Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.misc -> Re: Oracle vs. SQLServer
You didn't mention if you are referring to Microsoft or Sybase SQL Server.
These are architecturally similar but will become more different when
Microsoft releases version 7. I'll address some differences between
Microsoft and Oracle SQL Server an Oracle DBA should be aware of.
SQL Server allows multiple databases per instance. Each database consists of data and log components. The log is usually placed on a separate device (file) so that it can be dumped (archived).
The logs to provide the combined functionality of the Oracle redo logs and rollback segments. However, there are no 'groups' of transaction logs; if one fills, the committed transactions must be dumped so the log can be truncated. Users performing DDL will get errors until the transaction log dump is complete. Consequently, a transaction log needs to be sized large enough for the largest transaction ever run in the database, taking into account the frequency of log dumps.
SQL Server provides read consistency by locking data - it does not retrieve 'before image' data. Consequently, readers may block writers. SQL Server will escalate locks to a less granular level (page, extent, table) as the number of locks increase so blocking is much more likely than in Oracle. It is important that applications promote concurrency by making transactions as short as possible and retrieving complete result sets following select queries.
Space management is much simpler with SQL Server. Extents are fixed size and there are no chained or migrated rows.
Hope this helps.
>snip<
>what are the major differences, especially between Oracle 7
>and SQLServer? From a DBA standpoint, what do I most
>need to be aware of?
>
>Thanks,
>Harry
Received on Wed May 20 1998 - 08:07:40 CDT
![]() |
![]() |