Tuxera Claims NTFS Is The Fastest File-System For Linux

[Wyatt]

The Linux v. Windows...err, I can't really call it a "debate"...discussion? Let's go with that. Well anyway, can we let that die or move it elsewhere?

I'm much more interested in Tuxera's prestidigitation. To be sure, I'd love to know what's going on here, and why what he said in the mailing lists doesn't seem to match up with what they're publishing on their very own website (which no one here seems to have visited?). To wit:

http://www.tuxera.com/products/tuxera-ntfs-embedded/performance/

Immediately, we see they're rather fond of embedded markets, so that answers that question (again). More pressing, as we all no doubt expected, they reveal very little about their testing methodology.

• They claim their NTFS driver consumes less CPU time than Ext3, but not by how much. It could easily be within margin of error, even.
• They make cursory effort at showing multiple test profiles, but the final test is mostly worthless because of the contrived synthetic nature of a RAM disk and 100% CPU time.
• They say nothing of the number of runs, have no error bars, and don't specify what kind of operation is happening. Marketing graphs have a tendency to be misleading in my experience.
• They make a lot of words about what they're doing, but they gloss over the technicalities. What version of NTFS is supported? Are these tests run with all features enabled?

From the outset, I find the LKML posts kind of immature, but I'm truly mystified that the information Tuxera themselves have provided as marketing material doesn't line up with the claims that have been made by a developer(?).

We simply don't have enough information to even make a hypothesis.

[smitty3268]

Yes, back on topic

I highly suspect they get this good performance by sacrificing safety guarantees. That was the first thing i thought of when i read that it was due to postponing writes to the drives. What happens when the power goes out before it gets written?

Furthermore, what kind of tests did this involve? Writing? Reading? Large files, small files, random access, streaming? It's unlikely that any single FS would prevail in all types of use. I've noticed that in Windows, at least, NTFS seems to really struggle when given tons of small files. It's a pretty safe bet that the numbers being thrown around here were a cherry-picked demo designed to show off the best numbers for his file system as possible. After all, he is trying to sell the thing and make money off it.

[kraftman]

Linux is also used as a base for firmware in a lot of consumer electronics, e.g., nowadays many modern television sets have ntfs support, so users can share movies between their PC and TV through a USB drive.

This makes sense. However, I prefer ntfs-3g or similar to some proprietary module.

[kraftman]

Which site, which browser/version? Websites exploit the browser, and sometimes if the browser uses OS libraries, vulnerabilities in the libraries themselves.

A lot of us want to believe you, but you gave us nothing to go by.

I have never heard about Linux going down this way. You have already explained how this works. This browser is usually IE and I guess it's using OS libraries. I'd love to know why there are no the Windows like viruses on Linux and OS X. There are some trojans which are simply programs that users run. Didn't you know it's sometimes enough to connect to the internet or LAN and you can be infected this way on Windows? That's why all this talk about its security is a bullshit.

[kraftman]

I have and it doesn't.

Listen up fanboys: it's not 1998 anymore. Windows is reasonably secure, even for regular users (I don't get calls from my mom asking me to fix a virus anymore since she's been on Vista with MS Security Essentials). Windows has a modern kernel and a modern filesystem. All of these tired arguments from the Windows 3.1 and 95 era need to die. They needed to die years ago.

Sorry fanboy, but it seems it's still in 95 era in some terms. I have to deal with infected windowses very often. The security is still damn weak on Windows. Oh, Linux, OS X, *BSD, Solaris - I can say all of these operating systems have modern kernels. When Windows will die the arguments (which are still true) will die with it.

[kraftman]

and i wish it were still true that Linux had a big security advantage. It's not. Those saying it does are like the Mac fanboys claiming that the lack of viruses prove how superior OSX security is, which is a joke.

Do you know any OS X virus that infects system when user connects to internet? I've only heard about trojan horses on OS X and nothing more.

[AlbertP]

Windows' lack of security is mainly caused by the huge amount of Windows viruses out on the net. Vista/7 are quite secure, but because there are so many viruses, they can never beat Linux.

Some people call an OS secure based on the number of infects. Other people based on the difficulty a virus has to infect a system.

[kraftman]

Windows' lack of security is mainly caused by the huge amount of Windows viruses out on the net. Vista/7 are quite secure, but because there are so many viruses, they can never beat Linux.

Some people call an OS secure based on the number of infects. Other people based on the difficulty a virus has to infect a system.

Exactly. However, I'm mainly interested in the ways systems can be infected and I don't judge Windows security basing on number of viruses - it's much more popular, so there are more viruses. In that case I consider it's fair to judge by basing on design and security policy. In example Win7 compared to Vista:

Windows 7 takes a step backwards by relaxing the restrictions enforced by UAC to make installing and running legacy programs easier, but at the cost of security.

[kokoko3k]

8 thread pages with many offtopic replies, and all based on speaking instead of a true benchmark.
Isn't all that useless?

[kebabbert]

Dubios in Enterprise

Sure, this thing is fast. But is it safe? Probably not.

Either something is slow and safe (lots of checksums and other safety calculations) or it is fast and unsafe (bypass all checks, no time is spent in doing checks). I dont see Enterprise use something that is unsafe. What happens if you cut the power, will the entire filesystem be corrupted then? No, I would not trust my data on such a filesystem. And you also consider that already the normal NTFS is unsafe and might corrupt your data:

http://www.zdnet.com/blog/storage/how-microsoft-puts-your-data-at-risk/169

And in this driver, they have bypassed even more safety than in normal NTFS? How unsafe is it?


Most probably, the developer can also modify ext4 according to same principles, and make ext4 faster (but unsafe). But, I would not use an unsafe version of ext4. I prefer slow and safe, rather than fast and maybe corrupted data. I would think: is my data silently corrupted now? How can I know that my data is still intact? There is no way of knowing that! (see the link above). In the link, there is a PhD thesis talking about silently corrupted data.