... > The problem it seems to me with criticizing ZFS as > not much different > than WAFL, is that WAFL is really a networked storage > backend, not a > server operating system FS. If all you're using ZFS > for is backending > networked storage, the "not much different" criticism > holds a fair > amount of water I think.
A more fundamental problem is that there are several different debates going on in this one thread. The comparison with WAFL is primarily about the question of just how 'novel' ZFS's design is (leaving aside any questions about patent enforceability) and especially about just how 'unique' its reliability approaches are for environments that require them. In a nutshell, while some COW approaches predate both WAFL and ZFS, WAFL was arguably the first to come up with the kind of 'write anywhere' approach that ZFS also heavily relies upon and to the best of my knowledge WAFL was also the first to incorporate the kind of in-parent-verification that has played such a prominent part in the integrity discussion here. Another prominent debate in this thread revolves around the question of just how significant ZFS's unusual strengths are for *consumer* use. WAFL clearly plays no part in that debate, because it's available only on closed, server systems. However, that highlights > what's special about > ZFS...it isn't limited to just that use case. The major difference between ZFS and WAFL in this regard is that ZFS batch-writes-back its data to disk without first aggregating it in NVRAM (a subsidiary difference is that ZFS maintains a small-update log which WAFL's use of NVRAM makes unnecessary). Decoupling the implementation from NVRAM makes ZFS usable on arbitrary rather than specialized platforms, and that without doubt constitutes a significant advantage by increasing the available options (in both platform and price) for those installations that require the kind of protection (and ease of management) that both WAFL and ZFS offer and that don't require the level of performance that WAFL provides and ZFS often may not (the latter hasn't gotten much air time here, and while it can be discussed to some degree in the abstract a better approach would be to have some impartial benchmarks to look at, because the on-disk block layouts do differ significantly and sometimes subtly even if the underlying approaches don't). - bill This message posted from opensolaris.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
