[Paper Review] BFR: a Bloom Filter-based Routing Approach for Information-Centric Networks

Ali Marandi, Torsten Braun+2 more|Open Access at Essex (University of Essex)|Feb 1, 2017

Caching and Content Delivery16 references4 citations

TL;DR

This paper proposes BFR, a Bloom Filter-based, fully distributed routing scheme for Named Data Networking (NDN) that enables efficient content discovery without topology knowledge. By having origin servers advertise content using compact Bloom filters, BFR reduces signaling overhead and outperforms flooding and shortest-path routing in average round-trip delay and robustness to topology changes, while maintaining low false positive impact on routing accuracy.

ABSTRACT

Locating the demanded content is one of the major challenges in Information-Centric Networking (ICN). This process is known as content discovery. To facilitate content discovery, in this paper we focus on Named Data Networking (NDN) and propose a novel routing scheme for content discovery, called Bloom Filter-based Routing (BFR), which is fully distributed, content oriented, and topology agnostic at the intra-domain level. In BFR, origin servers advertise their content objects using Bloom filters. We compare the performance of the proposed BFR with flooding and shortest path content discovery approaches. BFR outperforms its counterparts in terms of the average round-trip delay, while it is shown to be very robust to false positive reports from Bloom filters. Also, BFR is much more robust than shortest path routing to topology changes. BFR strongly outperforms flooding and performs almost equal with shortest path routing with respect to the normalized communication costs for data retrieval and total communication overhead for forwarding Interests. All the three approaches achieve similar mean hit distance. The signalling overhead for content advertisement in BFR is much lower than the signalling overhead for calculating shortest paths in the shortest path approach. Finally, BFR requires small storage overhead for maintaining content advertisements.

Motivation & Objective

• To address the challenge of efficient content discovery in Information-Centric Networking (ICN), particularly in Named Data Networking (NDN), where traditional routing methods incur high signaling or latency overhead.
• To design a fully distributed, topology-agnostic, content-oriented routing scheme that reduces signaling and storage costs for FIB population.
• To evaluate whether Bloom filter-based content advertisement can outperform flooding and shortest-path routing in terms of communication cost, delay, and robustness.
• To analyze the impact of false positive errors in Bloom filters on routing accuracy and performance in real-world scenarios.

Proposed method

• Origin servers advertise their content objects using Bloom filters (BFs), which compactly represent sets of content names with minimal storage and signaling overhead.
• Each NDN node receives and stores BF advertisements from all paths toward the origin server, enabling FIB population without topology information.
• Interests are forwarded using the multicast strategy, propagating in parallel over all paths toward the origin server, enabling early satisfaction from in-network caches.
• The scheme relies on Longest Prefix Matching (LPM) in the FIB for Interest forwarding, using BF-based content advertisements to populate FIB entries.
• False positive errors are modeled and evaluated to assess their impact on routing correctness and performance.
• Performance is benchmarked against flooding and shortest-path routing using metrics like round-trip delay, communication cost, hit distance, and signaling overhead.

Experimental results

Research questions

• RQ1 Can a Bloom filter-based content advertisement scheme reduce signaling and storage overhead in NDN while maintaining efficient content discovery?
• RQ2 How does BFR compare to flooding and shortest-path routing in terms of average round-trip delay and communication cost?
• RQ3 To what extent do false positive errors in Bloom filters affect routing accuracy and performance in BFR?
• RQ4 How robust is BFR to dynamic network topology changes compared to shortest-path routing?
• RQ5 What is the impact of content popularity (α) on false positive error effects in BFR?

Key findings

• BFR outperforms flooding and shortest-path routing in average round-trip delay, achieving lower latency for content retrieval.
• BFR reduces signaling overhead by approximately three times compared to shortest-path routing, which requires frequent path computation updates.
• The normalized communication cost for data retrieval and total communication overhead in BFR is nearly equal to that of shortest-path routing, while being significantly lower than flooding.
• All three schemes—BFR, flooding, and shortest path—achieve similar mean hit distances, indicating comparable routing efficiency in terms of hop count.
• False positive errors have minimal practical impact: only 1.73% of Interests were routed toward wrong origin servers under the highest false positive probability (28.3%), with all Interests still successfully satisfied.
• BFR is highly robust to topology changes, significantly outperforming shortest-path routing, which requires constant path recomputation.