Linux中Interbase用作大型数据库性能如何(100分)

Linux中Interbase用作大型数据库性能如何
准备用Interbase6.0，有没有哪位大哥用过？

没有人用Interbase6.0么?

听真正的高手说，10G没有问题

注意:Interbase 不是大型数据库.能处理大数据量不一定就 是大型数据库.
是综合因素.简单的表关系,并发访问少,就是数据量大点IB也没问题.

好像关心的人不多 []
我自己查到了一点点
【这是一位台湾朋友对IB6与MS-SQL7.0之间作的比较】
Sorry to write in English, because I didn't install Win98 Chinese version.
It seems like not much comparison(比较) between MS-SQL 7.0 with Interbase 6.0.
Since Inprise annouced Interbase to be Open Source, I recently(最近) really have a chance to work more on the IBX component (Interbase Express).
The last project I did for a customer:
1. Delphi BDE MSSQL Driver, with MS-SQL 7.0
2. 1.5 GB of Source Data + 2 GB of processed data = 3.5 GB
3. HP LC3 Server With 4GB HDD (By Using NT4.0 Compress Directory to reduce(减少) data size)
4. 256 MB RAM
5. PIII 350 Single Processor (处理器)
6. 10/100 base Switch
Now the month end takes 2-3 hours to finished the processing ! It seems to me the only way to increase the performance is to upgrade the Hardware !
As MS-SQL 7.0 support(支持) simultenous access to multiple Harddisk, Load Balance between multi-Processor, Cached Index to larger volume of Memory. With all these features, it seems like my hardware configuration isn't powerful enought to power up MS-SQL 7.0.
But lets be mroe reality(真实) ! we can't propose the customer to buy such a hardware, it takes 2 to 3 times expensive(花费) !
After I try on Interbase Express, it was great !
MS-SQL 7.0
===========
Using BDE / ADO Insert 1,000,000 records take more than 90 minutes Delete 1,000,000 records take more than 100 minutes, the delete operationdo
n't using stored procedure ! (Only when using Truncate is much more faster !) When Start a transaction, commit as soon as possible, the longer I start, the slower it became !
Interbase
==========
1. Disable Force Write for the Database
2. Database Page Size 4096 KB
3. IBX With TIBSQL Direct Access
** All the testdo
ne under 1 Start and Commit Transaction !!
When Insert 1,000,000 records using Local Interbase, 8 min 42 sec !
When Insert 1,000,000 records using stored procedure 3 min 13 sec !
When Delete 1,000,000 records using TIBSQL in Client-Server 48 Sec !
When Edit 1,000,000 records by Stored procedure 3 min 47 sec !
With this few test, I regain confidence(恢复信心) on Interbase, and they even support cascade（层叠） delete and update. The trigger（触发器） we can write code before or after (MS-SQL 7.0 Only Support After Action) Action.
For the first time I use MS-SQL 7.0, I find it easier to modify the table and store procedure, but with Interbase, I must plan it carefully, any changes to the table, stored procedure or any meta-data that need to recreation（消遣、娱乐） will not be allow（允许） if some-how any relation（关系） that current （当前） metadata（元数据V定义存储在数据库中数据的形式的数据,可认为是 关于数据的数据） refer（提交） by other metadata, this is consider（考虑） the most trouble some thing when I use interbase.
But when I plan out my design properly （适当）, this problem has been minimized ! And I am sure as Interbase go Open source, there will be tools to overcome（克服） this problem.
MS-SQL 7.0 5 users (USD 1428) (Malaysia Selling price)
Interbase 6.0 (As far as I know - Free) (Open Source Licience Agreement)
MS-SQL 7.0 can handle very large volume of data (which is more than Terra byte), support replication, OLAP and full text search.
But for my requirement(需求) ->
1. Data volume up to 100 GB
2. User connection up to 100 users
3. Month End process for 100 GB of Data within 1 hour
I am sure Interbase is more than enough !

关键是维护!LINUX和InterBase的维护人员都不好找,如果客户有专人进行维护最好,
如果没有就比较难办了!

linux还不能算大型数据库！
interbase 6在win和linux下面运行都不是很困难，不过效率真的就一般了

多人接受答案了。