PHP mysql为统计数据选择引擎类型

Question

I am trying to develop a interface for a networking system. Basically, a software puts all traffic datada through UDP and an other software collects, and inserts data into mysql (clients doesnt have mysql server, only server which has user interface access has mysql server). there are about 50 clients has really high traffic and insert operation on mysql server is really fast. but i am wondering about performance issues couse i am planing to print stats in realtime with push server or something, also, these traffic stats must be saved for a long time, so what kind of database engine i should use? myisamm, innodb, archive? anybody can explain to me please? also i am open for suggestions too

Answer 1

I think the most important thing you need to do is set up an environment where you can do some testing and profile different approaches.

MyISAM can be wicked fast for the sort of thing you're talking about where you're only adding rows onto the end of the table and running selects against it - no deletes or updates.

You'll need to archive older rows at regular intervals and I believe the archive storage engine might be a good fit for this, but I have never had occasion to use it and do not speak from experience there. I doubt it would be usable as the main table for the app, though. You could consider some sort of partitioning eg where you store results in different tables by month.

The book High Performance MySQL should probably be your first port of call.

Sounds like an imposing project - good luck with it.

Answer 2

Here are the details of innodb and myisam

• myisam is better for high read volumes, innodb for high update volumes due to table vs row locking.

• innodb is journaled, and can recover from crashes where myisam can't, much like NTFS vs FAT file systems.

• myisam has full-text indexing, innodb doesn't.

• innodb has transaction support, commits and rollbacks, myisam lacks these.

so if you don't really care about updating and recovering tables, and don't need transaction support you can use MyISAM. To solve problem with large amount of data you can consider using MySQL partitions and then periodically archive older entries in backup tables with archive engine.
Also you can consider delayed inserts if you are really concerned about reading speed.

UPDATE :
For MyISAM tables, if there are no free blocks in the middle of the data file, concurrent SELECT and INSERT statements are supported. Under these circumstances, you very seldom need to use INSERT DELAYED with MyISAM. insert-delayed

Answer 3

您应该构建每天更新的统计表，您可以查询所需的统计信息。