I confess that I didn't realize we lacked client side caching in our
NFSv4. I thought client side caching was one of the significant
benefits that NFSv4 brought to the table (mainly to compete with the
likes of AFS and DFS).
If we don't have client side caching in NFSv4, then I too worry that the
loss of functionality will leave us with a potentially significant
architectural hole.
Again, for folks on local networks, the network may not be the
bottleneck, but with VPNs and similar technologies increasing the global
reach of corporate networks, I think there is still a real business case
for client side caching.
So, please either provide information as to how this gap will be filled,
or evidence as to why it isn't necessary.
-- Garrett
Darren J Moffat wrote:
> Personally I agree completely with the EOF of the current
> implementation, having been directly involved in supporting it in Sun
> Service and sustaining.
>
>
> However I disagree that we no longer need such functionality. I also
> find it quite sadly ironic that other operating systems are only now
> starting to implement something like cachefs just we Solaris is about
> to pull the rug out from underneath it.
>
> See: http://en.wikipedia.org/wiki/CacheFS
>
> NFS+CacheFS+Kerberos was a good competitor for AFS if we drop the
> CacheFS part we only have base NFSv4+Kerberos and IMO that level of
> caching is not good enough.
>
> I believe we need a persistent (ie local disk/ssd) cache integrated
> with NFSv4 before this EOF can be approved. However if there is
> business evidence that CacheFS is no longer required functionality
> (rather than no longer used because it doesn't work with NFSv4 or is
> buggy) I'd like to see that.
>
> Basically from an architecture view point this EOF leaves us with a
> gap that I think needs to be filled.
>
> --
> Darren J Moffat
