From a basic level you cannot compare the NetApp to the Clariion, you have to compare it to the Celeara (EMCs NAS Product). You have to remember that NetApp is a file level device, and as a file level device it is pretty good, mostly because of its user interface and ease of use. If you look into the their method of data read/write, you will find that the random write anywhere process (even though fast and attractive in the beginning) presents a real problem as the device space fills up, and the data is continually being changed. It is a simple matter of Physics.

You write in any open space, you do not reorganize files to continuous blocks, which the means you are writing fragmented files (or data segments), as the unit reaches 60-70% utilization of space, the problems start to become apparent. To read a file the head has to move to the appropriate space to read the next segment of the file. This means physical movement. As the files become more fragmented (because the NetApp) does not restructure the space or move data for efficiency, the need for the head to move to read the file becomes a burden. Thus NetApps are very ineffective after 60-70% space utilization.

The other point is to argue the proper place for data. Block data performs best on block level devices (SAN), file level data performs best on file level system (NAS) , with out getting into all the detailed reasons, as simple on is this. A file stored on a block level device can be rendered unreadable based upon a small amount of block level errors. Alternately a Block level application (such as a data base), operates inefficiently on a file level system (or obvious reasons), and the file level system does not have the protective features, etc. to handle a block level failure (that could be critical to a database).

Every manufacture has its benefits and disadvantages… In Storage, the all use the same spinning disk, it’s the software that make the product work, and for diversity it’s the offerings the company has…

Their filesystem is in fact not the reason for their commercial success.

Their commercial success is a result of relatively high prices, great margins, great sales people, great support and great hardware.

Pretty simple.

How about they make it available as binary? A loadable kernel module for their own customers (that works fine with FreeBSD)? It doesn’t require the distribution of source code. How about doing what they’re doing now and suing people that use it to compete with them, but make it fine for customers to use?

How about simply doing what someone like Nividia does when they ship a driver for a video card on FreeBSD?

Are those source code distributions?

Nope.

Guess where they sit?

/kernel/

I didn’t say that NetApp owed it to the world to give away its core technology.

I didn’t say that they even owe it to their block storage customer.

I said that considering the lack of a decent filesystem to put on top of their block storage, they shouldn’t be surprised that we flocked to ZFS. And now if they’re actually going to sue and end up being successful in taking ZFS away from us (that’s their goal right? Why else do it?), then yes, they would owe us WAFL.

“You should have given WAFL or a WAFL-lite file system to FreeBSD and then all of us could happily use it on top of iSCSI or fibre-channel block storage.”

I don’t think this is fair.

NetApp is WAFL. It’s the source of all their power — just about every unique feature owes its existence to this layer of block indirection. It was invented over a decade ago, and was truly unique. Innovation is surprisingly rare in computing — most companies trumpet it, but few deliver. WAFL is a true innovation — why should the company give it away?

You say that NetApp “could have figured out how to do it while making … even more money, and without spawning a bunch of FreeBSD storage appliance clones.”

This is a hollow compliment. How would one actually achieve this? If the WAFL innovation is the source of NetApp’s success, why should they give it away, and why do you assume they would prosper without it?

I agree that both companies should “kiss and make up”, and I think that’s the inevitable conclusion, but there’s no reasonable argument to say that NetApp somehow owed it to the world to give away its core technology.

@Dick Davies, at this point? Well NetApp went nuclear on Sun, and if they couldn’t work it out before Sun’s response, then Sun had no choice really but to go nuclear back. ZFS is that important. However, they both need to stop their respective superiority things that’s been going for years and years, and realize that neither provide nor have provided a full solution. Going back and forth like this doesn’t help customers.

While the CX are dual protocol systems the NS series are multi-protocol systems. (NAS/ISCSI/FC) The gateway products serve a different customer segment.

That should be “its” – drop the apostrophe. What you are saying here is “NetApp cannot sell it is filers,” which does not make any sense.

And yes, I agree: NetApp wouldn’t exist if Sun had a better clue about what its customers wanted, historically, and I’ve always been surprised that Sun didn’t try harder to acquire NetApp.