Keynote Speaker

Dr. P. Trinatha Rao

Sr. Associate Professor

Department of ECE, School of Technology

GITAM University, Hyderabad Campus

Hyderabad- 502 329, India

E-Mail: trinath@gitam.edu / polipalli9@gmail.com

Full Bio : Click Here

Title of Abstract : Energy Efficient Software Defined Framework

Software Defined Networking (SDN), a state-of-the-art approach in advanced Fourth Generation (4G) and Fifth Generation (5G) networks works on the separation of the control and user plane which adds a higher degree of freedom in steering traffic and maintaining service continuity.OpenFlow is direct implementation of SDN principles used to provide
network programmability. The SDN controller contains control functions and uses Open Flow or other protocols to control and configure forwarding devices in the data plane simplifying the network management tasks. OpenFlow is a widely used southbound interface between the SDN controller and the forwarding device.
Decentralized SDN architecture is formulated with SDN as the centralized controller having  multiple Content Delivery Network (CDN) servers deployed as proxy servers to create a distributed network managed in multiple data centres by forwarding data packets to the nodes in a particular area deployed. The motivation of CDN is to offload the SDN controller
responsibilities and thereby increase the performance and end to end user’s global availability, scalability, reducing bandwidth cost.
An energy efficient routing mechanism based on OpenFlow concepts is conceived considering the position of the nodes placed in the network. Unlike the conventional algorithms, the formulated Ambience Awake Routing (AAR) mechanism does not form a tree structure but multicasts beacons to the neighbouring entities based on their positions.
Location centred routing protocol (LCRP) uses the geographical location of the nodes for selecting the next set. Unlike the approaches in literature LCRP does not depend on any predefined arrangement of the nodes if the packet is being unicasted. It takes its motivation from the Greedy Perimeter (Karp et al., 2000) and face-2 (Prosenjit et al., 2001) routing
techniques. According to the regular location oriented routing techniques, it is assumed that  every single device is aware of its location by utilizing the global position services, the connecting devices in the neighbourhood is updated by the CDN controller in the flow tables and the next hop-node is decided by the mechanism for routing put in place. No concept of
flooding happens in unicasting or multicasting of the packets like the traditional routing techniques as there is no tree formation, which saves on the computational time and eliminates redundancy.
At the splitting junction of the device, replicas of packet are transited along all the connecting outlets. It is to be ensured that the length of the outlet path connecting destination devices is  to be short and the hop-counts are minimal. LCRP is entirely based on the local awareness of the position of the node but not on hop-count or on shortest branching. Greedy Multicast  Forwarding (GMF) is used to find the next set of hop nodes. The hop nodes are the neighbouring nodes which have a feasibility of accepting the packet and routing via them. OF devices like programmable routers and switches between the source and recipients replicate the data packets and then forward numerous duplicates wherever the path to destination nodes diverges. Only a set of nodes in the nearest geographical area are considered rather than the entire network area as potential next hop nodes based on their degree of connectivity (DOC). The DOC is determined by the battery energy level and power computation of the device to increase the probability of connectivity and thereby avoid packet loss. 
The proposed OpenFlow enabled SDN frameworkwith decentralized server for managing mobility and gateway functions proves to cost efficient when compared to the existing research works.The Ambience Awake routing algorithm exploits the position based routing principles and ensures that the neighbouring nodes are selected considering least  energy consuming entities resulting in a superlative mechanism in comparison to the existing techniques.