Samba's Tridge clusters code and crowds

Author of and contributor to the Samba file server, Andrew Tridgell, talks to Computerworld about Samba and his upcoming Linux.conf.au talk

Andrew "Tridge" Tridgell, the man behind the Samba file server and self-confessed TCP packet molester talks with Dahna McConnachie about Samba and some of his other favourite pastimes. He also discusses the recent deal that Microsoft signed to release previously secret code and data to the Samba project. Tridge will be speaking more about clustered Samba at the upcoming linux.conf.au.

Microsoft signed an agreement late December to release previously secret code and data to the Samba and the open source community. What will this mean for Samba, and how much will it speed up development time why/why not?

This should help speed up development of Samba quite a lot. It is very time consuming to work out all the details of a new area of these protocols by hand using protocol analysis off the wire, so being able to instead just read a set of docs is a big time saver!

How much will the agreement result in the increased interoperability between Windows and Linux/Unix?

For the sections of the protocols we already implement, Samba already has quite good interoperability. Where it will make more of a difference is allowing us to expand into new areas where we previously lacked features that Windows servers have. For example in the area of Active Directory replication, and remote indexing.

How much difference will it make to your work?

I'm expecting it will make a big difference! It will take years to implement all the new stuff we found out about in these documents though. It is nearly 15,000 pages of very technical documentation, and just reading it takes a very long time!

Can you elaborate a little on what your linux.conf.au talk will be about?

With the addition of CTDB, Samba can now be setup as the most scalable NAS solution available. I've been working on this for a couple of years, and talked about an early prototype last year, but it now has reached the stage where it is ready for production use, and we have some large production customers using clustered Samba for real work.

With clustered Samba you have the opportunity to replace dozens or even hundreds of traditional large NAS boxes with a single cluster, which makes management of huge data sets much easier.

Page Break

What are the major benefits that clustered Samba using CTDB offers?

High availability, so when a node fails it just keeps going, plus high scalability, so you can have tens of thousands of users on one Samba server.

What difference could CTDB make to the enterprise?

Enterprises usually have hundreds of NAS boxes, which makes data management a huge headache. Now they can have one super NAS box.

Who should consider using CTDB and why?

People with large data sets who need a highly available, very fast storage server for Windows or Unix clients.

Apparently you enjoy molesting TCP packets. Would you say this is your favourite activity? Does anything else come close, and if so what?

Yes, I love looking at the details of network protocols on the wire! I do love other things too (Sailing, travel, bushwalking etc), but it's great having a job where I get to do one of my favourite things.

You are a very popular speaker at the Linux conference. How many times have you spoken at the event and why do they keep asking you to return?

I have given at least one talk every year. I love giving talks, and I usually find something new to talk about, so maybe that's enough to keep people interested.

For more information on CTDB read: http://ctdb.samba.org/ and also http://samba.org/~tridge/ctdb_movies/ which shows you some movies of CTDB in action.