NFS is designed to give users high performance, transparent access to server file systems on global networks. Some of the most important design principles are summarized below:

• Transparent Access: Users and applications can access remote files as if they were local. They are not required to know whether the files reside on the local disk or on remote servers.

• Portability: NFS is machine and operating system independent. This allows it to be ported easily to multiple OS and hardware platforms from PCs to mainframes.

• Fast Recovery From Failure: NFS is designed to recover quickly from system failures and network problems, causing minimal disruption of service to users.

• Network Protocol Independence: NFS has the flexibility to run on multiple transport protocols instead of being restricted to just one. This allows it to utilize existing protocols today as well as new protocols in the future.

• Performance: NFS is designed for high performance so that users can access remote files as quickly as they can access local files.

• Security: The NFS architecture enables the utilization of multiple security mechanisms. This allows system administrators to choose the security mechanism that is appropriate for their distributed file sharing environment instead of being restricted to one solution. This also allows NFS to utilize new security mechanisms in the future.

These features are implemented within the NFS client/server framework which reduces costs by enabling heterogeneous resource sharing across the global enterprise. Servers make their file systems sharable through a process called exporting. Clients gain access to these file systems by adding them to their local file "tree" via the mount process. The NFS protocol provides the medium for communication between client and server processes over the network.

A New Revision to the NFS Protocol--NFS Version 3

In 1992, a group of vendors including IBM, Digital^TM, SunSoft^TM and others came together and began the process of defining the next revision to the NFS protocol. The result was NFS Version 3. NFS Version 3 offers substantial enhancements over NFS Version 2. These enhancements include, but are not limited to, the following:

• Improved client write throughput

• Reduced server load resulting in increased scalability and performance

• Improved support for systems utilizing Access Control Lists (ACLs)

• Support for large (multi-gigabyte) files on NFS servers.

More detailed information on the features mentioned above is covered in the following sections: "Improvements to Write Throughput with NFS Version 3" on page 15, "How NFS Version 3 and TCP Work Together" on page 10, and "Access Control Lists (ACLs)" on page 32.

Due to its many benefits, the industry is rapidly converting to the NFS Version 3 protocol. However to address backward compatibility with the installed base, it is possible to implement NFS so that both protocols are supported concurrently. For example in the upcoming Solaris 2.5 release, NFS clients and servers will be able to support both Version 2 and Version 3 concurrently. Clients and servers can negotiate which protocol to use based on what they both support. Figure 1 depicts this capability.

Figure 1 Concurrent support for both NFS Version 2 and Version 3 on Solaris 2.5

A complete description of the NFS Version 3 protocol can be found in the NFS Version 3 Protocol Specification available from SunSoft or by anonymous ftp from several locations on the Internet including:

• ftp.uu.net:/networking/ip/nfs/NFSV3.spec.ps.Z
  
• bcm.tmc.edu:/nfs/nfsv3.ps.Z
  
• gatekeeper.dec.com:/pub/standards/nfs/nfsv3.ps.Z