Sun Microsystems has thrown its weight behind PostgreSQL, announcing plans to distribute and support the open-source database.
Sun Microsystems has thrown its weight behind PostgreSQL, announcing plans to distribute and support the open-source database.
It’s good to see that a reputable, respectable company is starting to put its eyes on FLOSS. It makes it look better in the eyes of companies.
Plus it’s a great database that deserves support.
Finally i think they are doing something right. I like Java. Perfer PostgreSQL over MySQL. And OpenSolaris looks promising. ( They have a new paper about ZFS, their new file system. Dont know if any expert would mind giving some insight into it )
So is sun finally getting back on track?
That’s a pretty similar post to what I was going to write…with all the boneheaded decisions they’ve made lately, backing the best OSS (IMO) RDBMS is a no-brainer. I wonder if they’ll make any contributions to the project…they have some pretty talented engineers over there.
This kind of endorsement pretty much secures Postgre’s future in the OSS realm…I hope to see entry over on TPC soon, which would seal Postgre’s viability as an alternative to Oracle/MSSQL.
“They have a new paper about ZFS, their new file system.”
Arstechnica has a quick write-up on it:
http://arstechnica.com/news.ars/post/20051117-5595.html
I cut my teeth on Sun/Solaris workstations. This is the first time in a decade I’ve gotten seriously interested in their products.
Did this just happen after Oracle said they would support Sun’s OS as the one they would prefer to support their Database on?
Sun has been making noises about distributing PostgreSQL for at least a couple years. This is only the culmination of that.
http://www.theregister.co.uk/2003/04/28/mcnealy_on_project_orion_su…
http://www.zdnet.com.au/insight/software/0,39023769,39180150,00.htm
http://www.computerworld.com.au/index.php/id;116679278;fp;16;fpid;0
As I understand, they were looking at Ingres, however, due to the ‘who/what the hell is that’ factor, I think they weighed up the features of both and probably decided that they were better off going with PostgreSQL.
NOW the big question, are the going to start integrating their existing software and PostgreSQL together – their clustering and grid software are two that come to mind when I think integration; also, is SUN willing to put their muscle behind PostgreSQL and implement all the SQL standard features as well as making it as scalable as Orale, Sybase and SB2.
Admittingly, Oracle and DB2 run on the back of their respective names, however, hopefully people will be willing to give SUN’s offer a chance – just because a product from Oracle costs millions, doesn’t actually make it a good product, in some cases, it just means that you’ve just been had by some really good marketing and smoothing talking by a sales representative.
Postgres doesn’t need Sun ruining it, thank you. Postgres is already a lot more ‘SQL standard’ than Oracle, in fact Oracle is probably the LEAST SQL standard complaint of all the DBs. Postgres is already pretty damn powerful and scalable, the only thing it really lacks is
1) fancy OLAP, data mining, manipulation, etc tools.
2) it’s not Oracle.
the only thing it really lacks is
1) fancy OLAP, data mining, manipulation, etc tools.
2) it’s not Oracle.
3) Multiway replication, and systems for making distributed solutions fully transparant with regard to storage location and transactions.
How will they ruin it? if they’re contributing things to the project and there are still the current maintainers in charge, I see nothing wrong with SUN providing ‘extra hands’ in terms of programming – also in terms of scalability, SUN already has the reputation and knowledge required to know technologically and inregards to customers, to make products scale (marketing them, well, thats a different situation – they seem to be like DEC, great products but crap marketing; a company run by engineers).
SQL wise, I am referring to the latest additions to the standards, PostgreSQL shouldn’t go saying, “well, another product is even crapper, there fore, that justifys the current status quo here” – it should be, “this is where we are, who cares about the competition, we’ll keep pushing forward for 100% SQL compliance regardless of where our competitors maybe in regards to compliance”.
More high quality free software getting more legitimacy and support. More choice. More quality. Thanks Sun.
It’s not as if you can’t compile PostgreSQL on Solaris, or even get a Sunfreeware PKG, Sun bundling it with the FOSS version of their OS doesn’t make my day, especially as their only motivation can be to turn people away from Linux!
If they could lend some of their developers to the project and give it some new non-Solaris-specific features then I would be pleased, but otherwise Sun are just using the FOSS community to make Solaris look better.
Now a FOSS DB2 client for Solaris (that’s not written i Java!) *THAT* would make my day…
“It’s not as if you can’t compile PostgreSQL on Solaris, or even get a Sunfreeware PKG, Sun bundling it with the FOSS version of their OS doesn’t make my day”
It still means they will contribute by having people working on postgresql, which could mean a considerable boost in its developement.
Perhaps, perhaps not. PostgreSQL is BSD, and therefore, they are under no obligation to contribute those changes.
Hopefully, they will. I’ve been maintaining a Binary installer of PostgreSQL for OS X for the past year or so ( just finished building up a Universal Binary version of 8.1.0 ). It is an outstanding engine, and it just keeps getting better. There is no question that Sun’s energy and money would go along way towards further improvements.
That’s not the important part about the announcement though.
Sun’s name still carries of good bit of legitimatacy. Sun’s formal announcement of support of the product itself has more value in marketing and mindshare than any money they throw at the project. This is the kind of step that it takes to get to the next step, where people realize that PostgreSQL can compete with MSSQL and Oracle on the low end, and really only starts to not hold it’s own in massivelty scaled implementations…
The point is that Sun has said they will support PostgreSQL… that’s a big deal.
So if you are a business, and want to depend on PostgreSQL you can get a support contract from Sun.
As for new features…. give it time, who knows how things will turn out.
I got this meail today form the postgres announce mailing list.
It does look promising…
Here is a copy of what was sent:
——————
Community members:
Today, Sun announced that they would be supporting and shipping PostgreSQL as a part of Solaris distribution. The press release should be here very soon, if it’s not already:
http://www.sun.com/smi/Press/sunflash/web_sunflash.html
This should come as no surprise to the hackers in our community who’ve noticed an increasing number of @sun.com addresses on the pgsql-hackers list in the last few months. Sun’s support for PostgreSQL will involve:
help in optimizing for Solaris; DTrace; Solaris packages; sponsorship of community members and projects; and direct contributions/advice from Sun
engineers on database performance. More later as this stuff actually gets rolling.
I think almost everyone here will agree with me that Sun is a very welcome addition to our community of corporate supporters.
—
–Josh Berkus
Josh Berkus
PostgreSQL Project Core Team
http://www.postgresql.org
(all opinions expressed are my own; I do not speak
for the Project unless specifically noted.)
—————————(end of broadcast)—————————
TIP 5: don’t forget to increase your free space map settings
… do not fit the “throughput computing” paradigm implented by the UltraSPARC T1 (aka Niagara) processor.
Oracle on an UltraSPARC T1 system would give poor bang for the buck.
Carsten
Oracle will be pleased.
Seriously though, it’s a real no-brainer decision. Postgres is a quality database with a lot of great features, and it is perfectly sensible for Sun to include it in order to make their products look and actually be better. It’s what Red Hat has been doing for years, and it could really spell some trouble for Oracle. Finally, a reason to use Sun and Solaris!
I don’t mean to be a troll, I am very pragmatic in regards to what software I use, sometimes is FOSS sometimes is Propietary. I prefer what suits best to the situation.
I really do think that Open Source is here to stay, and I like PostgreSQL a lot, IMHO is the best Open Source DBMS. That is why I am worried about the meaning of this news. If they are just going to support PostgreSQL then it is OK.
But if they are going to mess with the way PostgreSQL does it job, then we have a problem. I don’t understand why people like the backing of a big corporation (Novell, IDM, Sun, etc…), I guess they don’t see the big picture which is that corporations support FOSS because is good for THEM not for FOSS. Just to point out an example, see the mess Novell has done with SUSE.
I fully understand that big name backing is necessary to access the enterprise market, but if that means following IBM/Novell/Sun agendas, then I believe the trade-off is too risky.
But if they are going to mess with the way PostgreSQL does it job, then we have a problem. I don’t understand why people like the backing of a big corporation (Novell, IDM, Sun, etc…), I guess they don’t see the big picture which is that corporations support FOSS because is good for THEM not for FOSS. Just to point out an example, see the mess Novell has done with SUSE.
You DO realize that Fujitsu has been funding and working on PostgreSQL for quite some time. Fujitsu is HUGE company.
Also, recall that pretty much EVERYTHING done in FOSS is done “for THEM”, THEM being whomever it is that is doing the work. Those that are doing the work are doing it because they a) enjoy it, b) scratching their own itch, c) getting paid for it or d) all of the above.
Nobody works on FOSS because they don’t want too (unless there are far east coding slave camps somewhere).
So, now, the same folks that have been working with Oracle to make it perform better on Sun hardware under Solaris will be looking at Postgres to make it better on Sun hardware and Solaris. Being as Sun is solidly on track for opening up a vast majority of its source code, I wouldn’t worry at all about Postgres.
“But if they are going to mess with the way PostgreSQL does it job, then we have a problem. I don’t understand why people like the backing of a big corporation (Novell, IDM, Sun, etc…), I guess they don’t see the big picture which is that corporations support FOSS because is good for THEM not for FOSS. Just to point out an example, see the mess Novell has done with SUSE. “
Novell is not a good comparison at all. Novell is a sinking ship without a rudder and a labotimized captain. Novell has long been a sinking ship, and kept doing stupid things both before the SuSE acquisition. Novell has also shown clearly they don’t know how to use their new Linux/Ximian toy, and monetize it.
Sun, by contrast, inspite of some of the dumb things they’ve done from time to time, and in spite of their loug mouthed execs, has really, really gotten it’s act together lately, and really seems to be turning things around. This is for the following reasons:
1) Technology. Java 5.0 is a huge improvement. NetBeans and Sun Java System Application server are awesome (I’ve tried ’em out -they’re fantastic). Solaris 10 is a big improvement. Niagra looks very promising. The AMD galaxy servers are doing well.
2) Free and Open source software support. Most of Sun’s software you can get for free, at least minimal developer versions. Solaris 10, the full version, is free of cost, and only costs money for support. OpenSolaris is generating lot’s of interest, and is gradually developing a great ecosytem. OpenOffice has made big improvements. I already mentioned NetBeans and SJSAS – both completely awesome, and totally free of cost (except for the enterprise version of SJSAS). And now they’re supporting the highly regarded PostgreSQL db. All great stuff.
3) Acquisitions. Sun has made some good strategic acquisitions – the last big one being that storage company (name escapes me) which was profitable and compliments Sun’s portolio perfectly.
4) Partnerships – many with other hardware providers, and most visibly with Google. Many yawned about the Google toolbar / OpenOffice / Java download intermingling. I think it was a great move – it will ensure even more downloads and deployment of both Java and OpenOffice, as well as Google toolbar.
5) Cost cutting. Let’s face it, Sun got fat during the ’90’s dotcom boom years, and was very very slow in making necessary cuts afer the bust and after Linux started eating up the WebServer market (a former cash cow for Solaris). But Sun sold off a lot of their pricey office real estate, and made the unfortunate but needed layoffs.
Personally, I come more from the Linux camp (not that anyone needs to be in any particular “camp” – it’s all about good tech, right?), and I’ve often been a big critic of Sun, and Java (I used to hate Java, now I love it) on these forums. But Sun has most certainly changed my perception of it. They’re doing some really good stuff, while betting the farm (they had to), and I give them major props and hope for their success.
Edited 2005-11-18 21:21
It would be nice if they made it better integrated into StarOffice/OpenOffice. As it is now you can select “Connect to MySQL” and get everything preconfigured, but if you do postgresql you have to do much more yourself, e.g. such as getting the right JDBC/ODBC driver.
I was in on the press conference when Sun announced its support for Postgres on Solaris are here are the main points:
DTrace Performance Tuning
DTrace Provider
Solaris Service Manifest
Container Aware
Sun Studio Optimizations
Sun Cluster Postgres Agent
Sun Support 24×7 Worldwide
They also said that this could be just the start for other OSS databases. Postgres was a better fit for Solaris because of certain enterprise features. It will be available for Soalris X86 and SPARC.
They also said that this could be just the start for other OSS databases. Postgres was a better fit for Solaris because of certain enterprise features. It will be available for Soalris X86 and SPARC.
What’s key about this statement, notably “this could be just the start”, is that it leverages Suns goal to work not just as a manufacturer, but also as a services company. With OSS getting more and more popular, the “professional support” market is just getting started. There’s little to stop Sun from creating “experts” in different OSS projects.
But rather than having to support “everything”, they can cherry pick the popular applications that suit them and simply support those on top of Solaris. That’s why they’ll, say, support Gnome and not KDE, or Postgres and not MySQL.
If the demand for MySQL support on Solaris seems to grow, then, they might rally around that banner, but MySQL already has a well know, and formal support organization compared to Postgres (which has some companies, but it’s not “obvious”).
By dedicating resources to support particular applications, and to make them run better on Solaris, it makes the entire package that much better. If the whole package is better, someone may choose Solaris because it runs Postgres better. And if they’re running Solaris, they may well look at Sun to support not just Solaris, but Postgres as well.
Suns goal right now is to lower barriers to choosing Sun software and, indirectly, Sun support and services.
Adopting popular OSS packages makes them look just that much better.