Benchmarking a 1TB database

In this series, Paul will be exploring a range of different data configuration techniques and measuring their impacts against baseline measurements ... Source : Paul Randal - SQLSkills.com

Physical Database Storage Design

This article provides a guide for physical storage design and gives recommendations and trade-offs for physical hardware design and file architecture ... Source : Microsoft.com - Technet Article

Multiple files for multiple CPUs

A common practice for multi core boxes is to create 1 file per cpu core. Paul addresses the merits of this in this blog post ... Source : Paul S. Randal - sqlskills.com - blog post

How many files should a database have?

In part 1 of this posting, the SQLCat team answer this question from an OLAP workload perspective ... Source : SQLCat.com

How Many Data Files Should I Create for a User Database?

Auto-shrink - turn it OFF! 

In this blog post, Paul talks about the dangers of AutoShrink - among others, unexpected performance hits and index fragmentation ... Source : Paul Randal - SQLskills.com - Blog Post

Autogrow; The Performance Impact of a Large Number of Virtual Log Files

It's a well known best practice to avoid autogrowth operations by presizing database files. This post points out in graphic detail (as only Linchi can) the reasons for this recommendation ... Source : Linchi Shea - SQLBlog.com

Filegroups - Dividing Data for Performance 

In this blog post, James talks about the use of filegroups and discusses the way he typically sets up databases using 3 filegroups - Primary, Data and Indexes ... Source : James Luetkehoelter - SQLblog.com - Blog Post

Data and Logs on the Same Drive?

Best practice dictates the separation of data and logs on separate drives, but what if drive separation isn't possible? Brent walks us through some situations in which placing both of these resources on the same drive may be appropriate ... Source : BrentOzar.com - Blog Post

Drive Letter Limitations

Linchi blogs about Drive Letter requirements for SQL Server Instances and its potential effect on large consolidation projects in the future ... Source : Linchi Shea - SQLblog.com - Blog Post

Mixed Page Allocation Contention and Trace Flag 1118

Trace Flag 1118 is used to force uniform extent allocation in place of mixed extents, and is often used in avoiding allocation contention in TempDB. Despite SQL 2008's improvement in mixed extent allocation, this trace flag is still available as an option, as Bob points out in this post ... Source : Bob Dorr - CSS SQL Server Engineers - Blog Post

Data Compression: Strategy, Capacity Planning and Best Practices

While the new data compression feature in SQL Server 2008 delivers obvious benefits, there's a number of considerations before implementation, most notably the CPU impact. This article covers both the benefits and impacts of compression by different workload types ... Source : Sanjay Mishra - Technical Article - MSDN.com

Data Compression Tests

In this 4 part posting, Linchi highlights the effects (good and bad) of data compression. Part 1 covers compression times using various MAXDOP settings. Part 2 looks at compression rates for a number of real customer databases, Part 3 focuses on insert overhead, and Part 4 demonstrates the potential for massive read performance increases with compression enabled. As usual, Linchi includes scripts and graphs the results, making for an excellent read. I'm sure the figures will change with the RTM release, but interesting results none the less ... Source : Linchi Shea - SQLblog.com - Blog Post

Data Compression - Page Vs Row

Sunil from the Storage Engine Team blogs about the differences between page and row compression in SQL Server 2008 ... Source : Sunil Agarwal - SQL Server Storage Engine - Blog Post