While cloud allows users to provision resources on-demand and gives an illusion of unlimited capacity, in reality the underlying resources are limited and require resource sharing which could cause inefficient cloud operation. This is especially important for NFV clouds, where service degradation may occur if additional resources are not available adjacent to the resources for a network service that may experience expected or unexpected spike in traffic.
OPNFV Promise project was initiated to solve this problem by introducing resource reservation for future usage. Since OpenStack is considered to be the defacto VIM for NFV, OPNFV Promise team is collaborating with OpenStack Blazar project and Product Working Group to address this issue.
This session will aim to provide an overview of how Promise and Blazar address the following
- Accept future resource usage requests
- Provide guaranteed access to reserved resources
- Manage resource usage conflicts
In this session, attendees can learn why resource reservation is important, not only for NFV workloads but for general cloud operators. We will provide details on the results of collaboration between OPNFV Promise and OpenStack Blazar project teams which includes the following:
Generalized user story for capacity management for cloud providers & operators (via OpenStack PWG)
- Gap analysis between this user story and *real* OpenStack features
- Collaboration to update OpenStack Blazar code
Since NFV is just one use case for OpenStack, OPNFV Promise team has agreed to expand the scope to generalize the requirements while developing the user story in collaboration with Blazar team. Attendees can learn how Promise project has translated its specific requirements to generalized OpenStack requirements and how they are being addressed by the community.
