Rank-Based Modeling for Universal Packets Compression in Multi-Modal Communications

TL;DR

This work addresses the challenge of efficiently compressing heterogeneous network traffic in multi-modal settings by introducing ByteTrans, a byte-level Transformer that predicts the next byte and encodes the resulting rank sequence with entropy coding. Grounded in information-theoretic principles, ByteTrans transforms byte streams into highly predictable rank patterns, enabling lossless compression that substantially reduces data size while remaining adaptable across devices. The paper demonstrates that ByteTrans achieves compression ratios around 46–50% and outperforms the baseline zlib, with validated deployments on edge devices and servers. The approach promises practical impact for IoT and sensor networks by lowering bandwidth requirements and enabling scalable, universal packet compression without modality-specific predictors, while also offering a path toward dynamic device-aware model adaptation.

Abstract

The rapid increase in networked systems and data transmission requires advanced data compression solutions to optimize bandwidth utilization and enhance network performance. This study introduces a novel byte-level predictive model using Transformer architecture, capable of handling the redundancy and diversity of data types in network traffic as byte sequences. Unlike traditional methods that require separate compressors for different data types, this unified approach sets new benchmarks and simplifies predictive modeling across various data modalities such as video, audio, images, and text, by processing them at the byte level. This is achieved by predicting subsequent byte probability distributions, encoding them into a sparse rank sequence using lossless entropy coding, and significantly reducing both data size and entropy. Experimental results show that our model achieves compression ratios below 50%, while offering models of various sizes tailored for different communication devices. Additionally, we successfully deploy these models on a range of edge devices and servers, demonstrating their practical applicability and effectiveness in real-world network scenarios. This approach significantly enhances data throughput and reduces bandwidth demands, making it particularly valuable in resource-constrained environments like the Internet of Things sensor networks.

Figures (9)

• Figure 1: The overview of ByteTrans framework.
• Figure 2: (a) Architecture of multi-head attention. (b) Architecture of byte Transformer decoder.
• Figure 3: Packets length distribution (KDE).
• Figure 4: Performance of the predictive models.
• Figure 5: compression ratio under various conditions.