Of course, you just can't abandon something that you have been enjoying for 10 years. And that would be a great pity. Yet as a strong FreeBSD advocate I feel that there is a whole forest behind this big tree and some of us, including myself, have been stacked in that tree. Please let me explain.
I have been following FreeBSD fanatically since 6.0-RELEASE. 8 years later in 9.1-RELEASE I am trying to find what has changed compared to the competition. The answer is a lot but the real question is what is the trend today, what do we need more or what are the possible competitive advantages that would make us choose FreeBSD.
The trend can be divided into two different categories:
I will start with Storages since I have been involved a lot lately. ZFS was marked production ready in FreeBSD 8.0-RELEASE. Since then, a lot of things have changed that affected the future of ZFS. The acquisition of Sun from Oracle having the most impact.
To make a long story short, a different group was created under the umbrella of Illumos, developing ZFS aside from Oracle. I was against that idea from the beginning because I always believed that the wheel should not be reinvented. We have now reached to a point that we have two different ZFS versions.
Now, besides an engineer I am also an entrepreneur. My experience has taught me that large corporations that want to make money are willing to cooperate and/or fund large projects. Why? Because they get to use that technology in return. Therefore, FreeBSD could have benefit from that if only there was the right approach.
Back to the engineering path, we assembled two identical SMB servers, like the ones we often use, and we tested ZFS on Linux and ZFS on FreeBSD using Samba and AFP transfers. Same ZFS version and same hardware. The performance on the Linux (CentOS) box was always better to my surprise. We were trying to simulate a real world scenario where six clients would transfer data from and to the storage at the same time. It turned out that the Linux storage was able to complete the transfers 5-10% faster. Then, we did the same using Solaris11 ZFSv34. The performance increase was 30% compared to FreeBSD.
Point (2), virtualization. I honestly believe that KVM has progressed so much that it is not even worth comparing it with anything else. If you add Cloudstack to the picture you get a full hybrid cloud infrastructure. Someone mentioned BHyVe earlier. Ask yourself this question, when will it be production ready and by the time it becomes production ready how far will the competition be?
Point (3), package management. It is very disappointing that even with PKGNG on the way I can't create a simple Wordpress installation.
Point (4), I don't understand why -STABLE is not being delivered over binary updates. Why can't we have 9.1.1, 9.1.2 for example. Why only security updates and not important binary patches that fix various errata issues?
This is not just someone (censored) complaining here for no reason. I care for FreeBSD and I would like to see the project advance. But we need to wake up, step away from the romantic view and face reality.
Best
