Thread: Oracle / PostgreSQL comparison...
* Rodrigo E. De León Plicet (rdeleonp@gmail.com) wrote: > Any comments? Sure, they've never bothered to actually look at the data. Consider that for quite a while Oracle essentially refused to admit that their could *possibly* be bugs in their system (see: Unbreakable Linux, or whatever that foolishness was). They also ignored remotely-exploitable privilege escalation problems for *years* (oh, well, your databases should be behind firewalls with only "trustworthy" people who can access them directly... yeah, right). PG provides patches on a *very* consistent basis, based on need, and even more so, core works with the CVE process and provides patches and new releases accordingly (they don't just spring fixes on people..). Next, PG doesn't even use the same basic technology as Oracle regarding how transaction isolation and versioning works. Oracle using rollback segments to store 'old' rows in, while PG uses a Multi-Version Concurrency Control (MVCC) system. They're fundamentally different things, so the notion that PG is somehow a 'reverse engineered' Oracle is complete bunk. Adhereing to the same standard *doesn't* make something reverse enginered. Additionally, there's *tons* of features that are in PG which aren't in other RDBMS's, like, I dunno, *freakin' readline support*. Have you ever tried to actually *use* sqlplus on a regular basis? It's *horrid*. psql is miles ahead of sqlplus when it comes to a reasonable RDBMS client, this guy even admits that (immediately after saying PG hasn't got any features that other RDBMS's have...). Yes, you can use the rlwrap hack w/ sqlplus, but, seriously, when is Oracle going to bother investing in their principle CLI again? Never? Seems that way. Even mysql's CLI is better than sqlplus, and they own the code to both now.. I love how he finishes with the claim that Oracle "keep their finger on the pulse of where IT is headed", right after admitting that their client is actually a huge piece of junk. Thanks, Stephen
Attachment
> http://cglendenningoracle.blogspot.com/2011/06/oracle-vs-postgres-postgresql.html > > Any comments? Amusing. " What kind of support is available if we have a problem? What is their patching schedule to address exploits and known security defects? If there is a bug, how responsive is the organization to resolving the issue? These are questions that imply a need to have an organization behind your technology. In this case, clearly PostgreSQL is not an appropriate choice, because you won't get acceptable answers to these questions for any open source software. " Um. What? Let's look at this: "What kind of support is available if we have a problem?" With PostgreSQL, I can correspond with *the people that wrote the code*. They're friendly, responsive, and are very reasonable about looking at possible bugs. Every time I have emailed the community lists for help, I have received reasonable answers within a few hours (usually *under* an hour). I've never found a bug, but I've seen plenty of them squashed, as the developers themselves admit a mistake and announce a fix. With Oracle, I get to call in to a call center to open a ticket to look at the problem to assign it to a support person to return to the call hours later to assume it's not a bug to maybe file a problem report to perhaps ignore it for months/years. All for the low, low price of $XX,000/core, depending on the magic 8-ball discount or saving throw you roll with the Oracle Sales Army. With PostgreSQL, it just works. I don't have to spend a day or two or three adjusting my system to hell and high water, including symlinking libraries back and forth, ignoring system patches that "interfere" with Oracle. Interfere? I'm sorry? You're an *APP* on *MY* server. *I* do not serve *you*, Larry. It's also obvious that the author has never even looked at the commercial support available for PostgreSQL, for so much less money. Am I a DBA by profession, to answer these questions authoritatively? No, I am a hobbyist DBA for my own uses, and I manage servers for DBAs. But it doesn't take an expert to see that the author of this article has likely become too comfortable working for companies that enjoy paying a half million dollars for licensing a couple of database servers that could be done with PostgreSQL (assuming application support, of course) for a tenth of that cost (hardware included). Thank you, PostgreSQL folks. I love your software and appreciate your great support. Benny -- "You were doing well until everyone died." -- "God", Futurama
On 06/23/2011 10:28 PM, Stephen Frost wrote:
I stole some inspiration from this comment for my own response, which I just posted to the site. I'll save a copy here in case the author becomes so embarrassed by his mistakes he deletes it:
The idea that PostgreSQL is reverse engineered from Oracle is ridiculous. Just a look at the vast differences in the MVCC implementation of the two; Oracle's redo logs vs. PostgreSQL WAL are completely difference designs.
As for there being no unique features in PostgreSQL, that's completely wrong too. A good example is how deep the transactional DDL features go--Oracle has started to catch up recently, but PostgreSQL still has a lead there. The ability extend the type system and indexes with your own custom items are also better in PostgreSQL than any other database. This is why the PostGIS add-on (built using the type extension facility) is busy displacing installations of the weaker Oracle Spatial at installations all over the world right now.
As for support, there are half a dozen companies in the world you can buy PostgreSQL support from at a fraction of the rate Oracle charges for it. I routinely fix bugs in the database itself within hours of report for my customers, as part of a service contract, which is an option on top of the free community support. Because PostgreSQL is open-source, there are multiple vendors available who provide this service. With Oracle as a closed source product, there can only be one who is capable of offering this quality of support. And that single source vendor has quite a history of squeezing as many dollars out of its customers as its can. Since there is choice among PostgreSQL support companies, you'll never get into that position with it.
Next, PG doesn't even use the same basic technology as Oracle regardinghow transaction isolation and versioning works. Oracle using rollback segments to store 'old' rows in, while PG uses a Multi-Version Concurrency Control (MVCC) system. They're fundamentally different things, so the notion that PG is somehow a 'reverse engineered' Oracle is complete bunk.
I stole some inspiration from this comment for my own response, which I just posted to the site. I'll save a copy here in case the author becomes so embarrassed by his mistakes he deletes it:
The idea that PostgreSQL is reverse engineered from Oracle is ridiculous. Just a look at the vast differences in the MVCC implementation of the two; Oracle's redo logs vs. PostgreSQL WAL are completely difference designs.
As for there being no unique features in PostgreSQL, that's completely wrong too. A good example is how deep the transactional DDL features go--Oracle has started to catch up recently, but PostgreSQL still has a lead there. The ability extend the type system and indexes with your own custom items are also better in PostgreSQL than any other database. This is why the PostGIS add-on (built using the type extension facility) is busy displacing installations of the weaker Oracle Spatial at installations all over the world right now.
As for support, there are half a dozen companies in the world you can buy PostgreSQL support from at a fraction of the rate Oracle charges for it. I routinely fix bugs in the database itself within hours of report for my customers, as part of a service contract, which is an option on top of the free community support. Because PostgreSQL is open-source, there are multiple vendors available who provide this service. With Oracle as a closed source product, there can only be one who is capable of offering this quality of support. And that single source vendor has quite a history of squeezing as many dollars out of its customers as its can. Since there is choice among PostgreSQL support companies, you'll never get into that position with it.
-- Greg Smith 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us "PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books
On 06/24/2011 09:14 AM, Rodrigo E. De León Plicet wrote: > Here: > > http://cglendenningoracle.blogspot.com/2011/06/oracle-vs-postgres-postgresql.html > > Any comments? He's been working with Oracle too long, and forgotten that it's a database not a career and a lifestyle? More seriously: it's all about thinking habits and values, to the point where there's almost no point arguing. The author has a completely different view of business, risk management, product evaluation, etc. Most notably in terms of worldview, there's the strong assumption that you need a single owning organization to get support, and that the organization will actually provide acceptable support at all let alone at a reasonable price. Now, my own experience is hardly comprehensive, but I've usually had much *better* support from 3rd parties than from a vendor, particularly in areas where the vendor tries to lock you in to their own support arrangements. Almost like they don't really try if they can lock out the competition... When you're talking with someone that invested in an organization and a viewpoint, you're unlikely to get them to examine alternative possibilities. It's interesting that the author fails to cover perhaps the most important point about product selection: You should select your database based on your needs, not your ideology. You should determine your needs, then evaluate available options according to those needs. You might select even a different database for different things if your needs for those two things differ enough. For example, I prefer working with PostgreSQL, but it wouldn't be the database I'd choose if I needed a super-fast mostly in-memory sharded system for XML/JSON document storage because that's not where its strengths lie. I'm quite surprised the article didn't mention the importance of having "somebody to sue if it goes wrong", which is a comment I often see made re Oracle. It's lucky they didn't try to stress that point, because evidence of succesful suits by customers against Oracle are ... hard to come by ... and Oracle's army of lawyers makes it unlikely that suits would succeed. Despite that and the fact that a suit is unlikely to award damages sufficient to make up for your losses unless you can show negligence or willful misbehavior, people keep on talking about needing to have somebody to sue. -- Craig Ringer
> > http://cglendenningoracle.blogspot.com/2011/06/oracle-vs-postgres-postgre > > sql.html > > > > Any comments? > > Amusing. > > " > What kind of support is available if we have a problem? What is their > patching schedule to address exploits and known security defects? If > there is a bug, how responsive is the organization to resolving the > issue? These are questions that imply a need to have an organization > behind your technology. In this case, clearly PostgreSQL is not an > appropriate choice, because you won't get acceptable answers to these > questions for any open source software. > " > > Um. What? Let's look at this: > > "What kind of support is available if we have a problem?" With > PostgreSQL, I can correspond with *the people that wrote the > code*. They're friendly, responsive, and are very reasonable > about looking at possible bugs. Every time I have emailed the > community lists for help, I have received reasonable answers > within a few hours (usually *under* an hour). I've never found > a bug, but I've seen plenty of them squashed, as the developers > themselves admit a mistake and announce a fix. > I've got support and even a permanent fix in HEAD for a problem I reported a while ago in under 24 hours. Oracle would never ever do that, particularly not for free (as in beer - which reminds me, we should collect a couple bucks on this list and send a beer and pizza truck over to Tom :-)) ) Back in the day I've worked for Sun and IBM in tech support. Both companies stressed on large customers (i.e. BMW would get their own highly skilled technician stationed on site). All I can say is: "it's amazing what kind of support you can get for 3.3 Million a year" :-) Uwe
This message has been digitally signed by the sender.
Attachment
> On 06/24/2011 09:14 AM, Rodrigo E. De León Plicet wrote: > I'm quite surprised the article didn't mention the importance of having > "somebody to sue if it goes wrong", which is a comment I often see made > re Oracle. It's lucky they didn't try to stress that point, because > evidence of succesful suits by customers against Oracle are ... hard to > come by ... and Oracle's army of lawyers makes it unlikely that suits > would succeed. Despite that and the fact that a suit is unlikely to > award damages sufficient to make up for your losses unless you can show > negligence or willful misbehavior, people keep on talking about needing > to have somebody to sue. The article also forgot to mention what I call the "chair cutting problem". Most decisions in IT departments, particularly in larger companies, are based on the probability to take the grunt when something goes wrong. Therefor, no IT Manager who wants to keep his job will buy anything else than what the CEO thinks is the market leader. If you buy the market leader's product, you take no risk. Clearly you made a good choice, so only the manufacturer is to blame if something goes wrong (and we all know, sooner or later something will go wrong). I've seen this over and over in 20+ years of IT consulting. Let's buy Oracle, Microsoft, IBM - clearly that's the good choices. Well, not always, but it sure is a choice that ensures the safety of the CTO. I've had one - ONE - client who accepted my recommendation to use PostgreSQL for their not all that large custom system. Ever since we've implemented the software and taken it to production the server and software functioned flawless and they don't even have anyone to check on the server (no DBA or sysadmin there). Sure, they're relatively small with 50+ employees, but for the little money it cost they got something that has been running fine since 2003. Last time I checked the server was up close to 500 days. They didn't use my recommendation for mostly linux servers (they did well before the housing slump, but ever since are short on money). Instead they hired a "sysadmin" company - which happens to be a OEM for HP. By the end of the day they spent close to half a million to replace 2 linux servers with 8 windows servers - which don't even work as well as the old servers did (although, that may be an issue with the "sysadmin company" - they want to sell more and are probably not interested or incapable of proper system administration). Oh, well, all I can say is I'd trust PostgreSQL with my payroll any day of the week (and actually do, as these days my main business - not IT consulting anymore - is running on PostgreSQL and again has been pretty much flawless for 3+ years now - with over 100.000 transactions a day) Uwe
On Thu, Jun 23, 2011 at 7:14 PM, Rodrigo E. De León Plicet <rdeleonp@gmail.com> wrote: > Here: > > http://cglendenningoracle.blogspot.com/2011/06/oracle-vs-postgres-postgresql.html > > Any comments? That is quite possibly one of the most ignorant opinion pieces I've ever read. The janitor in Dilbert is significantly more qualified to give opinion on pg versus oracle than that guy.
Stephen Frost wrote: > I love how he finishes with the claim that Oracle "keep their finger on > the pulse of where IT is headed", right after admitting that their > client is actually a huge piece of junk. I guess that was just a typo. Shouldn't it read "[Oracle can] keep their fingers on the throat of where IT is headed"? Yours, Laurenz Albe
> I love how he finishes with the claim that Oracle "keep their finger on > the pulse presumably, he means, the jugular ... -- john r pierce N 37, W 122 santa cruz ca mid-left coast
On 06/23/2011 10:28 PM, Stephen Frost wrote: > I love how he finishes with the claim that Oracle "keep their finger on > the pulse of where IT is headed", right after admitting that their > client is actually a huge piece of junk. > Oracle is able to keep their finger on the pulse of their customers, because they have their hands where they can firmly squeeze their...uh, wallets. -- Greg Smith 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us "PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books
On Friday 24. June 2011 03.14.39 Rodrigo E. De León Plicet wrote: > Here: > > http://cglendenningoracle.blogspot.com/2011/06/oracle-vs-postgres-postgresq > l.html > > Any comments? I think he got a point in «Oracle as the second largest software company in the world» which is a killer argument from the PHB point of view. They're big because they're big. regards, Leif
Le jeudi 23 juin 2011 à 18:14 -0700, Rodrigo E. De León Plicet a écrit : > Here: > > http://cglendenningoracle.blogspot.com/2011/06/oracle-vs-postgres-postgresql.html > > Any comments? > There is a previous post by the same author : http://craigglendenning.blogspot.com/2009/03/i-fight-to-stay-focusedand-often-lose.html First two paragraphs. No further comments. -- Vincent Veyron http://marica.fr/ Logiciel de gestion des sinistres et des contentieux pour le service juridique
On Friday 24. June 2011 06.01.31 Greg Smith wrote: > The idea that PostgreSQL is reverse engineered from Oracle is > ridiculous. Maybe he believes that SQL was invented by Oracle? regards, Leif
I would just like to add my voice to those praising the community support provided by this list. I am not a DBA, and merely tinker with a few databases, mostly on the web. As such, my questions have occasionally bordered on the very silly, but I have always had them answered courteously, helpfully and with a great deal of patience, in stark contrast to MySQL lists which tend to veer towards the sarcastic, the minimal and the obscure. Kind Regards, Dave Coventry