Throughput Gain (TG) in Communication Networks

• Throughput Gain (TG) is a metric that compares the data capacity of enhanced systems to a baseline, expressed as a ratio or percentage.
• It is widely used in optical, wireless, cognitive, and distributed networks to assess the impact of advanced amplification, beamforming, and resource allocation techniques.
• TG evaluation combines theoretical modeling with simulation metrics, highlighting sensitivities to parameters like SNR margins, power control, and system density.

Throughput gain (TG) is a comparative performance metric used to quantify the relative increase in data-carrying capacity or throughput of a communication system, protocol, or network architecture when transitioning from a reference (baseline) design to an enhanced, optimized, or differently configured system. TG is typically defined as the normalized difference between the throughputs of the two systems, with the result expressed either as a dimensionless ratio or as a percentage. Across diverse research domains including optical networks, wireless communications, cognitive radio, and distributed networking, TG provides a rigorous method to evaluate technological advances, architectural optimizations, and system-level interventions.

1. Mathematical Formalism and Definitions

For a baseline system with throughput $T_0$ and an improved system with throughput $T_1$, throughput gain is generally defined as

$\mathrm{TG} = \frac{T_1 - T_0}{T_0}$

yielding a dimensionless quantity (often multiplied by 100 for percent gain). The specific mapping of $T_1$ and $T_0$ depends on context:

• In optical fiber communication: $T_1$ and $T_0$ may refer to the end-to-end (Shannon) capacities (in bit/s) of networks with/without hybrid Raman amplification (Buglia et al., 2023).
• In wireless or mmWave systems: $T_1$ and $T_0$ correspond to achievable sum rates with and without intelligent surfaces, beamforming, or optimized resource allocation (Li et al., 2023).
• In network-wide resource allocation or cognitive radio: $T_1$ and $T_1$0 may denote spatial throughput or carried traffic under different access, sensing, or interference mitigation regimes (Banaei et al., 2011, Chen et al., 2021).
• In distributed random access: $T_1$1 and $T_1$2 can be the average packets/slot for schemes with multi-level random power control versus classical slotted ALOHA (Kumar et al., 2020).

This metric is always referenced to a rigorously computed baseline which may be a traditional configuration, an unoptimized mode of operation, or a theoretic maximum under certain constraints.

2. TG in Optical Networks and Physical Layer Design

In ultra-wideband optical fiber links, TG is used to quantify improvements due to advanced amplification, such as hybrid (distributed Raman + erbium-doped fiber amplifiers), over classic lumped amplification. Using the generalized GN-model under Gaussian modulation, the total end-to-end link capacity is

$T_1$3

where $T_1$4 is the channel count and $T_1$5 is the symbol rate per channel. TG is obtained as

$T_1$6

For a 10.5 THz, 117×57 km system employing PSO-optimized pump settings, Raman+EDFA amplification yields $T_1$7 Tbit/s versus $T_1$8 Tbit/s, giving $T_1$9 (Buglia et al., 2023). The gain is physically attributed to lower cumulative ASE noise and the ability to operate at lower launch powers, moderating nonlinear impairments.

In flexible optical networks, TG encapsulates the benefit of “just-enough” SNR margin and channel-spacing optimization over traditional excessive-margin baselines: $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$0 where $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$1 is throughput after joint provisioning/channel-spacing optimization and $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$2 with excessive SNR margin. Sustained TG up to 50% is observed at low loads, with gains persisting even in spectrum-constrained regimes (Chen et al., 2021).

3. TG in Wireless and IRS-Assisted Networks

In mmWave systems augmented with intelligent refracting surfaces (IRS), TG evaluates enhancements from programmable environmental reconfiguration. For IRS-assisted high-speed train links: $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$3 where $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$4 is the sum-rate with jointly optimized beamforming, IRS phasing, and power allocation, and $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$5 is the direct-link baseline. The sum-rate per TDMA frame is given by

$\mathrm{TG} = \frac{T_1 - T_0}{T_0}$6

Numerical studies demonstrate that TG ranges from $\mathrm{TG} = \frac{T_1 - T_0}{T_0}$710% for smaller IRS ($\mathrm{TG} = \frac{T_1 - T_0}{T_0}$8) up to 42% for large IRS arrays ($\mathrm{TG} = \frac{T_1 - T_0}{T_0}$9), and is also sensitive to train speed and cluster size (Li et al., 2023).

In THz/RF multi-hop routing, TG compares end-to-end throughput of THz and RF routes. Utilizing stochastic geometry,

$T_1$0

where $T_1$1 is the stepwise-optimal throughput for technology $T_1$2. Under feasible power/hop configurations, TG can exceed 10-20× at moderate distances and power, reflecting the bandwidth advantage of THz despite pronounced path loss (Lou et al., 9 Aug 2025).

4. TG in Cognitive Radio and Overlay Networks

In multi-tier cognitive radio with overlapping resource use, the metric of interest is often the asymptotic multiplexing gain (AMG), from which TG is directly derived. For large overlaid Poisson networks, the sum spatial throughput scales as $T_1$3. TG for tier $T_1$4 is defined as the ratio of its multiplexing gain under overlay operation $T_1$5 to the stand-alone gain $T_1$6: $T_1$7 Explicit expressions are available under different density regimes (e.g., $T_1$8 or $T_1$9) and are sensitive to parameters such as spectrum sensing radius, ALOHA access probability, and cross-interference range. TG$T_0$0 always, with proper design able to restore TG$T_0$1 for the primary via aggressive sensing or throttling secondary access, while the secondary can attain near-standalone scaling at low density (Banaei et al., 2011).

5. TG in Distributed Random Access and Cross-Layer Protocols

In distributed MAC protocols (ALOHA, IRSA), TG quantifies the throughput enhancement when multi-level random power control or non-orthogonal protocols are introduced: $T_0$2 where $T_0$3 is the mean successful packet decodings per slot (for load $T_0$4) and $T_0$5 is either the unit-throughput barrier or classical ALOHA peak throughput. With only two power levels,

$T_0$6

Simulations show $T_0$7 (ALOHA baseline) and up to 2× under IRSA with three power levels, demonstrating the feasibility of breaking the unit throughput barrier via physical-layer diversity and SIC (Kumar et al., 2020).

6. TG under Advanced PHY/MAC Protocols: AMC and HARQ

Throughput gain also serves as a primary metric when comparing advanced error-control strategies. For instance, layer-coded HARQ (L-HARQ) evaluated against conventional adaptive modulation and coding (AMC) plus hybrid ARQ, TG is defined as

$T_0$8

where $T_0$9 and $T_1$0 are average throughputs with L-HARQ and traditional HARQ, respectively. Relative gains of 14–32% (and SNR reductions of 1.5–3 dB at fixed efficiency) are reported for practical turbo-coded systems with moderate HARQ round counts ($T_1$1), especially in fast-fading or outdated-CSI regimes (Jabi et al., 2018).

7. Design Implications, Sensitivities, and Regime-Dependent Behavior

TG encapsulates critical physical and network-layer phenomena:

• In optical/fiber networks, TG arises from noise figure reduction and launch-power optimization; maximal when distributed (Raman) gain is concentrated to offset fiber loss, but subject to cubic nonlinear penalties if launch power is excessive (Buglia et al., 2023).
• In wireless and THz systems, TG is power-, distance-, and relay-density dependent. It can be highly sensitive to environmental constraints, as in IRS configuration or THz link budget.
• In cognitive overlays, TG quantifies the trade-off between primary protection and secondary utilization; adaptive spectrum sensing or access limitations can recover lost TG in vulnerable regimes (Banaei et al., 2011).
• In distributed random access, achieving high TG depends on power allocation strategy, degree distributions, and the effectiveness of SIC; theoretical upper bounds are approached with multi-level capture provisions (Kumar et al., 2020).
• In cross-layer design (AMC/HARQ), TG benefits are maximized when protocol parameters are tailored to channel time-variability and coding/retransmission are jointly adapted (Jabi et al., 2018).

The regime in which TG is evaluated—resource over- or under-provisioned, interference-limited, dense vs sparse topologies—directly determines the attainable gain, and often reflects inherent system-level constraints. As such, TG serves as a unifying comparative performance indicator in system optimization, architectural evaluation, and network design across communication technologies.