[ Full disclosure: I work for iXsystems™ . Opinions and views are mine. ]
The story began …
It was 2011. A friend of a friend called me out of the blue. He was rambling about his company’s storage needs. I recalled vividly that he wanted 100TB, and Dell and HP (before HPE) were hopeless doing NAS (network attached storage) in an Apple environment. They assembled a Frankenstein-ish NAS and plastered a price over MYR$100K around it.
In his environment, the Apple workstations were connected to dozens of WD Cloud Book storage (whatever it was called back then), daisy chained via Firewire to each other. I recalled one workstation had 3 WD “books” daisy chained together. They got the exploding storage needs but performance sucked. With every 2nd or 3rd user, access to files were at a snail pace, taking up to more than 2 minutes to open a file sometimes.
At that time, my old colleague at Sun was fervently talking about ZFS and OpenSolaris™. I told him about this opportunity, and so we began. It was him who used the word “crafter”. “We are not building“, he said, “we are crafting“. He was right.
Open Source Storage Economics – Instant Profit
We doubled our money in that opportunity with a price that was 25% cheaper than Dell or HP. The Supermicro® box that we crafted was running OpenIndiana with the napp-it GUI, both open source and craft-ready.
Given our enterprise backgrounds (both of us were Sun Enterprise E10K systems engineers) and deep storage knowledge, we recognize that this win was made possible by open source storage software.
The Builders, The Assemblers, and then The Crafters
Open source projects mean different things to different people. Many, especially the non-practitioners in the specific field, would assume there will be coding involved. There will be some degree of putting the different modules together to have an acceptable working solution. Open source storage software projects will also have the same set of perceptions.
I cannot fault them. Writing the codes for a new storage file system is a huge undertaking, and many presume open source storage would require coding. That would be the misnomer to the building a storage. The usual assumption would be one would not be able to recoup the costs of development for years and unable to get to value to the customers, let alone gaining profits from it.
My foray into assembling vanilla Gluster in 2011 and OpenStack® Swift in their early days was better than the coding part but nonetheless, difficult. Too many hoops and hurdles to overcome. And many storage software would fall into this assembler category at the time, including Ceph*. And the presumption thinking open source storage economics costs money and time led to many pre-maturely deciding you could not make money with Open Source storage. They were wrong.
* The reason I did not go for Ceph was its very “encapsulating” architecture. Too many things to assemble. Too many things within things. I studied Ceph architecture more in-depth than Gluster or OpenStack® and decided it was not for me. For a longer explanation of my reasons, send me a message
The 3rd category in Open Source Storage is Crafters. The early story of my partner and I crafting an OpenSolaris™ ZFS software-defined storage with x86 whitebox for our first customer in 2011 proved that Open Source storage economics could bring instant value. We built an (almost) enterprise, production-ready storage array with Supermicro® without coding, without any assembly.
The instant value of TrueNAS® hiding in plain sight
The non-Open Source practitioners would presume that Open Source Storage would be:
- Command line and full of codes
- Assembly required
- Unreliable and for small requirements only
- Cannot scale and lacks performance
Unbeknownst, there are several open source storage projects conspicuous with instant value. They are hiding in plain sight. Here are my favourites:
You can say that I am biased because I work for iXsystems™ but I have been working on TrueNAS® CORE and MinIO several years before I started at iXsystems™ less than 2 months ago. I have been in the storage technology industry for 28 years, and have become a connoisseur. I have come to appreciate the really good ones, and these 2 top my list of the best. They have instant value.
Value of crafter-ready Open Source Storage
I have to be specific. Not all Open Source Storage are the same. They have to be “crafter-ready” to have instant value. And smart people can see that.
- TrueNAS® CORE is the world’s most downloaded open source storage OS with 10+ million downloads. Why is that so?
- MinIO is the world’s fastest growing object storage with 400+ million Docker pulls. Why is that so?
And these projects have strong adoption, which feeds back into the value lifecycle, strengthen its base, new solutions and more.
Gathering from the experiences of TrueNAS® CORE in the field, many of the demands to scale, to serve the enterprise are ploughed back into iXsystems™ TrueNAS® Enterprise offering. The value of Open Source Storage is unprecedented compared to the proprietary storage software. That value can be seen from iXsystems™ latest flagship storage array, the TrueNAS® M60.
I can almost hear the Illumos community leader (Hey Garrett!) ranting about vendors and users who take the OpenSolaris™ code and not contributing back. This was something I caught in someone’s conversation many years ago.
But it is not like that. The Open Source community is a strong ecosystem, one that grows beyond the market and profits. The core values of Open Source define the intrinsic value proposition far and beyond. And it is vital to keep the ecosystem thriving and moving forward.
The TrueNAS® CORE project has been fortunate to have iXsystems™ owning FreeBSD® development and corralling the community to keep building forward and at the same time, co-helming the stewardship of the OpenZFS project as well. That powerful combination of FreeBSD® and OpenZFS opens up to the future of TrueNAS® version 12.0 and beyond.
Long term and short term, instant and further into the future, Open Source Storage, the crafter ones I have described, have much more values far reaching.