A Novel Algorithm for Detection of Elephant Flows: Landmark-LRU with Recycle

Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 235)


Along the trend of ever increasing Internet traffic, flows that occupy network bandwidth heavily over long period of time become serious threats. Elephant Flows, the nickname for these flows, have critical impact to network performance in terms of both delivery time and fairness. They raise complicated issues for management of Internet traffic and services. Identifying these flows efficiently in core network has been an important problem. The Least Recently Used (LRU) and the Landmark LRU schemes are two of the most efficient methods exist. We propose a cache update algorithm, to further improve the Landmark LRU scheme. The proposed scheme improves the accuracy of detecting elephant flow while maintaining efficiency of Landmark LRU. We verify our algorithm by analysis and simulation on wireless network traces captured on university environments.


Elephant flows Traffic measurements Algorithm QoS (quality of service) 


  1. 1.
    Mori T, Uchida M, Kawahara R, Pan J, Goto S (2004) Identifying elephant flows through periodically sampled packets. In: Proceedings of the IMC’04, 4th ACM SIGCOMM conference on Internet measurement, pp 115–120Google Scholar
  2. 2.
    Smitah A, Kim I, Reddy ALN (2001) Identifying long-term high-bandwidth flows at a router. In: Proceedings of the high performance computing, pp 361–371Google Scholar
  3. 3.
    Che L, Qui B (2006) Landmark LRU: an efficient scheme for the detection of elephant flows at internet routers. Commun Lett, IEEE 10(7):567–569CrossRefGoogle Scholar
  4. 4.
    Estan C, Varghese G (2002) New directions in traffic measurement and accounting. In: Proceedings of the ACM SIGCOMM’02, pp 323–336Google Scholar
  5. 5.
    Achanta PG, Reddy ALN (2003) Design, analysis and experience of a partial state router. In: IEEE INFOCOMGoogle Scholar
  6. 6.
    Che L, Qui B (2005) Traffic measurement for large internet flows using multi-stage landmark LRU Cache. TENCON 2005 IEEE Reg 10:1–6Google Scholar
  7. 7.

Copyright information

© Springer Science+Business Media Dordrecht 2013

Authors and Affiliations

  1. 1.Department of Computer ScienceSangmyung UniversitySeoulKorea

Personalised recommendations