packet Radio Service (GPRS) represents a monumental architectural pivot, marking the transition of cellular networks from legacy circuit-switched voice systems into the modern era of packet-switched internet routing.

Standard GSM is inherently circuit-switched. When a user makes a call or initiates a data session, the MSC allocates a dedicated, physical copper/fiber circuit through the network and locks a specific TDMA time slot on the radio tower exclusively for that user. That circuit remains locked for the entire duration of the call, regardless of whether anyone is actually speaking or transmitting data. This architecture is perfect for real-time voice, but catastrophically inefficient for internet data.

GPRS alters this by overlaying a completely separate, parallel packet-switched routing infrastructure onto the existing GSM radio towers. It introduces two massive new routing nodes. The Serving GPRS Support Node (SGSN) is the packet equivalent of the MSC/VLR. It handles complex mobility management, packet routing, logical link management, and authentication specifically for data traffic within its assigned geographical area.

The Gateway GPRS Support Node (GGSN) is the crucial boundary node. It acts as the colossal IP router that physically connects the walled-garden mobile network to the external public Internet. The GGSN is responsible for assigning IP addresses to mobile phones, executing deep packet inspection, managing firewall rules, and anchoring the Mobile IP tunnels.

The superiority of GPRS becomes glaringly obvious when comparing its statistical multiplexing to early attempts at speeding up data, like High-Speed Circuit-Switched Data (HSCSD). HSCSD attempted to increase speeds by dedicating multiple continuous TDMA time slots to a single user. However, internet traffic (like web browsing) is intensely bursty. A user downloads a webpage in two seconds, and then spends two minutes reading it. Under HSCSD, those multiple time slots remain locked and unusable by anyone else during those two minutes of reading, wasting massive radio bandwidth.

GPRS utilizes statistical multiplexing. Time slots are not permanently assigned; they are pooled. When a user clicks a link, the network dynamically assigns time slots for the few milliseconds required to burst the packets. The instant the download finishes, the time slots are immediately released back to the pool. While User A is reading their downloaded webpage, User B can instantly utilize those exact same time slots to download their email. This allows the network to serve vastly more users on the exact same limited spectrum.

Describe the architecture of the UMTS Terrestrial Radio Access Network (UTRAN). What are the specific functions of the Node B and the Radio Network Controller (RNC)?

The transition to 3G UMTS required the complete replacement of the legacy GSM Base Station Subsystem (BSS) to support the massive bandwidth and complex signaling demands of Wideband CDMA. The new architecture is called the UMTS Terrestrial Radio Access Network (UTRAN).

The foundation of UTRAN is the Node B, which serves as the UMTS equivalent of the GSM Base Transceiver Station (BTS). However, the Node B is significantly more computationally intensive. It handles the brutal mathematics of the physical layer. It is responsible for the massive spreading and despreading operations of the W-CDMA orthogonal codes, executing complex Forward Error Correction (FEC) algorithms, and interleaving data to survive burst errors. Crucially, because W-CDMA is intensely susceptible to the Near-Far problem, the Node B must execute critical inner-loop power control, sending commands to the mobile phones to adjust their transmit power 1500 times every single second to maintain perfect signal equilibrium.

Above the Node Bs sits the Radio Network Controller (RNC), the highly intelligent equivalent of the GSM BSC. The RNC acts as the absolute dictator of the radio spectrum within its domain. It manages sophisticated radio resource allocation, executes admission control algorithms to ensure the network doesn’t collapse under load, and handles the heavy cryptographic encryption of the data streams.

Most importantly, the RNC is the master orchestrator of Macrodiversity. Unlike GSM hard handoffs, UMTS utilizes Soft Handoffs. When a user is at a cell boundary, their phone communicates with two or three Node Bs simultaneously. All these Node Bs stream the received data up the backhaul fiber to the RNC. The RNC mathematically aligns these disparate, simultaneous data streams in time, evaluates their checksums, and dynamically selects the highest-quality frames from any stream on a millisecond-by-millisecond basis, synthesizing a pristine signal out of edge-case fading.

Explain the transition from narrow-band TDMA in GSM to Wideband CDMA in UMTS. How does the high chipping rate of W-CDMA (3.84 Mcps) provide resilience against multipath fading?

The evolution from 2G GSM to 3G UMTS represents a complete abandonment of narrow-band time-slicing in favor of massive, mathematics-driven spectral spreading.

GSM utilized a narrow-band approach. It carved the spectrum into tiny 200 kHz channels and physically separated users by isolating them in distinct time slots (TDMA). UMTS, utilizing Wideband CDMA (W-CDMA), discards time slots entirely. It allocates massive, continuous 5 MHz channel blocks. In W-CDMA, every single user within the cell transmits on the exact same 5 MHz frequency, at the exact same time. The users do not interfere with each other because their digital signals are multiplied by mathematically unique, orthogonal pseudo-noise codes, allowing the receiver to filter out everyone else as mere background static.

This transition to a massive bandwidth specifically engineered immense resilience against multipath fading, fundamentally leveraging the physics of the chipping rate. In W-CDMA, the data is spread using a chipping rate of 3.84 Mega-chips per second (Mcps). This extreme speed means the physical duration of a single chip ($T_c$) is incredibly short—approximately 0.26 microseconds.

This short chip duration is a powerful weapon against multipath reflections. When a radio signal bounces off a building, the delayed reflection arrives at the receiver slightly later than the direct Line-of-Sight signal. If this delayed reflection arrives separated by more than $0.26 \mu s$ (which corresponds to a path difference of roughly 78 meters), the receiver’s correlator can mathematically distinguish the reflection from the primary signal. Because the delayed reflection’s code no longer perfectly aligns in time, the correlator treats it as uncorrelated, orthogonal noise and effectively rejects it.

Furthermore, UMTS utilizes a RAKE receiver. The RAKE receiver actively hunts for these delayed multipath reflections (called “fingers”). Instead of rejecting them, it captures them, delays the primary signal to perfectly align the phases, and sums their energy constructively. The massive bandwidth of W-CDMA allows the system to turn multipath fading from a catastrophic cause of Inter-Symbol Interference into a measurable signal gain.

Compare a Basic Service Set (BSS) with an Independent Basic Service Set (IBSS) in Wi-Fi. Explain the operation of the Distributed Coordination Function (DCF) using CSMA/CA, specifically detailing the use of DIFS, SIFS, and random exponential backoff.

The IEEE 802.11 standard (Wi-Fi) defines two fundamentally different topological architectures for organizing wireless nodes.

A Basic Service Set (BSS) represents the standard infrastructure mode utilized in homes and offices. It revolves around a centralized Access Point (AP). The AP acts as the master clock, dictating timing synchronization through periodic beacons, routing all traffic between wireless stations, and serving as the physical bridge to the wired Ethernet internet backbone. An Independent Basic Service Set (IBSS) represents an Ad-Hoc mode. In an IBSS, there is no centralized AP. Stations communicate directly peer-to-peer, routing data among themselves in a decentralized mesh, typically used for temporary file sharing or emergency sensor networks.

Regardless of the topology, the critical challenge in Wi-Fi is avoiding collisions on the shared, unlicensed spectrum. This is managed by the Distributed Coordination Function (DCF), which operates using Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA).

The DCF protocol operates as a highly orchestrated, decentralized state machine. When a station has a packet to transmit, it first physically listens to the channel (Carrier Sense). If it detects any RF energy above a threshold, it assumes the channel is busy and defers. If the channel is completely silent, the station does not transmit immediately. It waits for a mandatory, mandatory silence period called the DCF Interframe Space (DIFS).

If the channel remains perfectly idle for the entire duration of the DIFS, the station enters the Collision Avoidance phase. It generates a random number and initiates a Random Exponential Backoff timer. This timer acts as a tie-breaker; if multiple stations were waiting for the channel to clear, the random timer ensures they don’t all transmit the millisecond the DIFS expires. The timer only counts down while the physical channel is idle; if someone else transmits, the timer freezes.

When the backoff timer finally hits zero, the station transmits its packet. Upon receiving the packet intact, the destination station waits for a much shorter silence period called the Short Interframe Space (SIFS), and instantly transmits an Acknowledgement (ACK) packet. The SIFS is deliberately designed to be strictly shorter than the DIFS. This physical timing guarantees that the receiving station can fire its ACK before any other station’s DIFS timer expires, granting the ACK absolute priority over any new data transmissions. If the sender does not receive the ACK, it assumes a collision occurred in the air, instantly doubles its maximum random backoff window (exponential backoff), and restarts the entire agonizing process.