ZFS in Leopard

The original “leak” occurred this morning when Sun CEO Jon Schwartz mentioned in front of a corporate media event for Sun’s latest blade servers that Apple will be announcing at next week’s WWDC (World Wide Developer Conference) that ZFS will be Leopard’s default file system.

The news spread across the grapevine like wildfire, coming close on the heels of similar rumors which rumor-mongers have been analyzing since last week.

Thus far we have been able to get clear-cut confirmation from at least three separate, well-established sources in Cupertino and fairly confident confirmations from several third party developers who are privvy to the “mainstream” build of Leopard which includes many features not included with developer seeds to date.

The “mainstream” build will be merged into the developer-test variant as of the WWDC Developer Preview 2 release (last year’s WWDC brought DP1 and there have been over 20 major seeds plus countless minor ones since then) and only a handful of features, if any, will remain secret after that according to the latest source reports out of Infinite Loop.

ZFS includes numerous features and optimizations found nowhere else — from powerful RAID enhancing features, unique data structure, industry-leading volume management tools, extremely efficient partitioning/instances handling, support for extremely large volume sizes (well in excess of 1024 Terabytes — e.g. Petabyte territory).

Of course it also supports the features offered by Leopard’s advanced HFS+ implementation including Journaling…..improves the functionality of Spotlight’s meta-data databases, and handles all real or virtual devices independently in what is called a “zpool” instead of the way other FS’es work by creating one “true” device and then assigining all other devices secondary/virtual status.

I/O performance, high-end server/cluster/workstation reliability & feature support, as well as failure-recovery and data security are all dramatically enhanced with ZFS. Early benchmarks show that Leopard’s powerful ZFS implementation can boost performance of large RAID arrays, Flash memory drives and many other disk types by anywhere from 5 to 50% or more (in the case of large software RAID arrays composed of powerful high-end drives; hardware RAIDs also lean to the higher end of that range but the difference may not be quite as dramatic in that case).

We’re very pleased by this news and can’t wait to hear all the details of how current HFS+ Journaling volumes will be updated to ZFS — apparently without the need for a reformat, according to Apple sources. Expect much more on this developing story in the days to come!


Share this post: Share this article on Facebook Share this Article on Twitter Add this Article to Stumbleupon Add this Article to Del.icio.us Add this Article to Digg Add this Article to Reddit Add this Article to Newsvine
This entry was posted in Apple Software and tagged , , . Bookmark the permalink. Both comments and trackbacks are currently closed.