A funny thing came up on my Twitter feed last week. There was an ongoing online voting battle pitting FreeNAS™ (now shall be known as TrueNAS® CORE) against Unraid. I wasn’t aware of it before that and I would not comment about Unraid because I have no experience with the software. But let me share with you my philosophy and my thoughts why I would choose TrueNAS® CORE over Unraid and of course TrueNAS® Enterprise along with it. We have to bear in mind that TrueNAS® SCALE is in development and will soon be here next year in 2021.
The real proving grounds
I have been in enterprise storage for a long time. If I were to count the days I entered the industry, that was more than 28 years ago. When people talked about their first PC (personal computer), they would say Atari or Commodore 64, or something retro that was meant for home use. Not me.
My first computer I was affiliated with was a SUN SPARC®station 2 (SS2). I took it home (from the company I was working with), opened it apart, and learned about the SBUS. My computer life started with a technology that was meant for the businesses, for the enterprise. Heck, I even installed and supported a few of the Sun E10000 for 2 years when I was with Sun Microsystems. Since that SS2, my pursuit of knowledge, experience and worldview evolved around storage technologies for the enterprise.
Open source software has also always interested me. I tried a few file systems including Lustre®, that parallel file system that powered some of the world’s supercomputers and I am a certified BeeGFS® Systems Engineer too. In the end, for me, and for many, the real proving grounds isn’t on personal and home use. It is about a storage systems and an OS that are built for the enterprise.
The Sun DNA
Back in 2010, my partner and I were working on OpenSolaris™ b134, OpenIndiana, the ZFS filesystem (which we absolutely swear by it) and napp-it, the ZFS GUI we wrapped the OpenIndiana ZFS storage with. We also tried out Nexenta® Core and the commercial NexentaStor as well but the per 8TB tier license was prohibitive back then. My first venture into starting my own company in 2009, NextIQ, was reselling Nexenta®.
After OpenSolaris b134 closed, Oracle® has killed our source of building a software defined storage array in our backyard. My partner left. I went back to employment.
It’s the file system man
A file system takes about 10-15 years to mature. And its usefulness for the enterprises and businesses alike is by its adoption in the storage operating system or storage operating environments. Proprietary file systems and storage OSes usually have more opportunities to succeed. That is why the NetApp® ONTAP and WAFL or Pure Storage® Purity OS among others dominate the enterprise storage market.
For open source file systems, the success factors have to take a different route. That is why the selection of a file system for a storage OS is vital for its market adoption success. I have reviewed some open source file systems that were meant for storage OSes in the past, including btrfs. My first btrfs review more than 8 years ago was promising but sadly, btrfs has not achieved the “production ready” status at all since then. This means that btrfs is not good enough for the enterprise, and thus just for playground stuff for now.
ZFS on the other hand, powers the largest enterprises. Datto®, one of the largest MSPs in the world, runs OpenZFS. Delphix, where ZFS creator, Matt Ahrens works, runs their solutions on OpenZFS. Tegile IntelliFlash, now part of DDN®, used OpenZFS. And there are many more – OS Nexus®, Joyent™, QNAP® (one of their storage families), QSAN™ – have OpenZFS as the core technology that powers their storage offerings.
ZFS is bad ass
ZFS is always on my mind. The passion and the conviction for ZFS file system after the Oracle® “generosity” to close source OpenSolaris continued. I talked to the Illumos guys, and got to know Garrett D’Amore. I spoke on the phone with Garrett a few times if I recalled. He was building DEY Systems with Jason Yoho and Richard Elling at the time, probably around the same idea we did with OpenSolaris and the enterprise storage array just 2 years back.
When it comes to OpenZFS, I can envision Samuel L. Jackson saying “That is a bad ass Mofo“.
FreeNAS vs Unraid- the result
By the way, Unraid won. Congratulations to them! They are probably very good in the market segment or personal segment they are in. However, I am looking at the real thing where businesses and enterprises put their confidence on TrueNAS® CORE and TrueNAS® Enterprise.
In the end, it is not about “My Refrigerator” is bigger or better than “Your Refrigerator” or vice versa. The real test is can this scale beyond your personal hobby project or small medium business investment into a technology worthy for the enterprise and even HPC (high performance computing)?
I found the online voting battle between Unraid and FreeNAS™ amusing. Maybe because it was in a space I was not used to, the space of home hobbyists and personal DIY hacks. But let’s be real. Real enterprises and businesses don’t need a flashy storage to have to prove itself. Real businesses want a storage that sits silently and reliably in the data center for years and years, does its job without fail and continue to provide value.
The true value of TrueNAS® CORE is exactly that because I have built them and installed them, and supported them for real businesses for the last 4-5 years. Yes, it has proven to be a boring storage array with a boring file system but that is the way real businesses want it to be. It works damn well!