30th September - 2nd October 2010

University of York

Premium sponsor
Gold sponsors
Silver sponsor
Bronze sponsors

Fast Track Foundations: Get Serious about Sequential I/O

Fast Track is a new reference data warehousing architecture provided by Microsoft. More than this it represents a new way of thinking about data warehousing. A Fast Track system is measured by its raw compute power - not by a DBAs ability to tune an index. Fast Track is an appliance-like solution that delivers phenomenal performance from a pre-defined, balanced configuration of CPU, memory and storage using nothing but commodity hardware.

Of particular interest in a Fast Track system is the way in which the storage and SQL Server are configured. To achieve the fantastic throughput without using SSDs requires some careful configuration.  This configuration is designed to make use of Sequential I/O to dramatically improve disk I/O performance.

Interested? If you have a large data warehouse that's seen better days or perhaps you are about to embark on a new warehousing project then you should be!  Fast Track is a great solution with a fantastic value proposition.

In this one hour session we'll aim to get under the skin of Fast Track and get some answers as to how it delivers such great throughput on commodity hardware. In the process we'll aim to answer the following questions:  
  • When might I need Fast Track? 
  • What is Sequential I/O?
  • How does Sequential I/O improve performance?  
  • What do I need to do to get Sequential I/O?  
  • How can I monitor for Sequential I/O ?
  • What may I need to change in my ETL to get the benefit of sequential I/O?

Allan Mitchell will be showing how to build SSIS packages to take advantage of sequential IO performance.

Still reading?  I'll save you a front row seat....

Speaker focus

James Rowland-Jones

JRJ.jpg
James is an Advisory Consultant at EMC Consulting. His focus has been the development and technical direction of major, mission-critical, pan-European projects in the enterprise. His experience includes data and system architecture, information analysis, logical and physical database modelling and design, highly available implementations, ETL process architecture and client tools implementations including a number of major bespoke applications. He has worked in IT since 1998 and with SQL Server since 6.5.

James Rowland-Jones's blog http://blogs.conchango.com/jamesrowlandjones



Downloads

No items to download

Other sessions from this conference

Denormalisation –having your cake and eating it.
Design patterns for SSIS Performance
Designing and Tuning high speed data loading
Event Driven Cube Performance and Security Monitoring
Getting Started with SQL Azure
High Availability in the Real World!
High performance functions
Improving Database Performance by Removing the Database
Inside the SQL Server Query Optimizer
Introduction to Performance Tuning Analysis Services 2008
Lessons Learned from 128 Core OLTP Testing
Make your DBA life better using SQL Server technologies - notes from field & best practices
Monitoring and Tuning Parallel Query Execution
Non-Functional Dependencies - A talktorial on getting performance and scalability right through proper Database Design and understanding Set and Relational theory
Optimizing tempdb and temporary object use
Performance Management for SQL Server with SCOM 2007
Performance Monitoring & Tuning - Documented Tips & Tricks
Performance Troubleshooting made easy with PAL
Photograph: SQL Server Crash Dump Analysis
Scaleout scenarios with Service Broker
SQL Server 2008 Reporting Services Scalability and Performance
Supercharge MDX performance using MDX Studio
Transaction Log Performance Tuning, Stop shouting Timber!