Why use SQL Server 2005 and not Oracle 10g?  
Author Message
Lmurraz





PostPosted: Mon Mar 05 11:58:22 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g?

I am proposing to *rewrite* an existing SQL Server 2000 (upgraded from SQL
Server 7) DB App with an VB6 front end.

My company is slightly biased towards Oracle 10g than SQL Server. What
would be the arguments for using SQL Server 2005? I think for this app, they
both can do the job equally well.

In terms of development efforts, I felt SQL Server 2005 would be quicker,
but how can I quantify this other than the following:
- 0% re-use with migration to Oracle
- additional effort of copying data from SQL Server to Oracle to do
reconciliation validation against existing app

SQL Server205  
 
 
Greg





PostPosted: Mon Mar 05 11:58:22 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g?

>I am proposing to *rewrite* an existing SQL Server 2000 (upgraded from SQL
> Server 7) DB App with an VB6 front end.
>
> My company is slightly biased towards Oracle 10g than SQL Server. What
> would be the arguments for using SQL Server 2005? I think for this app,
> they
> both can do the job equally well.

Better question they need to ask is "why Oracle?"

At this point other than non-Windows platforms, feature for feature the two
(Oracle and SQL Server 2005) are about the same.

I agree, taking an existing 2000 application and upgrading to 2005 is
probably FAR simpler and may be as simple as simply moving the database to
2005 and not having to do any code changes.


>
> In terms of development efforts, I felt SQL Server 2005 would be quicker,
> but how can I quantify this other than the following:
> - 0% re-use with migration to Oracle

Right there, a BIG win. If you have to rewrite code, you have to re-debug
it.

> - additional effort of copying data from SQL Server to Oracle to do
> reconciliation validation against existing app



--
Greg Moore
SQL Server DBA Consulting
sql (at) greenms.com http://www.greenms.com


 
 
patl





PostPosted: Mon Mar 05 12:12:32 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g?




> I agree, taking an existing 2000 application and upgrading to 2005 is
> probably FAR simpler and may be as simple as simply moving the database to
> 2005 and not having to do any code changes.

Can someone quantify these??

> > In terms of development efforts, I felt SQL Server 2005 would be quicker,
> > but how can I quantify this other than the following:
> > - 0% re-use with migration to Oracle
>
> Right there, a BIG win. If you have to rewrite code, you have to re-debug
> it.

Unfortunately, in our case this is not as strong an argument, because the
existing SQL Server structure and a lot of stored procs are not well written!

Also, I read up that both SQL Server 2005 and Oracle 10g integrate with the
Microsoft .NET Common Runtime? Does it mean, I can say for now, code every
stored proc for Oracle 10g in .NET (c# for instance), and the switch to SQL
Server 2005 is just a matter of changing a config file??
 
 
jbellnewsposts





PostPosted: Mon Mar 05 12:20:58 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? Hi Patrick



>
>

>
>
> > I agree, taking an existing 2000 application and upgrading to 2005 is
> > probably FAR simpler and may be as simple as simply moving the database to
> > 2005 and not having to do any code changes.
>
> Can someone quantify these??

The SQL 2005 upgrade advisor should tell most of the things that are needed
to upgrade the system.

John
 
 
Aaron





PostPosted: Mon Mar 05 13:30:17 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? > Unfortunately, in our case this is not as strong an argument, because the
> existing SQL Server structure and a lot of stored procs are not well
> written!

Okay, so if you move "crap" to SQL Server 2005, it's not going to suck any
less, and you can move right on to . If you need to migrate to Oracle,
you're going to have to combine the migratin of T-SQL to PL/SQL *and* try to
improve on the code itself.

> I can say for now, code every
> stored proc for Oracle 10g in .NET (c# for instance)

I highly recommend not doing that in any case. CLR is very powerful and at
the same time very dangerous. You do *not* want all of your database logic
buried in the CLR. There have been many discussions about this in the
groups if you want to peek at the archives.

--
Aaron Bertrand
SQL Server MVP
http://www.sqlblog.com/
http://www.aspfaq.com/5006


 
 
Aaron





PostPosted: Mon Mar 05 13:30:17 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? > Unfortunately, in our case this is not as strong an argument, because the
> existing SQL Server structure and a lot of stored procs are not well
> written!

Okay, so if you move "crap" to SQL Server 2005, it's not going to suck any
less, and you can move right on to . If you need to migrate to Oracle,
you're going to have to combine the migratin of T-SQL to PL/SQL *and* try to
improve on the code itself.

> I can say for now, code every
> stored proc for Oracle 10g in .NET (c# for instance)

I highly recommend not doing that in any case. CLR is very powerful and at
the same time very dangerous. You do *not* want all of your database logic
buried in the CLR. There have been many discussions about this in the
groups if you want to peek at the archives.

--
Aaron Bertrand
SQL Server MVP
http://www.sqlblog.com/
http://www.aspfaq.com/5006


 
 
LinchiShea





PostPosted: Mon Mar 05 14:36:05 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? I agree with Greg that the key question to ask is whether there is any
compelling reason to migrate to Oracle since it's already on SQL Server. In
addition, what is the skill set of the folks who'll be
upgrading/migrating/supporting the database.

Linchi



>
>

>
>
> > I agree, taking an existing 2000 application and upgrading to 2005 is
> > probably FAR simpler and may be as simple as simply moving the database to
> > 2005 and not having to do any code changes.
>
> Can someone quantify these??
>
> > > In terms of development efforts, I felt SQL Server 2005 would be quicker,
> > > but how can I quantify this other than the following:
> > > - 0% re-use with migration to Oracle
> >
> > Right there, a BIG win. If you have to rewrite code, you have to re-debug
> > it.
>
> Unfortunately, in our case this is not as strong an argument, because the
> existing SQL Server structure and a lot of stored procs are not well written!
>
> Also, I read up that both SQL Server 2005 and Oracle 10g integrate with the
> Microsoft .NET Common Runtime? Does it mean, I can say for now, code every
> stored proc for Oracle 10g in .NET (c# for instance), and the switch to SQL
> Server 2005 is just a matter of changing a config file??
 
 
patl





PostPosted: Tue Mar 06 03:34:18 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g?



> I agree with Greg that the key question to ask is whether there is any
> compelling reason to migrate to Oracle since it's already on SQL Server. In
> addition, what is the skill set of the folks who'll be
> upgrading/migrating/supporting the database.

Currently most (except myself) available resources are believed to be Oracle
10g+.NEt skilled. Someone suggested to me that migrating from SQL Server
2000 to SQL Server 2005 would be about 25% quicker than migration to Oracle
10g, and I wonder if I could quanitfy/validate that, other than the effort to
migrate from TSQL to PL/SQL?



> >
> >

> >
> >

> > Also, I read up that both SQL Server 2005 and Oracle 10g integrate with the
> > Microsoft .NET Common Runtime? Does it mean, I can say for now, code every
> > stored proc for Oracle 10g in .NET (c# for instance), and the switch to SQL
> > Server 2005 is just a matter of changing a config file??

What DB logic do I want to put in TSQL and what do I want to leave in .NET??
 
 
Arto





PostPosted: Tue Mar 06 04:25:44 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g?

> Also, I read up that both SQL Server 2005 and Oracle 10g integrate with the
> Microsoft .NET Common Runtime? Does it mean, I can say for now, code every
> stored proc for Oracle 10g in .NET (c# for instance), and the switch to SQL
> Server 2005 is just a matter of changing a config file??

I guess (I have read some articles but not tried myself) they do not
differ much on simple cases (ok, classes are not same, but still). But
why use C# on simple cases when you can use T-SQL / PL-SQL ? And on more
complex cases (making your own datatypes etc.) differences are quite big.


--
Arto Viitanen, CSC Ltd.
Espoo, Finland
 
 
patl





PostPosted: Tue Mar 06 05:09:26 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? If industry practices and MVP recommendations) is *not* to use .NET code
within DB Logic, then the competitive advantage SQL Server 2005 has over
Oracle 10g on .NET integration (e.g. In process vs. out of process) is
meaningless, isn't it?

I was just thinking if .NET code was used in DB Logic, it may help prevent
people being tied in to either SQL Server or Oracle??




>
> > Also, I read up that both SQL Server 2005 and Oracle 10g integrate with the
> > Microsoft .NET Common Runtime? Does it mean, I can say for now, code every
> > stored proc for Oracle 10g in .NET (c# for instance), and the switch to SQL
> > Server 2005 is just a matter of changing a config file??
>
> I guess (I have read some articles but not tried myself) they do not
> differ much on simple cases (ok, classes are not same, but still). But
> why use C# on simple cases when you can use T-SQL / PL-SQL ? And on more
> complex cases (making your own datatypes etc.) differences are quite big.
>
>
> --
> Arto Viitanen, CSC Ltd.
> Espoo, Finland
>
 
 
Aaron





PostPosted: Tue Mar 06 06:59:39 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? > If industry practices and MVP recommendations) is *not* to use .NET code
> within DB Logic, then the competitive advantage SQL Server 2005 has over
> Oracle 10g on .NET integration (e.g. In process vs. out of process) is
> meaningless, isn't it?

Nobody is saying don't use .NET. Like cursors and GUIDs, you just need to
use judgment when to use them, and not apply them like a blanket.

> I was just thinking if .NET code was used in DB Logic, it may help prevent
> people being tied in to either SQL Server or Oracle??

That's not quite the primary strength of .NET integration.

A


 
 
patl





PostPosted: Tue Mar 06 09:04:40 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? What is the primary strength of .NET integration?

What would be the additional efforts required of (improving and) migrating a
SQL Server 2000 DB to Oracle 10g over (improving and) upgrading to SQL Server
2005, besides the migration from TSQL to PL/SQL?


> Nobody is saying don't use .NET. Like cursors and GUIDs, you just need to
> use judgment when to use them, and not apply them like a blanket.
>
> > I was just thinking if .NET code was used in DB Logic, it may help prevent
> > people being tied in to either SQL Server or Oracle??
>
> That's not quite the primary strength of .NET integration.
>
> A
>
>
>
 
 
Aaron





PostPosted: Tue Mar 06 09:48:06 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? > What is the primary strength of .NET integration?

Doing things you *can't* do, or at least not very efficiently, in SQL. For
example, string parsing, regular expressions, user-defined aggregates. This
does not comprise of re-writing all data access logic to be in C# instead of
SQL. This will kill your server, because there are many areas where .NET is
*less* efficient than SQL (pretty much anything set-based).

> What would be the additional efforts required of (improving and) migrating
> a
> SQL Server 2000 DB to Oracle 10g over (improving and) upgrading to SQL
> Server
> 2005, besides the migration from TSQL to PL/SQL?

Don't know, haven't done it (thankfully). But I don't think anyone can
argue that SQL Server 2000 -> Oracle migration is going to be more work and
provide more obstacles than SQL Server 2000 -> SQL Server 2005 migration.
Whether or not you plan to actually improve the quality of the code itself
should be a separate issue and decided upon independently. If you really
think Oracle is going to save your application, then by all means, got for
it.

--
Aaron Bertrand
SQL Server MVP
http://www.sqlblog.com/
http://www.aspfaq.com/5006


 
 
jrpm





PostPosted: Tue Mar 06 14:37:20 CST 2007 Top

SQL Server >> Why use SQL Server 2005 and not Oracle 10g? As someone that lives on both Bill's and Larry's worlds, it is always easier
to upgrade within one world than to cross over. I have gone so far as to
upgrade a 2000 to 2005 and found it almost too easy.

But for a SQL Server to Oracle migration, there are a whole lot of factors,
some psychological/cultural, some technical. What it, to me, boils down to
is the T/SQL vs PL/SQL. Not the value of the two langauges, but the amount of
re-write needed. If you are fluent in both, then the volume. If you are not
fluent in both, which do you know best?



> > What is the primary strength of .NET integration?
>
> Doing things you *can't* do, or at least not very efficiently, in SQL. For
> example, string parsing, regular expressions, user-defined aggregates. This
> does not comprise of re-writing all data access logic to be in C# instead of
> SQL. This will kill your server, because there are many areas where .NET is
> *less* efficient than SQL (pretty much anything set-based).
>
> > What would be the additional efforts required of (improving and) migrating
> > a
> > SQL Server 2000 DB to Oracle 10g over (improving and) upgrading to SQL
> > Server
> > 2005, besides the migration from TSQL to PL/SQL?
>
> Don't know, haven't done it (thankfully). But I don't think anyone can
> argue that SQL Server 2000 -> Oracle migration is going to be more work and
> provide more obstacles than SQL Server 2000 -> SQL Server 2005 migration.
> Whether or not you plan to actually improve the quality of the code itself
> should be a separate issue and decided upon independently. If you really
> think Oracle is going to save your application, then by all means, got for
> it.
>
> --
> Aaron Bertrand
> SQL Server MVP
> http://www.sqlblog.com/
> http://www.aspfaq.com/5006
>
>
>