Directions EMEA 2010: SQL Server Tips & Tricks - Indexes

A session presented by a fellow-MVP:Jörg Stryk. The guy is unbelievable into SQL Server and SQL Server performance tuning on a NAV database (like Hynek is as well), so he's quite a good catch to be presenting a session about this topic, off course. May be you know Jörg from his book: The NAV/SQL Performance Field Guide.

Jorg started out by pointing out that a Key is not necessarily the same as an index. Key's are terms from the classic environment, while an index is something else. And it's indexes that he would liked to explain...

He compared it with the "yellow pages". If you want to search something, you go to the index, which points you to the right page. He even tore out the index pages to make clear what an impact "removing an index" can have. Furthermore, he pointed out there was still a way to search in a "somewhat" decent way, because the pages were physically stored in an arranged way. Furthermore, removing that physical structure (removing the clustered index) would mean ... Heaps of data where you have to search in .. Which is exhausting. He even was physically exhausted after toring out pages from the book ... so we got the message pretty well :-).

After that, he went into the real theoretical part, explaining the structure on SQL Server, and how you can see when there is an index problem .. having lots of reads.

To be able to do something about it, you have to understand how NAV reacts on SQL Server. For starters, how does a key translate to an index? The primary key, for example, will be the primary key in you SQL Server, plus, it will be clustered (by default) as well. All secondary keys will be made as indexes as well, completed with the primary key...

I suggest you check his blog as well. He even did this session in Munich and blogged about it. Just read here. Now that I noticed, it's needless to bother you with another description of his session ... easy for me ;-).




Good job, Jorg! Very well explained. This is really importand and/but basic stuff, so in my opinion, everyone should be able to understand this - and handling problems related to indexes, reads, ... . Especially because from next version, only SQL Server will be supported... .

Comment List