BalanceDN: Load-Balancing Allocation of Interest for Fast Discovery in Content Centric Networks

TL;DR

BalanceDN tackles slow, congested content search in Named Data Networking by introducing a DNS-like cluster of NDN resolvers that use load-balanced hashing to place and locate content. By avoiding blind flooding and enabling direct lookup, BalanceDN achieves faster discovery and reduces network resource consumption. The authors implement and evaluate the approach in ndnSIM across multiple topologies and content distributions, demonstrating significant reductions in hops and bandwidth versus traditional NDN routing. The work offers a scalable, non-flooding content-discovery mechanism with practical potential for deployment in future content-centric networks.

Abstract

In Named Data Networking (NDN), data is identified by unique names instead of IP addresses, and routers use the names of the content to forward Interest packets towards the producers of the requested content. However, the current content search mechanism in NDN is complex and slow. This mechanism not only creates congestion but also hinders practical deployment due to its slowness and cumbersome nature. To address this issue, we propose a methodology, called BalanceDN, that distributes content through the network such that sought content can be found quickly. BalanceDN uses a distributed allocation of resolvers as those used by the domain name system but differs in how content is distributed. Our approach avoids flooding the network with pending interest requests and also eliminates the need for blind search when the location of content is unknown. We tested our approach on ndnSIM; a simulation platform for NDN. The results show that the proposed routing scheme utilizes far fewer network resources compared to the NDN network when retrieving content. The proposed scheme accomplishes this performance gain by leveraging a load-balanced hashing mechanism to distribute and locate the name of the content on the distributed nameserver lookup service nodes.

Figures (11)

• Figure 1: Example of the flow of uploading data through the proposed NDN routing scheme.
• Figure 2: Architecture of the host-resolver.
• Figure 3: Pseudocode of the content discovery mechanism.
• Figure 4: Simple test network with 3 hops as the largest distance.
• Figure 5: Example of used discovery routes (green color) on the test network.