So after a ever so temporary but loud moment of insanity[1] having a decision made which I very strongly disagreed with (wanting to release online encrypted backup as closed source), we’re back in the world of freedom and the MySQL Server is (and will be) free and open source software (dual licensed, so you can buy a commercial license of the same thing).

[1] Addition (wanting to remove my use of the word): Marten (rightly) points out that although appreciating the new blog posts, he doesn’t appreciate having his decisions called insanity. He’s right. It’s the wrong way to put it. So, without wanting to censor or change history (instead preferring to illustrate my own stupidity and amazing ability to completely say the wrong thing every 6 months or so), I offer this clarification (that i have tried to express in about 3 drafts of blog posts, none of which have made the light of day as i was never really happy with them): the decision was made with all the right intentions (grow the company, end up producing more free software, making sales to enterprises easier, clearer differentiation etc) but it was one that I (and many others) rather strongly disagreed with. In the end, the dicision was made to have these parts as free software and I truly believe that this was made after more arguments were presented by myself (and others) about why having these parts as closed was a bad idea. It is quite the thing to make the decision to make modules for your free software product closed, it is about 15 steps higher to go back on it. I’ll share a phrase I used a few times when being a right nick-picker about things during employment contract negotiation this year (for MySQL Australia and then Sun): “Do I trust Marten? Absolutely. It’s the next guy. Remember, SCO was once Caldera and producing a linux distro and generally considered good.” So, that was more than I intended to write on the subject… but hopefully clarifies that I just thought the decision itself was bad, and am lucky enough to work at a place that encourages discussion when you don’t like things.

So, now I’m involved with writing up the worklog for encryption for the MySQL server native online backup. I also wrote most of the original worklog for compression of online backup (I implemented compressed backup and LCP for MySQl Cluster) as well as some proof-of-concept code (written in <5 minutes at 3am while jetlagged).

There are two main approaches to encryption: symmetric and asymmetric (public key). I think we should support both (but we’ll see what others think).

For symmetric (password based for those not up with the street lingo of crypto) we’re thinking of the following algorithms: 3DES, AES, Blowfish. Are there any others that people care about?

DES is obviously out as it’s not considered secure, and really, we should be helping users to get things right.

For public key: RSA and DSA are the obvious choices.

As for libraries implementing all of these? well….. I’m thinking about libgcrypt – it looks fairly nice and a bit similar to the kernel crypto api (which seems quite nice). Anybody got any other suggestions? Things you’d like to see? thoughts?

EDIT: Server not Service. We sell services, the server is free and open source. I fail.

I’ve hacked around a bit to get PBXT to compile in tree, and pulled in the Maria engine. Both are latest source.

So, want to try out Maria?

Want to try out PBXT?

Just want to do ./configure and go with it, just like building a normal MySQL Server?

Grab the -stew tree. Source tarball here:

mysql-5.1.24-stew1.tar.gz

(it’s based on something close to 5.1.24… and I’m going to switch some of my systems over to it rather soon… already done some good benchmarks on one of my apps).

feedback much appreciated.

UPDATE: Got x86-64 Linux. Try my binary tarball (built from above src tarball). Built on Ubuntu Gutsy (my laptop). So it may (or may not) work. If it kills your squirrel, not my fault.

Last night (okay… i’m posting thsi a bit later… so the other night), a group of us gathered around to hear about some work that had been done in getting a MySQL Cluster backend for OpenLDAP.

I’d heard a bit of rumors (where rumors is defined by somebody saying something on IRC and me being busy looking at other things) about this previously, but last night was the first time I a) saw it working and b) saw performance numbers.

Disclaimer: I am no LDAP expert.

So, what is it?

Normal LDAP can replicate asynchronously from one machine to another. You can even update on both and it has some conflict resolution. But… this costs in performance.

Normal LDAP can also replicate (asynchronously) to a remote location for read-only (e.g. make authentication go faster in Australia with the main LDAP server in the US).

The MySQL Cluster backend for OpenLDAP connects directly to MySQL Cluster, using rather optimised schema, indexes and coding (directly using everything that we’re good at – which was really awesome to hear).

So, the MySQL Cluster backed LDAP server is the 2nd fastest in the world. The fastest is OpenLDAP on a single machine. With MySQL Cluster, we’re not that much slower than a single box – but we have redundancy. So that’s totally awesome. The third fastest…. much slower than us.

This was one of the most awesome new things I’ve seen here.

After many years where it’s been under debate, Monty now really says he knows what MySQL is named after, and it is after My.

Rich Green talking at MySQL UserConf.

Dot point 2: GPL.

Let’s see….

LugRadio is possibly the best Linux/Free Software related podcast out there – and it has been for a very long time. LugRadio Live is a live event where lots of people gather together to love freedom.

This year, there’s a LugRadio Live USA – and it’s in San Francisco. So, are you going to be (or are able to be) around SF the weekend before the MySQL UC? If, there is no excuse not to drop in on LugRadio Live USA – it should be awesome! (oh, and it doesn’t cost much… like $10 or something)

Found this today: Windows Core Networking : WSAPoll, A new Winsock API to simplify porting poll() applications to Winsock.

Which means we should be able to squeeze some peformance out of Vista. Although, since poll() we’ve gone and moved on to much better things in the free world.

As part of porting MySQL Cluster to Microsoft Windows platforms, I’m going and redoing a lot of the portable sockets code in MySQL. My goal is to have enforcing of proper use of the portable functions be done by the compiler instead of the programmer – leading to less portability bugs.

Let’s face it: friends don’t let friends have to run Windows to not break the Windows port.

Good news is that so far I’ve found several bugs that could have presented them in real nasty ways at run time but have instead showed up as build errors – brilliant!