Slashdot Mirror
Red Hat DB = PostgreSQL - Confirmed
WeaselOne writes: "Okay, it's on the record. This Cnet story cites an Aberdeen analyst as confirming that the Red Hat database will, in fact, be PostgreSQL. Also has some interesting stuff about why they didn't just partner with Great Bridge."
--
Knowledge is, in every country, the surest basis of public happiness.
Hear Hear.
.. Q L". These were technical people, engineers and project managers, so I can't imagine trying to sell a customer on this, even if it did save them $5K over MS-SQL.
The branding point is really critical because, face it, PostgreSQL has really lousy name.
In a largely agnostic development shop, it seemed as if everyone had at least heard of MySQL and had certainly heard of RedHat, but were stumbling over "Post-Gres
"RedHat Database" (or "RHDB") might not be the greatest name either, but it's better than what they've got. (No offense to the Ingres, Post, and SQL people... Maybe change the name back to "Ingres" - some people remember it and it's easy to say)
--
Business. Numbers. Money. People. Computer World.
Just a note about the PHPBuilder.com article. Mysql has a problem with 30 concurrent users. This is a bit different than 30 users. 30 people can be connected at one time and mysql doesn't even sweat ... but when you get 30 concurrent queries, it starts to fall apart.
... if you get all your hits all within 6 hours think low end of the scale ... if it's pretty well spread out across the day, think high end of the scale). As most mysql users point out, mysql isn't meant to be used in super high load situations and this assessment backs that fact up.
For a pretty normal website, 30 concurrent connections can get you 500,000 - 1.5 million hits a day (depending on your traffic patterns
Now, I use mostly postgresql where I work (I need the transaction support and mysql didn't have it at the time). But I thought the current user point should be brought up.
All they are going to do is wrap a customized installer around someone else's hard work.
...
Gee, like they do with the kernel, X, mozilla, samba, kde, and gnome? Like what every distro does? Why should this be any different?
They're adding value to customers who want a DB with support. This is open source, they can do this, and I bet $20 that the Postgre team is happy to be part of the package, since this will expand their market and get more people interested in postgre
RedHat wants a DB. RedHat likes the GPL. There are good, GPL'd DB's available.
Where's the justification for reinventing the wheel?
PostgreSQL is PostgreSQL. If they don't intend to fork PostgreSQL, then why give it it's own name (Red Hat Database)? I'm not sure I understand the incentive here.
In the business world, what counts is service, support and reliability. RH is not selling software. They are selling service and expertise. PostgreSQL is a reliable but complex piece of software that requires knowledgeable people to support. That's where RH comes in. I am sure they have hired competent people who can help small to medium size businesses set up and operate their databases. Good move on RH part IMO.
ODBC, ADO, Oledb being de facto standards?!? Not in this world, they don't play even a tiny role. Only some client-side tools use ODBC via a JDBC-ODBC bridge, but such tools are highly discouraged (we rather generate CSV output from websites to push data into Excel, and don't want nor need ODBC. Writing back from a shaky product like Excel into a database of any importance is not allowed anyway). The only things that count are DB2 on the mainframe (with Corba interfaces) and JDBC databases on the rest (some DBI too for Perl). Since SQL Server doesn't provide good JDBC support (and 3rd party JDBC drivers for SQL Server are extremely expensive) it doesn't have a chance. Some smaller projects used SQL Server, but because of difficult integration they must be replaced everywhere (with Oracle, but Postgresql could also be an option).
Sorry Redhat but most IT manangers would choose MS in such a situation.
Right now, I doubt very much that Red Hat is aiming for or expects to get "most" IT managers buying their products. What thay _are_ aiming for is to get _more_ IT managers buying their products.
That is something which this development probably will do. I have very little doubt that RH will develop nice happy GUI interfaces etc for configuring RHDB, and that combined with Postgres not being crap, will result in more people using it.
So why PostgreSQL?
Red Hat needed to choose one that existed. Building from scratch, no matter how you look at it, would provide little return on investment.
They had a relationship established, with a group of developers very familiar with the product. Is it the best choice for everything, of course not. Hence the reason we have DB2/2, Oracle, MySQL...
What does Red Hat get out of it?
They have a core group of developers in place today. Zero cost of a learning curve for them and minimal for the developers they add. In addition, they get PR. Red Hat is a public company that is making money. The PR only helps.
What does the Open Source community get?
More dollars invested by someone who can work at it full time. In addition, we do not loose access to other DB choices. The development of MySQL, Oracle, sapDB, etc. will go on regardless of the announcement, either way. This will merely allow focus on this product for the reasons stated above.
To read this as a your SQL's better than mine, degrades quickly back to the NT vs. Linux sport that shows up in some newsgroup about a billion times a day. Many database options are good. Many are good at some things the others are not. I try to equate these choices like cars, we all want to drive different things for different reasons.
My experience is that Watcom SQL/Sybase SQL Anywhere/Sybase Adaptive Server Anywhere/whatever they're calling it today is the most ANSI compliant database. After all, didn't PostgreSQL just get outer joins in 7.1?
Nonetheless, we're using PostgreSQL for our current effort because of the price and the quality. It also has an amazing collection of built-in functions, and its flexibility for implementing triggers/rules is superb, if a bit obscure on occasion.
All Your Base Are Belong To Us!!!
Guns don't kill people -- people kill people.
But the guns seem to help a bit. (apologies to Eddie Izzard)
This page is a prime example of the immense value they add. Sure, an admin out to keep on top of security mail lists and such, but how many of these programs even bother to keep security notices going back for 2.5 years? A list like (together with rpm -q) is extreemly valuable for doing a quick check to make sure you didn't miss obvious security holes.
PJRC: Electronic Projects, 8051 Microcontroller Tools
If you take a close look at the benchmarking on the InnoDB site, you'll notice their comparison against Postgresql is only measuring speeds with a since query running against the database. Likewise for their MyISAM vs. InnoDB test, and their MySQL vs "market-leading database test".
So all of these suffer from the same thing that the default MySQL benchmarks do - they're meaningless save for marketing. There are very few real world applications that are going to have serialized queries, and those that do are likely not going to require a lot of raw speed.
If you'll notice, the only test that they use multiple connections with is their scalability test, in which they make no comparison to other databases, and that showed performance dropping sharply on selects after 50 clients, and didn't even bother giving performance on more than 20 for inserts (I'd be interested in knowing whether they were just embarassing or if the server actually crashed).
And none of the tests include a mixed load of selects and inserts. And they flat out admit their methodology was different for MySQL and the contender in one test. And they ran all of it on a single server. And none of the tests include large tables, joins, transactions, sub-selects, etc, etc. The number of faults with their benchmarks are innumerable.
Part of the beauty of the newer releases of Postgresql (aside from the improvements to the query optimizer) is their new "no-locking" MVCC (Multi Version Concurrency Control) system. Selects NEVER block inserts and vice versa.
This is a _huge_ win in terms of performance in real world situations; coupled with the row (or even cell level) locking pretty much guarante better scalability than MySQL, even if running a table type that supports finer grain locking (I believe the best they do right now is page level).
I like to think of MySQL as a sports car. It'll get to zero to sixty faster than an 18-wheel truck. But if you try to put any load on it, it won't even more.
(Note - the company I work for uses MySQL for a large production database. And it has worked even for a fairly large database under a somewhat heavy load. However, it is performing adequately, not well. And there are a lot of things we simply CANNOT do (e.g. a large query against the database for statistics purposes, since the select would lock all inserts and deadlock the database. Even under normal load its not uncommon for us to see 20-30 selects locked waiting for an insert to complete).
And to be honest, even if MySQL performed half as well as their documentation implied, I find the lack of features like sub selects, select into table, transactions (under the heavily tested table types), stored procedures, triggers, foreign keys and views make it a waste of my time and a waste of the programmers time.
We have dozens of programs and scripts that all have to access the database and they all have to work around the deficiencies in MySQLs capabilities. And, to be honest, not every attempt to do so is ultimately successful, resulting in hours wasted doing cleanup work (when it actually CAN be done without killing the database.).
I don't begrudge other people their preferences, but it would take a LOT of improvement and real benchmarks to convince me MySQL is a better database.
Matt
Doubtful there will be any immediate benefit beyond simply coupling Postgres with RedHat, other than the RPM integration you mentioned. In the long run though, consider how Microsoft has put SQL Server to work for them. Pretty much every product they label "Enterprise" somehow integrates in with their database server.
As RedHat goes struggling for the "Enterprise" market, having a known database running under the hood is going to become more and more critical. Anything they can do to simplify the process of getting that database up and running quickly is also a major selling point over other solutions.
Personally, I mostly use MySQL for stuff. I'm a simpleton, and it's nice and simple. For what RedHat is looking to do I doubt there's a better choice out there than Postgres. That, and if Postgres gets a nice influx of new dedicated developers this should be great news for both RedHat and Postgres users.
The line must be drawn here. This far. No further.
Gee, like they do with the kernel, X, mozilla, samba, kde, and gnome?
I know you're not attacking RedHat here, but I felt I should note that they do a bit more than just wrap stuff together. They've got a lot of folks doing development work on a number of those items mentioned.
I don't personally use RedHat here, but I do have a great deal of respect for them as a company. From what I've seen, they appear to be very straight shooters that have managed to not only lead the Linux market, but also maintain a great deal of repectability while doing so.
The line must be drawn here. This far. No further.
Everyone here is seems to be under the impression that redhat isn't going to do anything but repackage postgres. I don't believe its true. Postgres already comes wiht redhat doesn't it?
I believe they are going to rewrite large portions of it and optimize it and probably even move some functions to a kernel level dameon just like tux. SQL server is killing the mid-level database level and its microsofts pawn to stela marketshare. Database is almost a requirement for any real bussiness app server. SQl server is taking the roll and including all proprietary tools like oldedb, ado, odbc, and a whole bunch of vb tools to take away any competion in the mid ranger server market.
We need a quick solution thats scalable fast, and cheaper. many of redhat's r&d helped make kernel 2.4 as fast as it is today. They will do the same wiht postgres. In 2 to 4 years it may be quite powerfull and competitive to SQL server today. I doubt it will approach Oracle though. If Tux is an indication of what redhat plans to do, then we will have one hell of a great database. My guess is it will be a different database then it is today in 4 years. Microsoft is going to strangle the market if we don't get our act together soon.
For the trolls who say post is postgres is postgres, is like saying linux is linux is linux. Version 1.2 of the kernel really sucked compare to what is capable in the 2.4 release. At the time 1.2 kernel users thought it was good enough and were convinced its great. Today we know its laughable. But I expect such similiar changes with postgres if RedHat is in charge of accerlating it.
http://saveie6.com/
For Redhat, the idea of building a database would be a bad for bussiness on their part. First off, they would piss off Oracle and IBM and we need their support to move linux to the enterprise. Second, Redhat finally made a profit and why develop something with no instant return? If enterprise users need something powerfull Oracle or DB2 would be there. right?
.NEt/hailstorm :-) ), cheaper, and hopefully more powerfull. I assume redhat will come out with its own threading scheme and might put some limited functions in the kernel like what they did with TUX. SQL server is faster because its in the kernel and its supprising stable because they made sure to only put the indexing routines in the kernel and not memory managment in SQL server. We need real competition before ms guts Oracle from the bottom up.
.ocx for easy development for all the VB zombies, then I will be happy.
But after reading the CNET article and figuring out what market they want to go after it makes pefect sense. Microsoft's strategy is to scare IT managers buy showing a product get better and better to make them think its going to gut competion from the bottom up. Remember Scalablity day 5 years ago? Ms demonstrated transaction server handling a million hits a day with transaction server/wolfpack on NT4 on a dual pentiumpro 200. IT managers noticed it ran well in low end hardware and they thought "wow, if NT can do this on pc hardware, think about what it can do with higher end hardware! In 5 years unix will be dead!". It worked! Overnight NT4 became a best seller and many doubted that unix would ever survive. Today its laughable but the marketing worked and they are using a similiar strategy with SQL server.
SQL Server is practically the defacto standard already in any mid-range or even department database. MS is gutting them from the ground up. IT managers still have the all ms craze of 5 years ago and Oracle/sun or db2/as400 is pretty expensive and is used for large operations and is slower in midrange use. This is why Oracle bans the use of benchmarks. Sql Server is serious competition and is getting better. Sql server is supprising stable where I work and it went down only once in 3 years!
IT managers who buy redhat use it for a quick webserver installation or cheap department server. SQL server is cheaper to buy and implement for a bussiness database and it has support for odbc, ADO, oledb which are all becoming the defacto standards. Postgres as it is today has some limited odbc support, is not scalable, hard to set it up to handle a real load. I have no idea how Rob Malda even managed to run his site with only Mysql. Contrary to poular believe here on slashdot postgres, and mySQL are great for development but suck in real bussiness situations with hundreds of users. PHPbuilder.com did a benchmark and showed mysql having trouble with intense SQL calls with only 30 users. Since database access is essential for any web/intranet, messaging, or client/server app we need to have an alternative with great odbc, ado, oledb support as well as JDBC.
The problem with SQL Server is it doesn't support JDBC (wonder why) without an expensive third party extension. Servelts are pretty standard for retreiving data on intranet/internet servers. So if you have java servlets running on an intranet server you need a real database but only expensive oracle/sun or db2/as400 is available for real bussiness use( I assume 500-1000 users)?
Microsoft's answer is c#.net or VB.net applets running on SQL.NET wich is cheaper and fully integrated. You can also save development time by writing the apps in Visual basic and that alone could save for the cost of the servers. Sorry Redhat but most IT manangers would choose MS in such a situation.
With a real midrange database there can be a more economical solution thats more fault tollerant (if connection to the web dies, the server goes down in
IF redhat can accelerate postgres, and make it scale to many processors, support oledb, odbc and
http://saveie6.com/
To these kinds of people (which are unfortunately the people often writing the checks), when they hear PostGres, the name translates to a big fat "huh?". But if you take that software, gussy it up, and slap the word RedHat on it, and you instantly change that to "oh, RedHat's database".
I think its silly that such means must be gone to, but if this gets more opensource out into the enterprise world, then I'm all for it. Once people adopt the platform and the bumbling populace is more 'edumacated' about the Linux world, people will wake up and say "hey, I can just use PostGres!".
Damn the rusty wheels of corporate progress.
Yes sir, we folks at Red Hat, known for the Red Hat Linux Distribution, RHCEs and enterprise-level solutions, are now packing with every forthcoming Red Hat distribution a Red Hat configured database solution for low- to middle-tier database needs. The RHDB will be specially customized for Red Hat systems, which are put out by Red Hat, the most prominent Linux distribution on the market today.
You laugh, because it seems obvious here, but I remember taking a look over at a Microsoft story on MSNBC, and found no less than 75 instances of Microsoft branding, using "Microsoft" or "MS" or "Windows". And this was for one story, on one web page.
Any new avenue that Red Hat can use to drop its own name is an opportunity to solidify mindshare. Entering the database arena with a core product that won't embarass them is a no-brainer. They've even got GTK+ experts in-house that can make the thing look pretty.
This could be a SERIOUS hit to MS Access and SQL server. And with Postgresql functionality built straight into php, the whole MS-IIS-ACCESS/SQLS-ASP combination can be easily matched in terms of quality, performance and reliability by Linux-Apache-RedhatDB-PHP combination, totally surpassed in terms of cost, and only lagging behind a little bit in the gui department. From a marketing standpoint, it makes the solution LOOK more cohesive (even if it isn't), and that's always been the selling point of going with an MS solution -- smart move by RedHat. And with the GPL on their hands, we can trust them not to be sluggish and proprietary in terms of responding to the community's needs -- good move for us.
--------
Bleah! Heh heh heh... BLEAH BLEAH!!! Ha ha ha ha...
Seriously, why not use sapdb? We know it's pretty robust, perhaps too much so? It seems like too good a product to just let it sit unused.
Bwahahaha. You must be kiddin' me. Once upon a time MySql was better - but only in terms of performance. But - MySql has been plagued by the fact that they never implemented foreign key support into their DBMS. The MySql documentation speaks of foreign keys as if they are a bad thing. Which is of course a joke because everybody who has studied DBMS's seriously, knows that foreign keys are an integral part the relational database model as it exists.
Now, I ask you, why would the MySql designers occupy such a religious statement?
Is it because they would never be able to implement foreign key support afterwards into a static design, where you just could not possibly fit it in without a rewrite of the core DBMS logic?
Mysql is mostly a way of storing your data in some kind of simple hashed form, accessible with SQL queries. That's why mysql is so fast with queries.
But I think this was never the meaning of the relation model. I can store my data in my own hashed format, and my queries would be just as fast, or faster. Mysql is just a packaged piece of software representing this simple kind of storage, which answer to only the most simplest of SQL queries.
Take postgresql on the other hand: foreign key support, transaction roll-back support, ability to plug-in your favourite query language (be it based on tuple or domain calculus or relational algebra like sql), etc etc etc. With the added bonus that PostgreSQL *smokes* mysql on every available benchmark that you can find, it's clear which is the better option.
After reading the story and other information in the past concerning this development, I don't understand what redhat is exactly trying to do here. Will redhat package the db on its on and sell the packaging + support? (like it does with linux) Why do this? Why not just increase distribution of PostgreSQL (i.e.: a download page on its site and add a couple of marketing dudes), add pay-for-support and put some caffeine drinking pizza eating coders on PostgreSQL? Seems as though RH is taking this too far. If they are going all the way, perhaps RH should use some stock to buy out one of the above mentioned PostgreSQL distros.
"There ought to be limits to freedom"
I am very happy to hear this. I write a database agnostic product and PostgreSQL is the only one that is consistantly ANSI compliant.
I'd like if someone could help clarify what is exactly happenning though. Great Bridge employs half the Postgres developers.
Half the core committee, please, not half of the whole community. (BTW, I'm one of the half, so take my comments accordingly.)
They claim to have a "commercially QA-tested" version of Postgres -- what exactly is this?
What it says. Great Bridge tries to make sure that releases it packages are solid. Some Postgres releases are, um, not so solid, despite the best efforts of the community. GB adds another layer of testing and double-checking.
Does anyone have any more information either on what Great Bridge's or Red Hat's versions of PosgreSQL have over the normal releases?
I can't speak for RedHat, but Great Bridge's policy is to ship recognized Postgres releases. GB offers a nifty graphical installer (also open source, but it's not in the standard Postgres distribution), and the CD carries a number of related tools, but there's nothing there you couldn't get off the net. What GB adds is knowing what you want and which releases you want of it.
Based on RedHat's past track record, I expect that they're going to do largely the same sort of things with Postgres as GB is doing: ship releases they believe are good, along with selected tools that complement it. This will be competition for Great Bridge, no doubt about it. But I'm not in fear of being on the street soon. As database specialists, GB should be able to continue to win customers over RedHat's generalism. Besides, RedHat choosing Postgres increases the credibility of the project all round, and should increase the size of the customer pool that we will all be splashing in. I'm optimistic that it's a win-win situation.