[OpenAFS] OpenAFS vs NFSv4?

[Ted Anderson]

On 4/28/2003 10:50, Rodney M Dyer wrote:
> As a long time user of AFS, I don't understand how some people don't get 
> the "global namespace" idea.  I also don't understand why this isn't 
> being promoted as -the- single most important criteria of the 
> versatility of AFS.  As I understand it, no other filesystem in 
> existance has this capability, right?  I mean there is of course 
> Transarc's DFS, which almost no one uses anymore.

This question about the acceptance of the global namespace idea is a 
good one.  I don't think it is quite as simple a matter as having the 
clients perform the location resolution task.  Pin-pointing exactly why 
the approaches used by NFSv4, or automounter, or CIFS/MS-Dfs are not as 
good is hard.

I think the reason may be that the global namespace is an enabling 
technology that doesn't accomplish much all by itself.
    * it provides location independence
       * no, that is due to the VLDB
    * it provides replication
       * no, that is implemented by the volser
    * it provides scalability
       * no, that is because clients do the VLDB lookup
    * it decouples the logical data and physical infrastructure
       * no, transparency of volume moves gives that advantage
    * if provides uniformity
       * no, mountpoints in the file system accomplish that
etc.  All highly simplified, of course.

But the namespace is part of making all these features useful in the 
first place and many of these features enhance each other.  For example, 
without replication of volumes at the top levels of the namespace 
(and/or aggressive caching of those directories) the scalability 
advantages of having clients do the VLDB lookups would not save the 
server hosting root.cell from getting killed with lookups.  Similarly, 
replication is not nearly so useful if it isn't smoothly integrated into 
the naming system so that users just pickup replicas automatically.

So I'd be tempted to claim that the global namespace, plus a certain 
critical mass of other features, is the key to AFS's versatility.  Other 
systems don't have enough support to make their namespace really work.

> These are only my opinions.  Personally I'd like to tell two friends 
> about global-namespace filesystems, then they tell two friends, and so 
> on, and so on.  I just don't get how most AFS administrators have just 
> let the idea die by not promoting it.  Sometimes I guess we just get so 
> used to the idea, it just seems natural that everyone would understand it.

The key to getting the idea to spread, is to create a clear and sharp 
exposition of the benefits which highlights the distinctions with other 
distributed file systems.  I don't think this message does the trick, 
but hopefully it helps.

Ted Anderson