Is Object Storage happening?

Here is an answer I wrote to my dear friend StorageBod’s latest piece on Object Storage. As my response turned out quite a bit longer than initially intended, I thought I’d post it here as well.

Hi Martin,

Very nice blog post as usual. Let me sacrifice my first sunny afternoon of the year in Belgium to write down some of my thoughts (rather than copy/pasting a data sheet  like Paul did). Oh and … NAB would have been so much better with you over there,.

Disclaimer: I am still in charge of the object storage product line at DDN. The below thoughts are my own, however, not necessarily the company’s.

“Yet, it isn’t really happening”. I expect to be punished for writing this, but Martin: you are WRONG! Object storage is happening very much. Just not necessarily out in the open. Implementing object storage is a lot more game-changing than buying a new filer from a supplier you haven’t worked with before. We would have hoped for a faster wide adoption of object storage, but if you look at other recent paradigm changes, we are on the curve: I gave my first cloud presentation back in 2007. It took about 5 years for cloud to be embraced by the enterprise market. You could say that object storage has been around for more than five years (Centera etcetera), but if you look at the recent generation of object storage and the more modern use cases, we are on schedule. I wish I could talk to you about a few of those double-digit petabyte projects I’ve been involved with in the past 2 years.

“It turns out the developers really want a POSIX file-system”. We have had the REST vs. CIFS/NFS discussion before. I believe you and I agreed that a file gateway is unavoidable in the process to wider object storage adoption. Many applications are not object-ready and while we are waiting for the ISV’s to adapt their software, an extra file system layer can do the trick. As a matter of fact, a lot of the discussions I had at NAB last week (see, here is where it helps to be at the show rather than reading 140 character synopses), both with partners and customers, showed that we are getting there: DAM and MAM players are going for direct interfaces with the object storage pool.  That said, in many cases I don’t see why companies would even try object storage. Object storage was designed for large sets (multi-petabytes) of mostly immutable data. Those terabyte-scale projects to test the water, often just to replace filers, are not the right use cases. 100 TB of object storage with a file gateway on top will not be a better solution than 100TB of regular NAS.

“I’ve not gone down the route of putting in an Object storage solution because finding one which is supported across all the tools in today’s workflows is near impossible”. I agree: object storage will never be supported by all tools in your workflow. As I mentioned before, object storage was designed for data sets of mostly immutable data.  As you would have seen at the DDN booth if you had visited NAB (sorry to rub it in), our object storage platform (WOS) was positioned for collaboration, distribution and archiving of media assets, but not for creation or post-processing. Those steps in the media workflow benefit from other platforms. The nice thing about DDN’s WOS is that it perfectly integrates with our other storage platforms so that each platform can be used for its best purpose.

“[…] to provide us with the sort of transparency we need to support complex digital workflows”. Maybe you should also think about making you workflows less complex?

“I regularly suggest that we put in feature requests to the tools vendors to at least support S3; the looks I generally get are one of quiet bemusement or outright hostility and mutterings about Amazon and Cloud.” I can’t imagine that you would let yourself be intimidated by that! There are plenty of examples – also in your industry – of large media deployments on S3. So it makes total sense to add these requirements. That said, deploying an internal object store probable makes more sense than using the public service.

“So give it 20 years or so and we’ll be rocking.” See? We agree. But 20 months will be more than enough for object storage to rock. Technically, if you’d count all the projects I can’t talk about, it already does.

“object-storage gateway into the base operating system” Isn’t that what WebDAV is? OSX does this just fine.


~ by tomleyden on April 15, 2013.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: