Re: troll
Not a technical support thread at all: it's an off-topic cross-posted debate (titled "troll") about whether open-source databases will displace commercial ones. Participants argue over PostgreSQL, Firebird and MySQL versus Oracle, DB2 and Informix, touching on R&D funding, partitioning/parallelism and optimizer features needed for large data-warehouse workloads, and predictions that Informix will fade after the IBM acquisition. No question is posed and no resolution is recorded — the thread ends as opinion exchange.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
> What would you call http://www.enterprisedb.com? Sure, they haven't > reached "commercial release" status now but they definitely look like > a commercial vendor to me! A number of companies are packaging postgresql, Netezza is one that has a 'data warehouse appliance'. But I haven't yet seen any of them really spend any cash to publicize postgresql, rather than their own product. Maybe that will start changing? I hope so - postgresql is the best open source database out there. And good enough that I think db2, oracle, and informix are in deep trouble in 5-10 years.
On 29 Jul 2005 06:19:00 -0700, "Buck Nuggets" <bucknuggets@yahoo.com> wrote: >I hope so - postgresql is the best open source database out there. And >good enough that I think db2, oracle, and informix are in deep trouble >in 5-10 years Please dream on. Informix has been bought by IBM and will disappear soon. DB2 will never get any substantial marketshare on any non-IBM platform. Oracle will always have more R & D funding at their disposal, and will be capable to have postgresql stay where it is now: in the toy department. -- Sybrand Bakker, Senior Oracle DBA
> Oracle will always have more R & D funding at their disposal, and will > be capable to have postgresql stay where it is now: in the toy > department. And there's no way that a free web server like Apache or a free browser like Firefox will ever compete against Microsoft's R&D. Or free languages like PHP, Python, and Ruby will ever compete against Sun's R&D. Unfortunately, the mythical man month prevents you from getting a linear productivity improvement from a linear increase in staff. So, Oracle might have a thousand people doing R&D - but probably only a dozen can work productively on the engine. What are the rest doing? Building XML extenders? Integrating purchased ETL software? Integrating java? Big deal. Face it - in spite of a wide collection of defects and limitations, mysql is picking a ton of steam. The fact that it easily corrupts data (allows invalid dates, truncates strings & numbers without warnings), has terrible performance for most application, etc - none of this matters. Many people don't know, don't care, and just buy it. And Oracle has lost their revenue. To a *toy* database. Even worse for Oracle - Mysql won't be a toy forever and postgresql isn't a toy now. Postgresql right now is sufficiently mature for use in fortune 100s. However, since it can't handle the most demanding tasks, these companies typically require multiple products - like Postgresql plus DB2/Oracle/Informix/Sybase/etc. And since the cost of maintaining two separate skillsets generally exceeds the licensing savings, it's usually (in my opinion) worth it for large companies to forego the occasional license savings and stick to a consistent commercial product line. Today. But small companies, with small data requirements? They're generally fine on postgresql right now.. And with every passing year - between Moore's Law, and the rapid evolution of these products we're going to see them absorb the smallest databases - until the only databases still using large commercial products will be the rare monsters. I suspect that this is one of the reasons that Oracle desperately needs to become an applications company.
"Buck Nuggets" <bucknuggets@yahoo.com> wrote: >But small companies, with small data requirements? They're generally >fine on postgresql right now.. Firebird (ex Interbase), my favourite toy db, can quite happily run 300GB db's with 100's of concurrent users. >And with every passing year - between Moore's Law, and the rapid >evolution of these products we're going to see them absorb the smallest >databases - until the only databases still using large commercial >products will be the rare monsters. >I suspect that this is one of the reasons that Oracle desperately needs >to become an applications company. A very cogent analysis, and one with which I agree. People are starting to wake up and smell the coffee, for the *_vast_* majority of apps, one doesn't need the power/capabilities of an Oracle/DB2/SQL Server/<insert expensive db of choice>. This IMHO is part of the reason that we're seeing M$oft releasing it's db engine for free for small apps, and the free db's are starting to nibble at the heels of the big boys. When every machine is 64 bit, the reasons for choosing a paid for db will really start to evaporate, and as you say, only the "monsters" of this world will run paid-for dbs. Paul... -- plinehan __at__ yahoo __dot__ __com__ XP Pro, SP 2, Oracle, 9.2.0.1.0 (Enterprise Ed.) Interbase 6.0.1.0; When asking database related questions, please give other posters some clues, like operating system, version of db being used and DDL. The exact text and/or number of error messages is useful (!= "it didn't work!"). Thanks. Furthermore, as a courtesy to those who spend time analysing and attempting to help, please do not top post.
"Buck Nuggets" <bucknuggets@yahoo.com> wrote: >I hope so - postgresql is the best open source database out there. And >good enough that I think db2, oracle, and informix are in deep trouble >in 5-10 years. Have you looked at firebird? Paul... -- plinehan __at__ yahoo __dot__ __com__ XP Pro, SP 2, Oracle, 9.2.0.1.0 (Enterprise Ed.) Interbase 6.0.1.0; When asking database related questions, please give other posters some clues, like operating system, version of db being used and DDL. The exact text and/or number of error messages is useful (!= "it didn't work!"). Thanks. Furthermore, as a courtesy to those who spend time analysing and attempting to help, please do not top post.
> Have you looked at firebird? I haven't worked with firebird yet, but everyone that I've met who has really seems to like it. Must be something to that. Firebird and Postgresql seem to be the two worthwhile opensource & free options. There's certainly room for two good products here. Especially since MySQL and MAXDB are neither free, nor very good (aside from marketing). > Firebird (ex Interbase), my favourite toy db, can quite happily > run 300GB db's with 100's of concurrent users. Sure, but the most common database of this size would be a data warehouse/data mart used for business intelligence. And the queries typically used in this type of application are table scans. Informix/db2/oracle can rely on partitioning to avoid scanning entire tables, split the work across cpus via parallelism, manage the joins in the fastest manner due to robust optimizers that understand start schemas, and even rewrite the sql to use summary tables when possible. The net result is that one of these huge databases can run the 300 gb database on tiny hardware, whereas an tiny opensource database would require a huge piece of hardware to get the same results. I'm sure that this won't be the case forever, but until the open source databases get some of these features and capabilities they'll never compete well with their commercial alternatives in the 100+ gbyte range. That's why I'm assuming that db2/oracle/informix have about 5-10 years.
On 30 Jul 2005 07:02:50 -0700, "Buck Nuggets" <bucknuggets@yahoo.com> wrote: > >That's why I'm assuming that db2/oracle/informix have about 5-10 years. Your assumption is incorrect. Does any of those 'free' products have any R & D budget to it's disposal? The moment an R & D budget is needed, those products won't be free anymore. Quite the same as what you see with Redhat and the likes. This means commercial products will always be lightyears ahead, and they will also have the power to start selling scaled-down versions. -- Sybrand Bakker, Senior Oracle DBA
Sybrand Bakker wrote: >>That's why I'm assuming that db2/oracle/informix have about 5-10 years. > Your assumption is incorrect. Does any of those 'free' products have > any R & D budget to it's disposal? The moment an R & D budget is > needed, those products won't be free anymore. Quite the same as what > you see with Redhat and the likes. > This means commercial products will always be lightyears ahead, and > they will also have the power to start selling scaled-down versions. I think this is an interesting point. I've yet to clearly see where the open source community blazed the technological trail. Thsi boils down to teh question how much research is funded (not necessarily exceuted) by companies. If those companies loose their source of income they won't fund more research. So there is a self limiting factor here. And no "tabbed browsing" is not technology ;-) I do see open source moving into the commodity areas. As long as there is a bleeding edge commercial software has it's place. Call it RAC, GRID, autonomous computing, queuing, advanced XML, information integration. I think the commercial database vendors will stay ahead of the open source crowd for the foreseeable future. Comparing marketshare between the faction in this environment is another can of worms. Just my 2 cents canadian Serge -- Serge Rielau DB2 SQL Compiler Development IBM Toronto Lab