Telecommunication industry traditionally deployed OpenStack in a datacenter or central office. Advent of 5G, IoT, Augment Reality (AR), Virtual Reality (VR) and universal Customer Premise Equipment (uCPE) requires applications processing closer to the customer, i.e., at the edge.
Edge brings up additional use cases that were not originally addressed by OpenStack hosted at the traditional datacenter deployments. The additional use cases include thin deployments (2-3 servers), 10K small to medium sites, zero touch provisioning, support of GPU and other special video processing hardware, complete software defined storage, single pane view of all logs and events, etc.,
We have explored several architecture options to run OpenStack at the Edge, including running OpenStack as containers, a regional controller, etc. OpenStack needs to be modified to extend it to run efficiently at the edge.
In this talk we will share our insights of deploying OpenStack at the edge.
- Use of special hardware or no special hardware at the edge?
- What are the real use cases of edge from 5G, IoT, AR, VR, etc.?
- Container versus virtual machines at the Edge, what make sense from business and economic aspects?
- Architecture recommendation to the OpenStack in support of Edge requirements.