Future Internet: Architectures and Protocols

Marie-José Montpetit, Serge Fdida, Jianping WangIEEE Communications Magazine 2019 · OpenReview

Abstract

The articles in this special section focus on future Internet applications. In the 50 years of its existence, the Internet has moved from an academic network to the supporting lifeline of the 21st century. The Internet has not only linked billions of users to one another, but revolutionized commerce, newspapers, television, and transportation. The relentless pace of technological growth is generating opportunities in all sectors of our life: entertainment, automotive, healthcare, education, and personal communications. The articles presented here provide insights on where the Internet is moving to in terms of architectures and technologies.

Introduction

The Internet is currently a victim of its own success. Designed over 50 years ago as an academic tool, its foundational client-server model is struggling to keep up with the demands of the 21st century. We are moving from a world of simple web browsing to a "network continuum" of billions of IoT sensors, autonomous vehicles, and data-hungry mobile devices. This transition necessitates a shift away from aging architectures toward a decentralized, "softwarized" model.

The Problem: Scaling the Client-Server Model

The traditional approach—relying on remote server farms for computation and storage—is no longer sustainable for modern applications. It creates latency bottlenecks and leads to inefficient resource use. The primary challenge is transitioning the Internet’s architecture to support "softwarization" (moving network functions from hardware to software) and "democratization" (pushing intelligence and control to the network edge).

Key Methodologies for the Future Internet

To address these challenges, several novel architectural approaches have been proposed:

  • Dispersed Computing: An evolution of edge computing that leverages geographically distributed resources to meet strict Quality of Service (QoS) requirements.
  • IS-Fog Framework: This combines Information-Centric Networking (ICN) with "social awareness" to better organize and orchestrate fog computing deployments based on human interaction patterns.
  • Cognitive Network Function Virtualization (NFV): A method for ICNs that uses the "cognition" of user content to enable on-demand configuration of storage and computation at the edge.
  • Thread/OpenThread: A novel multihop architecture designed specifically for low-power wireless IoT devices that lack high-power sources.
  • Software-Defined Quantum Key Distribution (QKD): Integrating quantum mechanics into software-defined networks to provide a new layer of data protection.

Key Insights and Results

The Network Continuum

The boundary between the "cloud" (server farms) and the "edge" (user devices) is disappearing. In this "network continuum," data filtering and processing happen wherever it is most efficient. This is supported by the rise of programmable hardware, which allows for fast, affordable data filtering at the source, reducing the need to send unnecessary data to the cloud.

From "Where" to "What"

There is a significant push toward Named Data Networking (NDN). Instead of the Internet asking "Where is this server?" (IP-based), it asks "What is this content?" (Content-centric). This shift has moved from theory to practice, particularly in securing smart home environments through specialized sign-on protocols.

Improved Agility and Efficiency

New architectures for the "Internet of Vehicles" (IoV) demonstrate improved agility and scalability by integrating SDN, NFV, and fog computing. Furthermore, data-driven intelligent networks are turning the "deluge of data" into a design advantage, improving resource utilization across the board.

Limitations and Future Outlook

While these technologies are promising, several hurdles remain:

  • Orchestration Complexity: Managing a massive, dispersed fog deployment requires more research into automated control systems.
  • Power Constraints: The IoT still struggles with power management; the lack of high-power sources in remote sensors remains a significant bottleneck.
  • Evolution vs. Revolution: A fundamental question remains: Will the current Internet continue to evolve incrementally, or will it eventually need to be replaced by a completely new architecture? The current research provides vital insights, but the final answer to this existential question is still being written.

Summary updated

Abstract

The articles in this special section focus on future Internet applications. n the 50 years of its existence, the Internet has moved from an academic network to the supporting lifeline of the 21st century. The Internet has not only linked billions of users to one another, but revolutionized commerce, newspapers, television, and transportation. The relentless pace of technological growth is generating opportunities in all sectors of our life: entertainment, automotive, healthcare, education, and personal communications. The articles presented here provide insights on where the Internet is moving to in terms of architectures and technologies.
BibTeX
@article{montpetit2019future,
  author = {Marie-José Montpetit and Serge Fdida and Jianping Wang},
  title = {{Future Internet: Architectures and Protocols}},
  journal = {IEEE Communications Magazine},
  year = {2019},
  url = {https://openreview.net/forum?id=6HP-Fp4C-el},
}
Figure