Jump to: navigation, search


Difficulty Hard
Topics testing, tempest
Mentor Debo~, Yathi, Boris?

The scheduler today is not very scalable when the size of the Openstack deployment becomes large. This project will first benchmark existing scheduler code and figure out ways to optimize the scheduler. One of the things that has been done is to use memcached like layer to store the data from nova. Another way is to connect the memcached layer to a forklifted scheduler and make it scalable (using multiple threads or scaling out).