We will add support for Availability Zone Awareness to HopsFS - that is, HopsFS will be enabled to run across data centers in the cloud within the same region, and will be made data-center tolerant (it will continue to work correctly even if a data-center fails). An example of a data-center is an Amazon Web Services Availability Zone. If HopsFS spans more than one availability zone, it would continue to function if an availability zone becomes unavailable.
We introduced LocationDomainIds in HopsFS similar to NDB LocationDomainId, that allows admins to assign namenodes and clients in HopsFS to their respective availability zone. HopsFS will guarantee that a client from a specific availability zone to be connected to a namenode from the same availability zone if there are namenodes in this availability zone.