Browsing SUNY Polytechnic Institute by Author "Gonuguntla, Raja Bhushan Rao"
A Technology Case Study on Integrating Open Stack with SDN for Internet Connectivity using BGPGonuguntla, Raja Bhushan Rao; Hash, Larry; Advisor (2016-12)There were many developments in Internet usage, which resulted in significant increase in Internet routing. With existing networking infrastructure, it is difficult to meet these requirements and causing more cost and less performance. Since network devices are hardware modules, processing them requires more power and more memory. However, if network protocols are developed using software modules, flexibility can be achieved in various programming applications and reduces dependency on hardware. The concept of using networking protocols as a software module can be explained using “Software Defined Networking (SDN).” With SDN, existing infrastructure can be integrated with various applications and centralized control protocols can be developed. One of the key components of SDN is integrating with Cloud Computing, where many applications can be built, which can be used for on-demand services. Integrating cloud computing with SDN will create dynamic networks and reduces infrastructure costs. In this paper, a case was considered for providing better internet connectivity by building public & private networks using Open source cloud technology (OpenStack) and existing distribution environments. For connectivity, BGP was used as routing protocol as it is known to be well- suited for such environments. Both public and private networks were integrated with SDN for centralized control. OpenStack was used to build various network topologies using different plugins through SDN controller. This method allowed to develop SDN controller with global view of OpenStack networks. The same controller was connected to distributed layers using Open Flow protocol. Since, both OpenStack and distributed networks were attached to SDN controller, centralized control of network protocols could be achieved. This model of centralized networks could be very useful in reducing costs and improving network efficiency, especially in large scale deployments.