Fast Failure Recovery Using Multi-threading in BGP

  • Gao Lei
  • Lai Mingche
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 181)


Since Border Gateway Protocol presents poor path diversity at the router level, it does not allow for a fast recovery and always produces a great amount of BGP churn upon session failures, making inter-domain routing system facing reliability challenge. In this paper, we propose a failure recovery scheme implemented on threaded BGP architecture to quickly react and recover from session failures. In our scheme, a failure recovery thread is designed to work with multiple BGP session threads in parallel, and it employs BFD protocol to quickly detect link failures and generate backup routes by our presented dynamic route generation algorithm in case of session failure occurrence. The failure recovery thread will notify BGP session threads to establish backup paths with related nexthops along the paths, and it will not advertise route change to all neighbors as traditional BGP does until the notification timer expires, thus effectively reducing BGP churn and improving the route stability. Finally, simulation results show that the execution time of dynamic route generation algorithm is only 160ms averagely, the increased number of update messages relative to BGP is 16.4% and 48.8% at most under two experiment scenarios, and the duration of packet loss reaches 0.82s and 2.63s on average compared to 6.5s of BGP, effectively improving the reliability of BGP.


Failure recovery TBGP Multi-threading Reliability 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Bonaventure, O., Filsfils, C., Francois, P.: Achieving sub-50 milliseconds recovery upon bgp peering link failure. IEEE/ACM Transactions on Networking 15(5), 1123–1135 (2007)CrossRefGoogle Scholar
  2. 2.
    Schrieck, V.V., Francois, P., Bonaventure, O.: BGP add- paths: the scaling/performance tradeoffs. IEEE Journal on Selected Areas in Communications 28(8), 1299–1307 (2010)CrossRefGoogle Scholar
  3. 3.
    Wei, Z., Wang, F.: Achieving resilient routing through redistributing routing protocols. In: ICC, pp.1–5 (2011)Google Scholar
  4. 4.
    Xiao, L., He, G., Nahrstedt, K.: BGP Session Lifetime Modeling in Congested Networks. Journal of Computer Networks 50(17), 1–12 (2006)Google Scholar
  5. 5.
    Wang, L., Saranu, M., Gottlieb, J.M., Pei, D.: Understanding bgp session failures in a large isp. In: IEEE INFOCOM 2007, pp. 348–356 (2007)Google Scholar
  6. 6.
    Sahoo, A., Kant, K., Mohapatra, P.: Characterization of BGP Recovery Time under Large- Scale Failures. In: IEEE ICC 2006, pp. 949–954 (2006)Google Scholar
  7. 7.
    Wang, H., Wang, C., Cai, S.: Enhance Internet Routing Availability with Multipath Interdomain Routing. In: IEEE ICCDA 2010, vol. 5, pp. 428–433 (2010)Google Scholar
  8. 8.
    Hu, C., Chen, K., Chen, Y., Liu, B.: Evaluating Potential Routing Diversity for Internet Failure Recovery. In: IEEE INFOCOM 2010 (2010)Google Scholar
  9. 9.
    Wang, N., Guo, Y., Ho, K., et al.: Fast Network Failure Recovery Using Multiple BGP Routing Planes. In: IEEE GLOBECOM 2009 (2009)Google Scholar
  10. 10.
    Ragha, L.L., Chag, K.V.: Multiple Route Selector BGP (MRS-BGP). In: IEEE ICWET 2010, pp. 304–308 (2010)Google Scholar
  11. 11.
    Dai, B., He, J., Wang, H., et al.: iRoute: A Scalable Inter-domain Multi-path Routing Framework for Multimedia Transmission. In: IEEE ICMT, pp. 4966–4969 (2011)Google Scholar
  12. 12.
    Ma, H., Zhang, J., Guo, Y., He, L.: Scalable Resilient BGP – Fast Recovery from Transient Inter-Domain Link Failures. In: IEEE IITA, pp. 980–984 (2008)Google Scholar
  13. 13.
    Watari, M., Hei, Y., Ano, S., Yamazaki, K.: OSPF-based Fast Reroute for BGP Link Failures. In: IEEE GLOBECOM, pp. 1–7 (2009)Google Scholar
  14. 14.
    Pei, D., der Merwe, J.V.: BGP convergence in virtual private networks. Presented at the Internet Measurement Conf., Rio de Janeiro, Brazil (October 2006)Google Scholar
  15. 15.
    Labovitz, C., Ahuja, et al.: Delayed internet routing convergence. In: Proc. ACM SIGCOMM 2000, Stockholm, Sweden, August 28-September 1, pp. 175–187 (2000)Google Scholar
  16. 16.
    Zhao, X., Pei, D., Massey, D., Zhang, L.: A study on the routing convergence of Latin American networks. In: Proc. LANC 2003, LaPaz, Bolivia, October 4-5, pp. 35–43 (2003)Google Scholar
  17. 17.
    Kushman, N., Kandula, S., Katabi, D., Maggs, B.M.: RBGP: Staying connected in a connected world. In: Proc. NSDI, pp. 341–354 (2007)Google Scholar
  18. 18.
    Wang, F., Gao, L.: A backup route aware routing protocol –Fast recovery from transient routing failures. In: Proc. INFOCOM 2008 (2008)Google Scholar
  19. 19.
    Ganichev, I., Dai, B., Godfrey, P.B., Shenker, S.: Yamr: Yet another multipath routing protocol. EECS Department, Tech. Rep. UCB/EECS-2009-150 (October 2009)Google Scholar
  20. 20.
    Wang, K., Gao, L., Lai, M.: A scalable multithreaded BGP architecture for next generation router. In: IEEE EMC 2011, pp. 72–77 (2011)Google Scholar
  21. 21.
    Katz, D., Ward, D.: Bidirectional forwarding detection. RFC 5880 (July 2010)Google Scholar
  22. 22.
    Delaunois, C.: Ghitle: Generator of Hierarchical Internet Topologies using LEvels,

Copyright information

© Springer Science+Business Media Dordrecht 2012

Authors and Affiliations

  • Gao Lei
    • 1
  • Lai Mingche
    • 1
  1. 1.Department of ComputerNational University of Defense TechnologyChangshaChina

Personalised recommendations