- 最后登录
- 2023-8-16
- 在线时间
- 1686 小时
- 威望
- 2135
- 金钱
- 50532
- 注册时间
- 2011-10-12
- 阅读权限
- 200
- 帖子
- 5207
- 精华
- 39
- 积分
- 2135
- UID
- 2
|
1#
发表于 2012-1-23 22:27:21
|
查看: 8834 |
回复: 19
Oracle tuning used to be largely focused on tuning file I/O. Now tuning is focused on wait events, many of which are file I/O related. Lately the evolution of server hardware is again changing how we look at tuning our Oracle databases. Some of these changes are:
1.
Spinning hard disks are obsolete, but refuse to die.
Hard disks should have gone away by now, but just when you think they are done for, they emerge faster with more space and lower cost.
2.
Solid State Disks have arrived.
SSDs are now cost effective to implement even on low to midrange servers. Although slower than memory, they are an order of magnitude faster than spinning disk.
3.
Memory costs are quickly dropping.
Having 4G of RAM was impressive only a few years ago. Now midrange servers routinely have 24-32G RAM. Today’s standard is a 64 bit OS. All this memory reduces file I/O considerably, but not entirely. The database still needs to get those changes on to a non-volatile memory, i.e. disk, SSD, etc.
4.
CPUs are becoming multi-core and more powerful.
With all that memory, you would think I/O would go away. However, modern CPUs are pushing the bottleneck back to the I/O subsystem.
What this means to the DBA is that file I/O is still important and still a key part of getting your Oracle database to optimal performance.
11g_disk_io_internals_ioug_2010_v1_ce1.doc
(255 KB, 下载次数: 58, 售价: 5 金钱)
|
|
|手机版|Archiver|Oracle数据库数据恢复、性能优化
IP卡
狗仔卡
分享
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
显身卡
发表于 2012-2-11 14:05:15
