Text 3072, 69 rader
Skriven 2005-03-10 11:33:10 av Ellen K. (1:379/45)
   Kommentar till text 3055 av Adam Flinton (1:379/45)
Ärende: Re: IDE market consolidates some more...?
=================================================
From: Ellen K. <72322.1016@compuserve.com>

He was surprised to hear how horrible Oracle was when run on a single computer,
even though it was an IBM Series 6 with 6 CPUs (for fewer than 500 users!).  
He previously had a setup with a couple hundred boxes and
100,000 users and load balancing and said it was great.   Maybe Oracle
is great under that scenario, but for 1100 users and maybe 100 GB of data, SS
is a WAY better choice.  And as I've said before, for a monster database I
would still rather go to DB2 or maybe Terabyte.

One suggestion I made was that if we go to an RDBMS other than SS, implying way
more administration time (especially if it's Oracle), I retain responsibility
for data architecture and back-end programming and we get another resource to
do the administrative stuff.   If the
database is properly designed and the box is properly configured, SS basically
manages itself.  My SS databases have each been tuned exactly ONCE, each has
over 600 users, one on a single-CPU box and the other on a dual, and
performance is thank G-d excellent.    Plus for 2005, the
Standard edition allows up to 4 CPUs and unlimited RAM and has failover
clustering -- at $6K retail list per CPU that is incredible value, especially
considering you get all the OLAP stuff too.

On Thu, 10 Mar 2005 13:59:44 +0000, Adam Flinton
<adam@NOSPAM_softfab.com> wrote in message <42305258$1@w3.nls.net>:

>Ellen K. wrote:
>> The gearshift statement is correct, but it's not a valid analogy.
>>
>> You'll probably be thrilled to hear that my boss told me today that we
>> might go to Java.  For me it wouldn't be hugely horrible since I am the
>> back-end person, but I really feel sorry for the front-end guys if we
>> do.   The only thing is that he thinks that also implies going to
>> Oracle, because five years ago at another company he found Java against
>> SQL Server to have lousy performance.  I pointed out that since then
>> Microsoft came out with their own JDBC driver for SQL Server, and
>> suggested we test Java against SS using the latest versions of Java,
>> JDBC and the MS JDBC driver, and make the same tests against Oracle to
>> see whether Oracle is really better (and if so by how much relative to
>> the huge extra expense)... and if SS performance seems too slow to at
>> least also test against DB2.
>>
>
>I take it you don't like Oracle <G>
>
>I can't say how well the SS jdbc drivers work coz I haven't used them.
>
>Adam
>
>> On Mon, 07 Mar 2005 18:42:42 +0000, adam
>> <adam_nopam_@nospammo_softfab.com> wrote in message
>> <422ca041$1@w3.nls.net>:
>>
>>
>>>Ellen K. wrote:
>>>
>>>>Development time is also a factor -- sure, you can WRITE a UI, but how
>>>>long does it take to write and debug, compared to an environment where
>>>>you can create a dependable and stable UI by drag-and-drop?
>>>>
>>>
>>>Depends entirely on if you know what you're doing. An automatic
>>>gearshift does not make a bad driver into a good one.
>>>
>>>Adam
>>
>>

--- BBBS/NT v4.01 Flag-5
 * Origin: Barktopia BBS Site http://HarborWebs.com:8081 (1:379/45)