FreeBSD 7 will be revolutionary.

A few weeks back, at the end of December, FreeBSD 7.0-RC1 was released. FreeBSD 7 will no doubt prove to be quite revolutionary. For one thing, this will be the first major FreeBSD release in a number of years. FreeBSD 6.0 was released in November of 2005, so there has been quite some time for the development of FreeBSD 7 to take place.

If you’re unfamiliar with what FreeBSD 7 will bring, I’d suggest that you look over the excellent What’s cooking for FreeBSD 7? Web page. As you can see, the amount of change FreeBSD 7 will bring is quite significant.

I’m particularly looking forward to FreeBSD’s support for ZFS. ZFS alone is a rather revolutionary filesystem originally developed at Sun for Solaris. It addresses many of the difficult issues we face when it comes to dealing with the huge datasets that are becoming all too common these days.

Having support for ZFS available in FreeBSD will be a major win for those who already have extensive FreeBSD infrastructure in place, but need the features that ZFS offers. So it will become possible for them to continue using FreeBSD, rather than having to integrate Solaris into their network infrastructure. This proves very helpful in maintaining an effective, reliable server environment.

Another major improvement will be the use of the new jemalloc userland memory allocator. While phkmalloc has served us well for some time now, jemalloc has been designed from the ground-up to take into account the multiprocessor systems that are becoming virtually ubiquitous today. For those of us who are dealing with heavily threaded software running on multiprocessor systems, there is evidence that jemalloc will bring some significant performance gains.

Also of importance are the improvements to the networking stack. With gigabit (or faster) network cards being the norm these days, FreeBSD’s support for TCP/IP Segmentation Offload (TSO) and Large Receive Offload (LRO) will no doubt prove to be very useful. Along with the new sendfile() implementation, and the improved sosend() functionality, we will likely see some large networking performance boosts.

There are, of course, many other improvements that have been made. There are security enhancements, improved audio support, libthr becoming the default threading library, updated support for executing Linux binaries, and the new SCHED_ULE replacement, just to name a few.

It’s time for those of us in the IT profession to start considering the use of FreeBSD 7. The new features and improvements offered by this release will no doubt have a great impact for many of us. We will be getting better support for storing huge amounts of data, networking performance improvements to help us better transmit that data, and userland performance improvements to better let us manipulate it. And we can’t forget to thank the many FreeBSD developers and contributors for all of the time and effort they have put into creating such an excellent operating system.

One Response to “FreeBSD 7 will be revolutionary.”

  1. eric Says:

    FreeBSD all the way!! It will simply rock the market :)

Leave a Reply

*
To protect against spam, please type the word in the picture. Click on the picture to hear an audio file of the word.
Click to hear an audio file of the anti-spam word