Linux & Unix

Some weeks ago, I run a complete set of filesystem benchmarks on a DELL R510 server equipped with 12x 2 TB SATA disk and a low-end PERC H200 controller ( you can read more here: http://www.ilsistemista.net/index.php/linux-a-unix/13-ext4-vs-xfs-large-volumes-with-low-end-raid-controller.html ).

Now I am able to replicate the same set of benchmarks against a much more powerful PERC H700 storage controller, equipped with 1024 MB of non-volatile cache memory (it use DRAM memory backed up by a battery unit).

Some months ago, I wrote an article comparing EXT3, EXT4, XFS and BTRFS filesystem performances with a Fedora 14 x86_64 installation done on a Dell Latitude D620 laptop. While the results were quite interesting (especially to evaluate BTRFS performance), they were collected on a consumer machine (a laptop), with consumer-grade processor and HDD. So, the results do not necessarily translate to server world in a linear manner – a very good filesystem for a single 2.5'' HDD can be inadequate for a multi-disk server machine, and vice-versa.

Today, thank to the “Center for Research Computing” at University of Notre Dame, and especially to Paul Brenner, Serguei Fedorov and Rich Sudlow, I am able to present you some filesystem benchmark results collected on a quite powerful Dell R510 server, loaded with 12 x 2 TB SATA disk connected to a low-end, inexpensive PERC H200 controller. The article will focus on EXT4 vs XFS performance, as EXT3 can not grow bigger than 2 TB and BTRFS is way too young (and unproven) to be considered in the server world. I hope that these data can help you to chose the right filesystem for your workload.

Modern operating systems are very complex piece of software: the provide many integrated functionalities, wrapping them onto a very nice desktop environment. In the latest years, advancements in computer speed let developers concentrate more on the realization of a consistent, easy-to-use and very attractive graphical user interfaces (GUIs), with many eye-candy effects turned on by default.

However, while in the latest years aggregate computer speed (throughput) has steady increased more or less at the rate predicted by the Moore law (note: this law primary refers to the number of integrated transistors, but I see it used everywhere as a performance meter also) single thread performance (latency) did not have this radical performance increment. This means that complex, mostly single thread applications as the GUI and, by extension, the entire graphical stack have to be very careful about their speed: if the user feel the GUI and/or other graphics slow, he will have a quite bad experience using the machine.

So, a fast and responsive GUI and graphics system are of the utmost importance for a comfortable use of a desktop computer. On the other hand, this kind of applications are quite hard to program, because the hardware give you no facilities to create nice graphics. So, while in the old days (20 years ago) GUIs and OS graphics were programmed in assembly language, today the hardware is abstracted under various libraries levels.

In the latests years, there was considerable ferment in the Linux community regarding which filesystem is best suited to accomplish its goal – to organize your files. In endless discussion, we can read about the alleged superiority of one filesystem over another one, however often these statements lack some objective data points to prove that.

As a typical Linux user can choice from a plethora of very different filesystems, I would like to give you some number to compare them. In this article, we will focus on the performances of these filesystems:

• ext3, which was the “standard” Linux filesystem since almost a decade;

• ext4, the high-anticipated ext3 successor;

• xfs, an high-performance filesystem originally developed by Silicon Graphics, Inc. for the IRIX operating system;

• btrfs, a new, next-gen filesystem developed with scalability in mind.