News Stay informed about the latest enterprise technology news and product updates.

Samba expert digs into 3.0 upgrade

Samba Team co-founder John H. Terpstra dances you deep into the Samba 3.0 upgrade.

Some key features make Samba 3.0 a better mousetrap than Samba 2.0 for Windows-Linux file and print sharing, says Samba Team co-founder John H. Terpstra, who adds that it's too bad the open source community can't come up with a better mousetrap for building and selling Linux desktops. In this interview, Terpstra points out some of the most useful new capabilities of Samba 3.0 for enterprises. Then he explains how and why the open source community is failing to persuade companies to abandon their Windows front ends.

Terpstra is co-editor, with Jelmer Vernooij, of a new book on the latest release of Samba, titled The Official Samba 3 How-To and Reference Guide. The book, due out in November, has been co-written by Samba Team's core developers and end users. It will be published by Prentice Hall PTR as part of Bruce Perens' Open Source Series line of books.

What are some core differences between Samba 2.0 and Samba 3.0?

John Terpstra: Samba 3.0 offers some really exciting new functions, like the ability to seamlessly integrate native and mixed-mode Active Directory for file and print sharing purposes, offering a sign-on for Unix and Linux clients into AD.

Administrators should look at the Unicode support in Samba 3.0, which gives them globalization support. This really was a lot more difficult to do with Samba 2.0. Samba 3.0 gives you stable Windows 2000-, Windows 2003-type networking protocols, which is very important for people living on the bleeding edge of Microsoft technology.

Also, an administrator can have multiple password back ends and multiple virtual file system modules, so that there's now a recycle bin on your Samba network drives. There is no need to ever lose a file again after it's been deleted.

What's new with Samba 3.0 in security?

Terpstra: We have more extended auditing facilities there now. We have secure channel support, so we have much more secure networking protocols. Samba 2.0 does not do secure channel work, but Samba 3.0 does. There's a lot more, too, but that's a key new capability.

What new Samba 3.0 features will be of value to Linux administrators and users?

Terpstra: There are many very subtle changes that are of interest to users of both Windows and Linux. If you are using Windows on an access control list-enabled platform, regrettably only Mandrake and SuSE are access control list-enabled out of the box. With Samba 3.0, you can assign extended-access control list users and groups on your network resources using Windows tools. You can use the server tools to manage your users and groups from within Windows, even though they are on a Samba domain controller or Samba server. For the Windows administrators there are a lot of goodies that are new to that environment that give far better controllability and far finer-grained access.

For the Unix administrator, Samba 3.0 provides the ability to implement LDAP as the back end and provide a truly distributed and much more scalable architecture than was possible with Samba 2.0. The multiple-password database and the virtual file system technology are of interest to Linux users because it gives them tools that allow them more control.

Speaking of the back end, I've talked to several CIOs who are moving to Linux on the back end but are sticking with Windows in the front office. Do you see that happening a lot?

Terpstra: Yes, and I see some risks there. Microsoft could change the SICS protocols on the Windows cards, which would immediately invalidate the use of Samba. That is what worries people who use Samba the most. I don't know that Microsoft is doing that, but it is a concern.

I think that one of the key things that keeps Microsoft on the desktop, apart from the close connection between Active Directory and Exchange, is Microsoft's office automation software. At this point, I believe that those who have played with and some other office automation programs are not convinced that they have the level of seamless interoperability that they are looking for. I think that's an inhibitor. I don't know that that's going to change soon.

That said, I can say Windows desktop interoperability with Linux via Samba 3.0 is excellent, better than it's ever been.

Do you think that advocates for enterprise usage of Linux desktops are jumping the gun, promoting this option before it's really enterprise-ready?

Terpstra: That's a difficult question to answer. Difficult in that I'm equally at home in both worlds, and if I were a dedicated Microsoft shop I would not find a compelling argument to change. The question is: How tightly are you coupled to that Microsoft desktop and the Microsoft Office automation tools?

There is not perfect interoperability between OpenOffice and Microsoft Office. For some companies, that little bit of difference is still a big, big obstacle. I think that what we really need to start seeing in the Linux world or in the open source world is a very much more compelling argument based on an alternative way of solving the desktop problem.

For those who use Linux, Samba is a great tool because it removes the barriers to interoperability for file and print services. It allows you to integrate your Unix and Linux systems and the data that's stored on them with Microsoft Windows desktop systems and data.

What will it take for the Linux desktop to gain wider acceptance and more enterprise adoption?

Terpstra: What I'm looking for from the Linux vendors is a more determined consumerist mentality, a willingness to deliver a consumer-oriented solution. The Achilles' heel, I believe, of the open source community is that it is not consumer-oriented.

I've done some research and asked open source developer community members, 'Why are you creating software?' and they say: 'Because I have an itch, and I scratch it.' In other words, they do it to solve their own problems and not someone else's problems. Here's an example: If I, as a user of open source technology, say that I want a certain open source technology, the community's typical response is, 'Well, get off your butt and do it yourself. You've got the source code, so go and do it.' If I respond to this by saying I'm not a programmer, their answer is, 'Well, learn.'

That kind of attitude is not going to convince corporate America to shift their desktop to an open source technology. Open source developers and companies need to deliver desktop products and server products at the request of users; those products must be easy and intuitive to use. Until they do, we can make all the arguments in the world that open source software is better and more stable, and that we've got better technology and offer freedom, but it's just not going to cut it.

FOR MORE INFORMATION: news exclusive: "Samba upgrade enables simple NT 4.0 migration"

Ask the experts

FEEDBACK: What's your Samba implementation like?
Send your feedback to the news team.

Dig Deeper on Linux servers

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.