TL;DR
This paper proposes a novel approach to utilize idle user bandwidth in Information-Centric Networking (ICN) by treating it as a network service. By integrating blockchain technology, they create an incentive mechanism that encourages users to contribute their bandwidth, leading to significant improvements in network throughput and reduced packet loss.
What Problem Does This Paper Address?
The internet is experiencing explosive growth in traffic, driven by bandwidth-intensive applications. This has led users to purchase more network bandwidth, yet a significant portion of this bandwidth remains idle for most of the time, representing a substantial waste of resources. Traditional IP networks struggle to efficiently utilize this idle bandwidth, often relying on application-layer solutions that introduce scalability issues and overhead. Information-Centric Networking (ICN), with its identifier-locator separation, offers a promising network-layer approach to aggregate idle bandwidth. However, a key challenge is incentivizing users to contribute their resources reliably and securely.
Methods
The core of the proposed solution lies in a fusion architecture of ICN and blockchain.
- ICN Integration: The authors leverage ICN's ability to decouple identifiers from locators. They propose treating user bandwidth contribution as a service, assigning it a unique User Bandwidth Sharing Service ID (UBSID). When network congestion occurs or ordinary ICN routers are overloaded, traffic can be rerouted to these user nodes for forwarding via an ID/NA cooperative routing mechanism. This approach aims to improve ICN's scalability and underlying resource utilization.
- Blockchain-Based Incentive Mechanism: To encourage user participation, a blockchain-based bandwidth trading market is established. They introduce a Proof-of-Network-Contribution (PoNC) consensus algorithm, where users can "mine" by forwarding packets. This mechanism incentivizes users to keep their devices online and contribute bandwidth.
- Node Selection Strategy: A crucial component is the selection of optimal user nodes for relaying traffic. This strategy considers factors like bandwidth, latency, and packet loss rate, using a formula that combines these metrics. The Name Mapping Resolution System (NMRS) plays a key role in managing these user nodes and their network addresses.
Datasets
The experiments were conducted using real-world network topologies from the Internet Topology Zoo dataset, specifically the INTEGRA and AGIS topologies. These are used to simulate the underlying physical network infrastructure.
Key Insights and Ideas
- ICN as a Network-Layer Solution for Idle Bandwidth: Moving beyond application-layer solutions, ICN's architecture is well-suited for aggregating idle bandwidth at the network layer, offering better scalability and integration.
- Blockchain for Trust and Incentives: Blockchain provides a secure and transparent platform to manage contributions and reward users, overcoming the trust issues inherent in decentralized resource sharing.
- "Mining" for Bandwidth Contribution: The concept of users "mining" by forwarding packets is an intuitive and effective way to incentivize continuous participation and resource contribution.
- Dynamic Node Selection is Crucial: Simply aggregating bandwidth isn't enough; intelligently selecting the best-performing user nodes based on real-time network conditions is key to achieving performance gains.
Key Results
The experimental results demonstrate significant improvements:
- Throughput: A maximum throughput improvement of 19.4% was observed when utilizing user idle bandwidth, particularly in high bandwidth demand scenarios.
- Packet Loss Rate: The proposed solution effectively reduces packet loss rate, especially at high sending rates, by offloading traffic from congested ordinary ICN nodes to user nodes.
- Latency and Bandwidth Utilization: Compared to existing methods like NKN, the ICN-based approach offers more stable end-to-end latency and a maximum bandwidth utilization improvement of 13.7%. The NKN approach, relying on DHT, showed unstable latency due to its routing algorithm's limitations.
Limitations and Future Work
The authors acknowledge that the calculation of network contribution needs further study, particularly regarding measurement security and the potential for cheating. They also note that existing distributed measurement methods can incur significant computational overhead.
Future work will focus on:
- Developing more efficient and secure methods for measuring network contribution.
- Exploring how to leverage ICN's inherent advantages to design more efficient proof algorithms for network contribution.
Summary updated