Abstract:
Due to the emerging trends on scalability, flexibility and automation in data centers, some of
the most recent technologies include Vxlan L2EVPN. The focus of this investigation is on the
deployment of a Vxlan-based data center network built on EVPN control plane. The work involves
configuring Spine and Leaf switches, setting up BGP peering sessions, and deploying
Vxlan overlays for better communication among network devices. The devices have been configured
using a single point of management which acts as a control node. The devices have
been configured using Ansible. Following, a comprehensive YAML files for configurations,
shows the process step by step on how to setup up network infrastructure. The effectiveness
of the implemented solution is measured by running different show commands such as “show
ip bgp” and “show bgp summary,” that give an overview of how well the network is configured
via Ansible, together with BGP relations between neighbors. Secondly, the analysis looks
into MAC learning in Vxlan and confirms the VLAN-VNI mappings by making use of “show
evpn vni” command as well as “show vxlan vlanvnimap”. Ping tests between end hosts confirm
the correctness of set Vxlan network, as they prove communication across VLANs and among
specified virtual networks. The thesis goes over the implications of these results and highlights
how MAC learning affects Vxlan packet forwarding. The research outlines recommendations
for improved network optimization through constant monitoring, further security updates and
continuous training of the network administrators. The results serve valuable implications to
practitioners in the Vxlan-based Data Center Networking, providing a theoretical manual for
deployments.
