NFSv4 future features?

- Replication/Migration is in WG charter now (highest priority)

- Performance

- RDMA use

- directory delegation

- namespace mechanism and policy  
  
- features to support clustered applications

- Proxy caching (what is the problem?)
  - What is the problem? availability or wan performance
  - Directory delegation can help enable proxy caching
  - Reduce latency at the edge
  - Lightweight accelerators in front of overloaded servers
  - Remote development sites
  - Enhancements for distributed caching
  - Simple NFS caches

- Delegation of an entire filesystem or directory delegation

- namespace ideas:
  - fs_locations attribute may be sufficient for building larger
    namespace mechanisms and the work left could be at the server
    (similar to the replication/migration work)

  - would be helpful to replication/migration for the client to have
    observability of replicas before contacting the server

- applications modifications to maintain NFSv4's new attributes

- Quality of Service mechanisms and policy

- MIBs

- Snapshot notification to applications to allow for consistent
  data at snapshot; win applications would have access to this feature

- Interaction between auditing (ACL) and delegation; audit record
  communication to the server

- DAFS operations that would be applicable to NFSv4

- Lock/Open resource limits at the server are not reflected in the
  delegation mechanism

- replication/migration should allow for filesystem split


Observations

	- Motivation, originally, for NFSv4 was to move as close to
	  AFS but not present the barrier to entry

	- Namespace not embedded into the protocol like DFS or MS/DFS