TRMAC: A Time-Reversal-based MAC Protocol for Wireless Networks within Computing Packages Journal: Nano Communication Networks
Ama Bandara, Abhijit Das, Fátima Rodríguez-Galán, Eduard Alarcón, Sergi Abadal · 2025
Read TRMAC: A Time-Reversal-based MAC Protocol for Wireless Networks within Computing Packages Journal: Nano Communication Networks by Ama Bandara, Abhijit Das, Fátima Rodríguez-Galán, Eduard Alarcón, Sergi Abadal (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-19487.txt.
Introduction
As chips pack hundreds of cores together, wires between them strain, and wireless links inside the package look attractive. Ama Bandara and co-authors present TRMAC in a 2025 paper, a protocol that decides which core may transmit at each moment. It exploits time reversal, a trick that focuses a signal on one spot.
Because on-chip channels barely change, the protocol measures them once and uses energy thresholds to coordinate access. Many transmissions can then share one channel in parallel, with no central referee. Simulations at both the signal and system level showed low delay and high throughput that scaled to hundreds of cores. The authors pitch the design for chiplet systems where reuse of space matters most.
Catalog record
| CREATOR | Ama Bandara, Abhijit Das, Fátima Rodríguez-Galán, Eduard Alarcón, Sergi Abadal |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-19487
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
TRMAC: A Time-Reversal-based MAC Protocol for Wireless Networks within Computing Packages Journal: Nano Communication Networks Ama Bandara, Abhijit Das, Fátima Rodríguez-Galán, Eduard Alarcón, Sergi Abadal Abstract As chiplet-based integration and many-core architectures become the norm in high-performance computing, on-chip wireless communication has emerged as a compelling alternative to traditional interconnects. However, scalable Medium Access Control (MAC) remains a fundamental challenge, particularly under dense traffic and limited spectral resources. This paper presents TRMAC , a novel cross-layer MAC protocol that exploits the spatial focusing capability of Time Reversal (TR) to enable multiple parallel transmissions over a shared frequency channel. By leveraging the quasi-deterministic nature of on-chip wireless channels, TRMAC pre-characterizes channel impulse responses to coordinate access using energy-based thresholds, eliminating the need for orthogonal resource allocation or centralized arbitration. Through detailed physical-layer simulation and system-level evaluation on diverse traffic, TRMAC demonstrates comparable or superior performance to existing multi-channel MAC protocols, achieving low latency, high throughput, and strong scalability across hundreds of cores. TRMAC provides a low-complexity, high-efficiency solution for future Wireless Networks-on-Chip (WNoCs), particularly in chiplet-based systems where spatial reuse and modularity are critical. With simulations we prove that TRMAC can be utilized for parallel transmissions with a single frequency channel with a similar throughput and latency as in using multiple frequency bands omitting the need for complex transceivers. This work establishes a new design direction for MAC protocols that are tightly integrated with the underlying channel physics to meet the demands of next-generation computing platforms. 1 Introduction Recent advances in High-Performance Computing (HPC), driven by the proliferation of CPUs, GPUs, and domain-specific accelerators, have led to the integration of many cores within a single computing package, and more recently, the disaggregation of functionality across chiplets interconnected via high-density packaging [ 1 ] . To fully harness the computational capabilities of such architectures, whether monolithic or chiplet-based, parallel processing has become indispensable, enabling the concurrent execution of tasks and efficient handling of large-scale data. However, this parallelism introduces significant challenges in inter-core and inter-chiplet communication, particularly in maintaining synchronization and ensuring coherent data sharing [ 2 , 3 , 4 ] . Consequently, precise and efficient communication mechanisms are critical to coordinating tasks and maximizing system performance [ 5 , 6 , 3 ] . Wireless Networks-on-Chip (WNoCs) have recently emerged as a promising complement to traditional wired NoC architectures, offering a path to alleviate the growing challenges of latency and power dissipation in massively multi-core and System-on-Chip (SoC)/System-in-Package (SiP) designs [ 5 , 7 ] . In these systems, communication is facilitated through an interposer-based environment, where antennas integrated on each core or chiplet establish wireless links within the package. Compared to conventional wired interconnects, wireless links offer substantial advantages, including single-hop data transfers, inherent broadcasting capabilities, and dynamic reconfigurability [ 7 , 8 ] . However, to be viable in high-performance environments, WNoCs must meet stringent requirements: data rates exceeding 100 Gb/s, latencies below 10 ns, energy consumption below 1 pJ/bit, and Bit Error Rates (BERs) below 10^{-10} , ensuring performance, efficiency and reliability [ 9 ] . While the use of wireless communication within chip packages offers compelling benefits, it also presents unique challenges. The confined and reflective nature of the chip environment leads to significant electromagnetic energy dispersion, resulting in high delay spread and Inter-Symbol Interference (ISI) that degrades communication at high data rates. Furthermore, simultaneous non-orthogonal transmissions over shared bandwidths introduce severe Co-Channel Interference (CCI), particularly in parallel wireless links. To ensure robust and efficient data exchange, it is crucial to adopt Medium Access Control (MAC) or multiple access schemes that begin at the physical layer to mitigate interference and avoid collisions. Figure 1 : WNoC in an interposer package with two parallel communication links. Data packets are transmitted using TR filters, which suppress CCI at neighboring nodes. Spatial channels are coordinated via TRMAC. Fortunately, the quasi-deterministic nature of the on-chip wireless channel, enabled by fixed package dimensions and known material properties, allows for accurate pre-characterization of the Channel Impulse Response (CIR). This predictability makes Time Reversal (TR) an attractive technique for improving communication robustness [ 10 ] . By exploiting the spatial and temporal invariance of the channel, TR focuses energy both in space and time at the intended receiver as shown in Figure 1 , thereby significantly reducing ISI and enhancing signal fidelity [ 10 , 11 , 12 ] . TR filtering employs the time-reversed version of the CIR to act as a spatial matched filter that concentrates energy at the target receiver. In reverberant environments with rich multipath effects, the uniqueness of each CIR results in stronger signal focusing and reduced interference. Additionally, TR minimizes CCI by exploiting the orthogonality of spatial channels between different transmitter-receiver pairs. This allows multiple parallel transmissions to occur simultaneously in both time and frequency domains, provided that the number of concurrent transmissions, which we refer to the Number of Parallel Transmissions (NPT), is carefully bounded. Although TR suppresses CCI, it does not eliminate it entirely, necessitating well-defined NPT thresholds, which remain stable due to the time-invariant nature of the channel. Despite enabling spatial multiplexing, TR does not inherently prevent transmission collisions. Moreover, TR-based communication is susceptible to the deafness problem, wherein transmitters and receivers lack awareness of ongoing transmissions due to the absence of explicit coordination, beyond the NPT constraint. To address these limitations, this work introduces TRMAC: a cross-layer, adaptive MAC protocol specifically designed to manage TR-precoded transmissions in chip-scale wireless environments. Existing WNoC MAC protocols typically rely on resource-sharing schemes such as frequency-division, time-division, or code-division multiple access (FDMA, TDMA, or CDMA) to mitigate interference in multi-link environments [ 13 , 14 ] . Some approaches adopt hierarchical hybrid wired-wireless architectures, but tend to emphasize overall system design while overlooking the effects of wireless channel impairments [ 15 , 16 , 17 , 18 , 4 ] . Table 1 summarizes recent MAC protocols in terms of architecture, multiplexing, scalability, and traffic awareness capabilities. Many of these schemes are tailored to specific architectures or applications, and do not scale well with increasing core counts or dynamic traffic. Additionally, they often incur extra overhead in area and power due to complex transceiver designs or control mechanisms, and may introduce latency through intricate scheduling or arbitration. Table 1 : Summary of the design traits of existing protocols Work Architecture Multiplexing Scalability Traffic Awareness Shamim et al. [ 4 ] Non-static N/A High Low Mansoor et al. [ 19 ] Non-static TDMA Medium High Franques et al. [ 20 ] Non-static N/A High High Mestres et al. [ 21 ] Non-static N/A High Low Vijayakumaran et al. [ 13 ] Static [ 22 ] CDMA High High Matolak et al. [ 14 ] Static FDMA, TDMA High Medium DiTomaso et al. [ 17 ] Static TDMA High High TRMAC (this work) Non-Static Spatial High High Despite significant progress, none of the existing MAC protocols account for communication over multiple spatial channels as enabled by TR. Most designs either assume global knowledge of ongoing transmissions or rely on orthogonal resource allocation to support parallelism. In contrast, TRMAC leverages the spatial diversity offered by TR to support simultaneous transmissions in both time and frequency. It adopts a cross-layer approach,