Re: In Japan with Josh Berkus

Transcript:

<introduction>

Josh: "Can people in the back hear me? Thank you for hosting me in Tokyo, it's a lot of fun for me to come over here. It is also an extremely exciting time to be a PostgreSQL developer. It's just amazing how something that was a hobby, a sideline, um a ... thing that I and people like Bruce did in their spare time has become a major business.

So, I'm going to go over a little of where we're going and where we've been.

On November 8th of last year at the Open Source Database Conference in Frankfurt Germany, Peter Eisentraut, another member of the Postgres core team anounced the availability of Postgres 8.1. This release was a milestone for us, a highpoint in a lot of ways. From the features, from the amount of adoption and excitement that it's created, news coverage, the news coverage of Postgres in general....

Now, we didn't start out with what we have in 8.1. As you know, Postgres has been under development for a long time, over 20 years. Since my involvement with Postgresql, started in 1998, I'm just going to talk about what we've done since we went on the Internet in 1996. In fact, our 10th anniversary of going on the internet is coming up in July, and we'll be holding a small conference for PostgreSQL developers, for contributors in Toronto.

So, when Postgres, actually as Postgres95 first became available for people to download, from Postgres.org, from Postgres95.org, I don't remember what our website was called. The first goal at that time was to stop it from crashing. A lot of that work was done by Bruce, who's in the back, and by Vim, one of our key developers at the time. Now, I of course waited and joined the community after Postgres stopped crashing. After that, the next thing we had to do was implement a lot of features that were considered standard on other SQL databases. Things like left joins, the schema object, and stored procedures. Once we were good enough in terms of implementing business features and standard database features, we focused on majing the database perform better. Because most of our users at that time were telecommunications companies and internet service providers and similar companies, we focused on what's called online transaction processing. And thanks partly to the design of Postgres with a few improvements, things like the free space map, we were quickly able to make Postgres measure up to even the largest commercial databases for online transaction processing. The other big thing that started in those years, and didn't peak till recently, was the port to the Windows operating system. That port was led by Bruce Momjian and involved a lot of Japanese contribuotrs to the Postgres database. Having conquered online transaction processing, in the last year to year and a half we've moved on to data warehousing and very large databases. And a lot of the features in 8.1 and some that will be coming in 8.2 will be about data warehousing.

Now, what's coming in 2007 and beyond I'm not quite sure - a lot of it is up to you. As an open source project, we go where our users and contributors want us to go. I have a feeling that the that place is going to involve specialty database purposes and application integration. But we'll see.

So, what's in 8.1 is a whole lot of features that we're pretty excited about. This includes major SQL features like 2-phase commit and user roles, large database and data warehousing features like index bitmap scan and table partitioning, faster transactions through improved multi-processor performance, shared row locks and faster GIST indexes. We also were able to take care of a couple of things on our todo list that some of our users have been asking us for a very long time. That includes more powerful and more standard functions and stored procedures. And integrating the autovacuum utility into the backend of the Postgres database. So, 2-phase commit, which took a couple of developers about 2 years, is heavily in demand by a small group, mostly in financial processing. The concept is pretty simple, instead of simply commiting a transaction on a single server, you have two phases where both servers coordinate committing a transaction together. The way we implemented it is when one server is ready to commit a transaction it sends a prepare to commit message to the other server, the other server acknowledges that with a prepared message, then when the transaction is ready, then when it's acknowledged, the master server sends a commit message. And the second server acknowledges it. Now the tricky part is what happens when one of these servers fails in the middle of this process. For example, what happens if this server fails, well the answer is that if this acknowledgement has not been received

... 18:58 into the recording. I'll get the rest transcribed and maybe put in on the new bizgres network site at http://bgn.greenplum.com.

- Luke

________________________________

From: pgsql-hackers-owner(at)postgresql(dot)org on behalf of Luke Lonergan
Sent: Thu 2/16/2006 10:57 PM
To: Satoshi Nagayasu
Cc: Bruce Momjian; PostgreSQL-development
Subject: Re: [HACKERS] In Japan with Josh Berkus

Arigato gozaimas!

- Luke

________________________________

From: pgsql-hackers-owner(at)postgresql(dot)org on behalf of Satoshi Nagayasu
Sent: Thu 2/16/2006 10:17 PM
To: Luke Lonergan
Cc: Bruce Momjian; PostgreSQL-development
Subject: Re: [HACKERS] In Japan with Josh Berkus

Hi all,

Josh's talk is now available at:

http://snaga.org/01_Josh_Berkus.mp3

This file is very long, and an interpreter's voice
to interpret into Japanese is also recorded.

If you want to learn Japanese, please try it! :)

Thanks.

Luke Lonergan wrote:
> Drink Sake and eat some Yakitori for us folks in the west. Maybe shake
> a robot hand or two while you're at it :-)
>
> - Luke
>
> On 2/16/06 2:14 PM, "Bruce Momjian" <pgman(at)candle(dot)pha(dot)pa(dot)us> wrote:
>
> FYI, Josh Berkus and I are in Japan to give some presentations. We
> return to the USA on February 23.
>
> --
> Bruce Momjian | http://candle.pha.pa.us
> pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
> + If your life is a hard drive, | 13 Roberts Road
> + Christ can be your backup. | Newtown Square,
> Pennsylvania 19073
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: Don't 'kill -9' the postmaster
>
>
>

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo(at)postgresql(dot)org so that your
message can get through to the mailing list cleanly

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster