Dumping SQL Server

  • Dump Truck

    This article on open source databases is from last year, giving you an idea of how long it takes me to write something :), but it's still interesting.

    Are open source databases, primarily MySQL and PostgreSQL, as reliable as SQL Server? Is the technology just as good?

    I think this is an opinion that depends on which side of the fence you're on. Most die hard Open Source advocates would say that their databases are not only equal, but better. Hard core SQL Server, Oracle, or DB2 DBAs might claim the proprietary solutions are miles ahead.

    I really like SQL Server and I've made a career of it. However I'm not really bigoted against the other platforms. I've looked at MySQL and PostgreSQL in the past (years ago) and they were both good products, they worked well, but they were also immature. Nowadays, MySQL 5 has added stored procedures, transactions, triggers, and a whole bunch of other features that I'd consider essential. There's even a parallel cluster server, which reads like what I'd want in SQL Server. A series of servers I can add or remove from a cluster and have all the servers appear as one to the users. Personally I think SQL Server is making a huge mistake not going in this direction, but maybe that will change with Windows Compute Cluster. PostgreSQL v8.2 is similar, with many of the same features and arguable a slightly more mature product.

    I do think that SQL Server has more features and extra subsystems that give it an edge in many situations. However I think if you are building a strictly OLTP system, any of these products would work. So which one do you choose?

    The one your staff is most comfortable with and has the most experience with.

  • The open source databases have become very solid. I think the choice now is driven by culture more than technology. There is a lot of the Microsoft culture out there and a lot of corporate types feel safer since they they know the can't get fired for buying Microsoft. Plus Microsoft tries their best to wrap everything up into a shiny visual studio package to get you to use the whole stack.

    On a pure technology basis, I've been very impressed with MySQL and have only fooled with Postgres just a little but it seems very solid. If you look at sites with very high technical demands, you'll find a lot of sites running on MySQL and running very well at much lower cost than SQL Server for the same level of performance. To me MySQL is aimed more at the SQL Server type crowd and Postgres is more at the Oracle crowd.

  • There's no way we'd dump SQL for anything else, we've pretty much tied ourselves in (much to MS pleasure im sure).

    I do agree with Steve that I wish MS would develop something like a parallel cluster for SQL Server, it would mean that we could stop spending 100K on a decent server and just buy smaller ones to spead the load onto.  Load balanced SQL Servers would be ideal.

  • For work stuff, SQL Server is where it's at - mostly because that's what they have established. However, personally, I could bounce around to any of the database servers, depending on what resources I have available and what my packages need. My personal site uses MySQL, but that's what my blogging package needs. And over the years, I used to think PostGreSQL was the better choice for bigger programs, but MySQL has grown up quite a bit and I'm impressed to see MySQL 5.

  • I personally have worked with PostgreSQL and quite enjoyed it. There were a few growing pains at first, but eventually it just became a rock solid platform for us to develop against. This was in a a Small business where they didn't have the cash to pay for per processor licensing for SQL enterprise. Actually they didn't give us a budget period. Due to MySQL's (at the time anyhow) lack of transactions sprocs etc we chose PgSql and were very happy. We were able to take an old decommissioned server toss redhat on it, load up PgSQL and the thing just flew.

    At my current job I agree that I'd never be able to move from a MS platform because it's just too ingrained, but if I were starting another development scenario from scratch PgSQL ( and more likely EnterpriseDB) would definitely be in the mix of products we'd look at.

    To help us help you read this[/url]For better help with performance problems please read this[/url]

  • As a DBA for several platforms at a hospital I don't have a choice as to which platform to use but the vendors we purchase solutions from are overwhelmingly using SQL Server.

    We only have 2-3 Oracle, 2-3 Sybase, 1 DB2, no open source, and close to 50 SQL Server solutions.

    Why do the vendors choose this platform?

    In my opinion it is because it is by far the easiest to set up and maintain and Microsoft provides better tools to interface with the server.

    Microsoft may not be the best under the hood but they certainly make it easy to do the things most people want / need to do.

    I've installed MySQL a couple of times over the years to see what it was like and it was not even in the same league as SQL Server when it came to installation, setup and maintenance.

    As another example: it is taking two DBA's and a system programmer to upgrade DB2 from version 7 to version 8.  The upgrade is broken down into multiple steps (compatability mode and then new function mode) which requires two system outages.  I can upgrade SQL Server on my own in a matter of hours with relative ease.  Granted DB2 may be able to outperform SQL Server for monstrous tasks of gargantuan size but most of the database apps I've seen are not that huge, so who needs the headache?

     

  • There are two areas needing to be addressed here and some have hit it right on.

    1. Business requires stability, performance, and technical support. Many businesses have a very low risk tolerance and leaning towards stability is predominant.

    With this in mind not many businesses will businesses will risk their business critical system on a new or unproven product. A new and or different development environment is moderate to high risk and incurs cost, in that new developers and support teams are required to retain, or to be replaced. This comes at a cost in dollars and in productivity.

    A higher risk is to change the database or the backend. If a new approach to how you get to the data is risky, how much so changing to a new place where the data lives. To release a RDBMS without stored procedures, triggers, and other programming on the data server is irresponsible and unacceptable to the small to large IT shop or business. It is almost as if you would like to have a car that has a frame, tires, and an engine knowing that this will run and take you somewhere. However if the car has no method of setting direction, no way of stopping, no instructions and no mechanic in the area knows enough to work on it you have something of very elevated risk.

    Most businesses are not ready to 'risk it all' on new unproven software. They operate with the principle of 'you get what you pay for' or 'you do not get something for nothing'. I can not blame them.

    2. Personal use does not dictate stability, performance, or technical assistance. Many DBA, developers and other IT folk love a challenge and see the use of this class of product as cheap and ready to use. Plus if there are problems with it all the more fun.

    Personal users do not care about the level of risk in many cases. They do not have deep pockets where price is not an object. And the idea of getting it done at any cost does not enter into the process at all. If it is fun, cheep, available and 'cool' then bring it on.

    There are probably a few other areas or categories that others can identify but these are the main players that come to mind.

    As usual have a great day!

     

    Not all gray hairs are Dinosaurs!

  • Miles,

    Great comments and the risk issue is a big one that people forget about. I have some friends that are whizzes with open source, especially the LAMP stack and they'd love to convert clients over. And it would probably run great. But what happens if they leave? Or they get tired or want a change or anything?

    A business has to be able to find people to run things and the fact is it's "easier" to find Microsoft people and it's easier for them to run the systems. There are relatively fewer DBAs that work on Open Source, and that's a legitimate concern. It's changing all the time and there are more and more people working with these systems. I wouldn't worry about finding a desktop or server admin to run Linux or Apache, but MySQL or PostgreSQL is another matter.

  • A lot of my projects involve using MS-Access to connect to Sql Server, but AFAIK there is no equivalent open source version of Access. I've looked at Kexi, Rekall and Base; they can import Access data into their own respective formats, but without pass-thru queries and VBA they're not viable solutions, so for the time being, Windows is the only platform choice in my world, and IMHO Sql Server is the best Windows db.

  • I think that both MS SQL and MySQL are great products. I have never tried postgreSQL.

    I feel that currently MS SQL is an easier system to use and manage with their Management Studio as long as the system is in-house. For remote site I administer it is much simpler to use MySQL because of PhpMyAdmin. Microsoft does have a remote administration tool which they built for SQL 2000, but it isn't near as nice or speedy as PhpMyAdmin. MS SQL has many more features which can be taken advantage of and are much more likely to be needed if the company is big enough to have their web infrastructure in-house.

    These opinions have been my current experience and are likely to change the more I learn.

  • I would think majority of the crowd is experienced with SQL Server or Oracle when it comes to databases. This may be due to agressive marketing technique from the parent companies. In addition there are trainings, certificates, awards if you become "experts" in using these. And we all would like to have those on our resume .... But the open source is lagging in that area.

    My understanding is when companies like Yahoo Inc, Google Inc. use the open source databases (check here http://fe13.news.sp1.yahoo.com/s/nm/20070425/bs_nm/ibm_mysql_dc) the product must be stable and secure. Definitely its not a product developed just by college student.

    As said above, one more drawback to open source is getting the right documentation or right help.

    When the open source gets right partner for marketing its excellent features and if it succeeds in attracting people like us definitely the corporate will go for open source product... why to pay so much for license

    Steve... thanks for the good topic. 

  • Hello all,

    I recently worked with a company that created insurance company policy management and claims software.  Our platform was SQL Server.  It was the second generation on the product.  The prior iteration was Foxpro. 

    Given years of history with SQL Server, 2000+ stored procedures, and a tight integration with the core software product there is no motivation for a migration.  Even if open source has some significant benefits those benefits would have to be awesome to justify the cost to migrate and the time taken away from the core development cycle.   

    Even if open source solutions offered the same scalability and performance, for existing solutions there is no clear benefit to switch and there is significant risk to product stability and customer perception.

     

    Keith

     

  • Use the right tool for the job. I've avoided using MySql because of the lack of referential integrity. With Version 5 they solved that problem as well as most other things that would have kept me away. Sql Server still pays my bills. On the other side of our house where the development is done under Red Hat they use Postgres. Some of our web servers use a MySql back end. That seems to be the niches that these databases fit into. My Inventory system would not overtax any of them at the current level but integrating the bar code system is much easier because it uses WinCE.


    Kindest Regards,

    Scott Beckstead

    "We cannot defend freedom abroad by abandoning it here at home!"
    Edward R. Murrow

    scottbeckstead.com

  • ( I've avoided using MySql because of the lack of referential integrity.)

    MySQL have got DRI(declarative referential integrity) for years but you have to pay for it in INNODB file, so with a 26 pages long algebra MySQL cost you money. lol 

    On a serious note I use Oracle so I can be employed, my race and sex are not  issues, it is what I know.

     

    Kind regards,
    Gift Peddie

  • What I want to know is the story behind that dump truck.


    James C Loesch

Viewing 15 posts - 1 through 15 (of 16 total)

You must be logged in to reply to this topic. Login to reply