Alleviating the topology mismatch problem in distributed overlay networks: A survey

Vassilis Moustakas, Hüseyin Akcan, Mema Roussopoulos, Alex Delis

Research output: Contribution to journalArticlepeer-review


Peer-to-peer (P2P) systems have enjoyed immense attention and have been widely deployed on the Internet for well over a decade. They are often implemented via an overlay network abstraction atop the Internet's best-effort IP infrastructure. P2P systems support a plethora of desirable features to distributed applications including anonymity, high availability, robustness, load balancing, quality of service and scalability to name just a few. Unfortunately, inherent weaknesses of early deployments of P2P systems, prevented applications from leveraging the full potential of the paradigm. One major weakness, identified early on, is the topology mismatch problem between the overlay network and the underlying IP topology. This mismatch can impose an extraordinary amount of unnecessary stress on network resources and can adversely affect both the scalability and efficiency of the operating applications. In this paper, we survey over a decade's worth of research efforts aimed at alleviating the topology mismatch problem in both structured and unstructured P2P systems. We provide a fine-grained categorization of the suggested solutions by discussing their novelty, advantages and weaknesses. Finally, we offer an analysis as well as pictorial comparisons of the reviewed approaches since we aim to offer a comprehensive reference for developers, system architects and researchers in the field.

Original languageEnglish (US)
Pages (from-to)216-245
Number of pages30
JournalJournal of Systems and Software
StatePublished - Mar 1 2016


  • Overlay network
  • Topology awareness
  • Topology mismatch

ASJC Scopus subject areas

  • Software
  • Information Systems
  • Hardware and Architecture


Dive into the research topics of 'Alleviating the topology mismatch problem in distributed overlay networks: A survey'. Together they form a unique fingerprint.

Cite this