روش نوين توزيع بار ترافيک مبتنی بر BGP در مراکز داده توزيع‌شده جغرافيايي
محورهای موضوعی : مهندسی برق و کامپیوتر
امیر اسمعیلی
1
(دانشگاه صنعتی امیرکبیر)
بهادر بخشی سراسکانرود
2
(دانشگاه صنعتی امیرکبیر)
کلید واژه: توزیع بار ترافيکیمراکز داده توزيع‌شدهBGPanycastشبکه‌هاي نرم‌افزار محور,
چکیده مقاله :
مسایل مربوط به مراکز داده به عنوان زیرساخت ارائه سرویس‌های فناوری اطلاعات همواره حایز اهمیت بوده است. از سوی دیگر در سال‌های اخیر، بنا به دلایل متعددی از جمله تفاوت هزینه انرژی در نقاط مختلف، امنیت بالاتر و تأمین کیفیت سرویس، مراکز داده توزیع‌شده جغرافیایی مورد توجه قرار گرفته است که در آن مرکز داده متشکل از چندین خوشه در نقاط مختلف است. مسأله اصلی در این نوع مراکز داده، توزیع کارای بار ترافیکی بین خوشه‌ها است. با وجود این که روش‌های متعددی برای این منظور پیشنهاد شده است ولی به دلیل متمرکزبودن توزیع‌کننده، سرعت پایین واکنش به تغییرات بار ترافیکی و یا محدودبودن به یک تکنولوژی خاص، راه حل کارا، عملی و جامع این مسأله به شمار نمی‌روند. در این مقاله معماری جدیدی پیشنهاد شده که در آن برای توزیع بار ترافیکی، مفهوم مسیریابی anycast با تنظیمات مناسب BGP و در نظر گرفتن ملاحظات عملی پیاده‌سازی می‌شود. نتایج شبیه‌سازی بهبود قابل توجه این روش نسبت به روش‌های موجود را نشان می‌دهد.
Today, hosting services in geographically distributed data centers is very common among service provider companies, because of more efficiency of energy consumption, high availability of the system, and providing quality of service. Load distribution is the main issue in the geographical data centers. On the one hand, there are several architectures to distribute load between different clusters, e.g., central load balancer, DNS-based systems, and IGP based schemes; one the other hand, the optimum traffic load balancing between clusters is a very challengeable issue. The proposed solutions have different facilities to distribute incoming traffic; nevertheless, they are vulnerable in terms of propagation delay, centralized load balancer failure, and maintaining connections. In this paper, a new architecture based on BGP and Anycast routing protocols in SDN based data centers is proposed to distribute traffic loads between clusters. Simulation result shows improvement in comparison to the existing techniques.
[1] H. Goudarzi and M. Pedram, "Force-directed geographical load balancing and scheduling for batch jobs in distributed datacenters," in Proc. IEEE Int. Conf. on Cluster Computing, CLUSTER'13, 8 pp., Indianapolis, IN, USA, 23-27 Sept. 201 2013.
[2] -, About Data Centers, www.tia942.org/content/162/289/About_Data_Centers (Accessed on 11 May 2017).
[3] H. Goudarzi, S. Hatami, and M. Pedram, "Demand-side load scheduling incentivized by dynamic energy prices," in Proc. IEEE Int. Conf. on Smart Grid Communications, SmartGridComm'11, pp. 351-356, Brussels, Belgium, 17-20 Oct. 2011.
[4] H. Uppal and D. Brandon, OpenFlow Based Load Balancing, Networking Project Report, University of Washington, 2010.
[5] D. Kreutz, et al., "Software-defined networking: a comprehensive survey," Proceedings of the IEEE, vol. 103, no. 1, pp. 14-76, Jan. 2015.
[6] N. Gude, et al., "NOX: towards an operating system for networks," ACM SIGCOMM Computer Communication Review, vol. 38, no. 3, pp. 105-110, Jul. 2008.
[7] A. Serra, D. Gaiti, G. Barroso, and J. Boudy, "Assuring QoS differentiation and load balancing on web servers clusters," in Proc. IEEE Conf. on Control Applications, CCA'05, pp. 885-890, Toronto, Canada, 28-31 Aug. 2005.
[8] H. Zhang and X. Guo, "SDN-based load balancing strategy for server cluster," in Proc. IEEE 3rd Int. Conf. on Cloud Computing and Intelligence Systems, pp. 662-667, Shenzhen, China, 27-29 Nov. 2014.
[9] F. Semchedine, L. Bouallouche-Medjkoune, O. Sayeh, S. Ayoub, and D. Aïssani, "DNS-based load balancing with cache for geographically distributed web server systems," in Proc. Global Summit on Computer & Information Technology, GSCIT'14, 6 pp., Sousse, Tunisia, 14-16 Jun. 2014.
[10] J. Dilley, B. Maggs, J. Parikh, H. Prokop, R. Sitaraman, and B. Weihl, "Globally distributed content delivery," IEEE Internet Computing, vol. 6, no. 5, pp. 50-58, Sep./Oct. 2002.
[11] S. Jain, et al., "B4: experience with a globally-deployed software defined WAN," ACM SIGCOMM Computer Communication Review, vol. 43, no. 4, pp. 3-14, Oct. 2013.
[12] Network Working Group, The Common Gateway Interface (CGI) Version 1.1, https://tools.ietf.org/html/rfc2526 (Accessed on 11 May 2017).
[13] H. Hashim and A. Manan, "Active anycast method for server load balancing," in Proc. Student Conf. on Research and Development, pp. 105-108, Shah Alam, Malaysia, Malaysia 17-17 Jul. 2002.
[14] R. L. Pereira and T. Vazao, "Network layer implemented anycast load balancing," in Proc. IEEE Symp. on Computers and Communications, pp. 387-393, Las Vegas, NV, USA, 1-4 Jul. 2007.
[15] Network Working Group, A Border Gateway Protocol 4 (BGP-4), https://tools.ietf.org/html/rfc4271 (Accessed on 11 May 2017).
[16] N. Gude, T. Koponen, J. Pettit, B. Pfaff, M. Casado, N. McKeown, and S. Shenker, "NOX: towards an operating system for networks," ACM SIGCOMM Computer Communication Review, vol. 38, no. 3, pp. 105-110, Jul. 2008.
[17] S. Kaur, S. Japinder, and S. Navtej, "Network programmability using POX controller," in Proc. IEEE Int. Conf. on Communication, Computing & Systems, ICCCS'14, 2014.
[18] A. A. Dixit, F. Hao, S. Mukherjee, T. V. Lakshman, and R. Kompella, "Elasticon: an elastic distributed sdn controller," in Proc. ACM/IEEE Symp. on Architectures for Networking and Communications Systems, pp. 17-20, Marina del Rey, CA, USA, 20-21 Oct. 2014.
[19] A. Krishnamurthy, S. P. Chandrabose, and A. Gember-Jacobson, "Pratyaastha: an efficient elastic distributed SDN control plane," ACM Workshop on Hot Topics in Software Defined Networking, pp. 133-138, Chicago, Illinois, USA, 22-22 Aug. 2014.
[20] V. Yazici, M. O. Sunay, and A. O. Ercan, "Controlling a software defined network via distributed controllers," in Proc. NEM Summit, Implementing Future Media Internet towards New Horizons, 6 pp., 2012.