Part I discussed performance and and some data integrity features of ZFS. Now for some more cool features and the StorageMojo.com conclusion.
Physician, Heal Thyself
On-disk bit rot is a real and continuing problem. For a lot of reasons, data can go wrong. The important thing is getting it fixed. As mentioned in Pt I, ZFS separates checksums from their blocks, so it can detect both corrupted and/or incorrect blocks. It also fixes them when found in the course of an I/O.
Even better, ZFS maintains a background process that traverses the metadata and verifies the validity of each block. This process is analogous to the ECC memory scrubbing that EMC used in the original Symmetrix line with its large, single point of failure cache.
Whether fixing a single block or replacing a failed disk, ZFS uses the tree-based checksum structure to ensure valid data.
Snapshot Copy, Cheap. Real Cheap.
What, that’s it? No Double Parity, no cluster support, and it needs to be ported to OS X. Those are the downsides?
Well if they are the only downsides and this is yet another world changing Sun idea (Shocking how they can never seem to make a penny out of any of those) then it pretty much screws Sun’s ambitions of making it in the storage industry. Any person with a free Solaris download and a disk pack can just eliminate the need to buy STK gear from Sun.
That’s the rub for this talented team, isn’t it? Sun, like most IT providers, has a commission sales force. If they want to send their kids to college they need to sell stuff. The commission on FREE won’t do that. So my advice to storage entreprenuers is to build and support ZFS-based storage systems. You won’t have to worry about competition from Sun.
I dunno, I’m not sold on the whole ZFS is going to clean array vendors clocks idea, though I probably would say that since I work for one. (Settle down PR folks, all opinions herein are my own and not those of my employer. Note to readers, no agenda in these comments just commenting on what I think as guy who sees a lot of storage environments and speaks to a lot of IT folks running those environments)
Lets look at Symantec customers for example, many of them have been using VxVM and VxFS for years on top of Luns presented from arrays, even though to listen to Veritas you weren’t supposed to have needed an array if you had bought their stuff. Why would it be any different for the person using ZFS? For the most part folks using Solstice Disksuite will just swap out SDS for this, and as an SDS user myself I’ll be one of them as lets face it it’s a pig, perhaps a lot of Storage Foundation for Solaris customers will do the same thing too unless it’s a corporate standard.
So while I can see a future where ZFS tanks Symantec revenues in non-VCS on Solaris shops unless Symantec can come back with something which raises the bar higher, I’m not sure folks are going to abandon fast centrally managed storage for whatever amount of disks they can daisy chain together themselves en masse.
I think they’ll do what they’ve always done, for performance apps the storage team will allocate out a bunch of luns which the admins or app guys will then scrape together in whatever fashion they choose using the volume manager on the host. That’s what they’ve been doing to date regardless of if it’s a bunch of local disks in a server or something sitting in a refrigerator sized array.
Good point, Mark. And I agree. Absent a compelling business reason to change, IT won’t. That is why they always find themselves dragged kicking and screaming into whatever the latest is.
Which is why I tend to focus more on the applications that will drive new storage models, rather than the technology. When a business unit sees themselves losing money or market share because someone else is using more cost-effective storage, then change starts.
I also believe that open-source storage needs a different business model than other open-source software. The data stored is just too important to distribute responsibility across several groups. Stay tuned for more thoughts on that. And thanks for the thoughtful comment, Mark.
ZFS will not be fully appreciated until it gets into the Linux kernel. Personally, I can’t wait.