QPM: Phonetic aware P2P search
Md. Rakibul Haque, Reaz Ahmed, Raouf Boutaba · 2009
Efficient discovery of information based on partially specified and misspelled query keywords is a challenging problem in large scale peer-to-peer (P2P) networks. This paper presents QPM, a P2P search mechanism for efficient information retrieval with misspelled and partial keywords. QPM uses the double metaphone algorithm to phonetically match misspelled query keywords with advertised keywords. For achieving bandwidth efficiency and similarity matching, QPM incorporates second order Reed-Muller code within the Plexus protocol having a logarithmic routing efficiency on overlay network size. QPM supports large scale networks and achieves better resilience to peer failure by maintaining redundant routing paths and by systematically placing index replicas. The concept presented in this paper is supported by necessary experimental evaluation.