MESSAGE
| DATE | 2010-01-04 |
| FROM | Paul Robert Marino
|
| SUBJECT | Re: [NYLXS - HANGOUT] MySQL issues
|
On 1/3/2010 8:33 PM, Ruben Safir wrote:
> On Sat, Jan 02, 2010 at 11:03:38PM -0500, Paul Robert Marino wrote:
>
>> well the truth is there is no danger of a monopoly because Oracle is
>> dieing in the data warehouse market and has been dead in the web market
>> for years.
>>
>
> http://www.crn.com/software/220000766;jsessionid=WOFRZXO4N2RGDQE1GHPCKH4ATMY32JVN
>
> This doesn't sound like dieing and dead to me FWIW.
>
>
well that article primaraly talks about there midleware products which
does not have any thing to do with databases except for the fact they
use one for data storage. for example one of their middle ware products
keep openldap in sync with active directory and uses a clustered
database to store its configuration information as part of its automated
failover however the information in the ldap directories themselves is
kept int their respective backend's.
Nasdaq uses green plumb, my employer (which i can not name on an open
forum for leagal reasons) uses Green plumb, Netezza, and another strange
product that's a combination flat file DB and ETL product in one the
name of which i cant remember for its dataware housing.
>> Oracle cant afford to drop MySQL because its their chance to
>> get back into the web market. In the data warehouse market databases
>> like Green Plumb and Netezza both of which are base on the BSD licensed
>> PostgreSQL code base have been dominating the market along with other
>> PostgreSQL variants. the nice thing is that these companies have been
>> contributing back to the PostgreSQL code base hence the rapid
>> development of new versions over the last 7 years the latest one being
>> version 8.4.2 released on 2009-12-14.
>>
> http://www.oracle.com/database/number-one-database.html
> http://www.informationweek.com/news/software/database_apps/showArticle.jhtml?articleID=207402230
>
> This doesn't sound like Oracle is losing much in market share, although
> I will take a look at these other products, just for fun. I happen to
> hate Postges. IMO, it is a PIA to use, and so is Oracle, for that
> matter. The major thing that attects me to MYSQL is that is
> unbelievably and in a common sense way, straight forward to understand,
> set up and use. And Monty makes that happen by being always available
> all the time and by making what I believe is correct design decisions.
>
>
>
well PostgreSQL has gotten easier to use in the last couple of years
since the creation of the autovacuum daemon in the 8.x series has
automated the table space maintenance for most users. beyond that the
advanced features like using perl as a embedded PSQL Language for custom
builtin functions can be a pain to learn they are very helpful and MySQL
cant do any thing like that. and i know there is a traditional
performance debate between MySQL and PostgreSQL but they are both better
at different types of queries. MySQL likes several small queries to
return a result where as PostgreSQL likes one big query to return the
same result, as a result query sets that take seconds on MySQL take
minute on PostgreSQL and large complex queries that take seconds on
PostgreSQL can literally take days on MySQL.
>
>> the funny thing is that these
>> PostgreSQL variants cost more then the oracles closest to equivalent
>> product but they are capable of performance rates that blow all of their
>> competition out of the water. Oh incidentally the other database
>> mantioned in the article EnterpriseDB is also a PostgreSQL variant.
>> As far as Oracle's the argument to change the license that's nothing
>> surprising. Oracle was of the opinion that they made GNU/Linux a viable
>> platform when they ported all of their database products (Oracle,
>> Sybase, and Informex) over and that the GPL community owed them; however
>> the GPL community has always felt that Oracle has proffited off of GPL
>> software but never given any thing back to it.
>>
>
> Two things, the idea of a conortium around a database product isn't new.
> Sybase was developed in just such a fashion decades ago...and where is
> Sybase these days. The Postgres database is developed under the BSD
> license, and that has always been a problem, IMO, and that is exactly
> what Monty currently wants with MYSQL.
>
> As for Oracle, both Oracle and Sun have made a big error which as time
> goes forward, hurts them measurably, which is that they haven't fully
> commited to Free Software. There one foot in and one foot out
> methodology has cost them huges sales and product development
> possibilities. There dependencies on license sales is a long term dead
> end for them, as it is of the record companies. As it is, in both
> companies, support is the driving force of their profitability.
>
>
Two clarification points
the database most people refer to as Postgres is actually named
PostgreSQL. PostgreSQL is the product of several forks of an original
database that was names Postgres. Postgres was a ground up rewrite for a
PHD thesis of Ingres by the original developer of Ingres.
PostgreSQL is maintained by a consortium called "PostgreSQL Global
Development Group" of which SUN last I heard was a member.
by the way Sybase is still VERY heavily used for financial systems for
performance reasons. simply it has the best insert rate per connection
of any RDBMS available which make it good for handling near live data
stores that require history lookups. note sqllite is not included in
that because it is not a RDBMS in that it is single user during write
operations and has no network capabilities.
>
>> Oracle has also felt that
>> that the development of MySQL pushed them out of the web market and that
>> the GPL community was biting the hand that fed them in the early days.
>> further more there has been a lot of speculation that Oracle will do one
>> of their old tricks, namely that after they change the license they will
>> fork a proprietary version branded with the oracle name and let the free
>> speech portion of the code die.
>>
>> On a side note the reason why Oracle lost the web market has nothing to
>> do with price.
>>
> Uh... I disagree. Cost was an enourmous factor, as it was with SCO.
> Who the hell was going to shell out a 100 grand to develope slashdot?
>
>
You would be surprised what companies are willing to shell out for
database support. beleave me if it weren't for performance issues
oracle would have a near lock on the web market and nearly every
opensource web app would support oracle as an option instead of just
MySQL or PostgreSQL
>> The primary reason is that Oracle does not support the
>> LIMIT or OFFSET SQL directives and use subqueries in their place which
>> makes it highly inefficient for the type of dynamic partial reports that
>> websites generate constantly (any time a result is split across multiple
>> pages such as an online retailer product listing that show 25 per page).a
>>
> I agree that that is also a big issue...and not the only PIA thing
> about Oracle.
>
>
>>
>> On 1/2/2010 8:37 PM, Paul Charles Leddy wrote:
>>
>>> Ya, I am extremely divided.
>>>
>>> This is a historic event in my mind. Especially with the split between
>>> Monty/RMS vs Moglen/Groklaw.
>>>
>>> Grand history. Take your time. You don't have to choose sides. None of
>>> us may know enough. Ever.
>>>
>>> And. I think it is a learning experience for me to see that a lot
>>> comes down to where the money is, more than I thought. At least, that
>>> is the counter claim on both parts.
>>>
>>> I assume money is the root of all evil. That's besides the point, but
>>> might be a good stand to take in the end.
>>>
>>> One big question for me is about monopoly. One of the means of
>>> maintaining the social system we have -- I am reading
>>> "Turbo-Capitalism by Luttwak, btw, I don't make these things up mysqlf
>>> -- is that no one should have an outright monopoly. That's agreed. We
>>> need to at least keep up appearances. So, remind me, what was the
>>> answer to the fact that Oracle will have a monopoly if they suck in
>>> MySQL?
>>>
>>> I am also skeptical -- if you are curious to know -- that MySQL's GPL
>>> status can't save in via forking. It would be historic, right? Correct
>>> that no other huge project has forked and survived?
>>>
>>> Anyways, I think it would take a LOT of work, but Monty could do it,
>>> if he wanted, and people would go w/ him. What am I missing here? The
>>> large corporate perspective? Definitely. Enlighten me. Crazy?
>>>
>>> Can anyone clarify the dual-license deal in a paragraph w/ less than
>>> 10 sentences? I am still unsure on that measure.
>>>
>>>
>>
|
|
