NORTHEAST SQL SERVER USER GROUP
Network, learn, ask a question, meet other folk, get fed - these are all things that happen at user group events. These events are a really great opportunity to socialise in an informal learning experience - if you want your own exposure then come and do a 1 - 5 minute nugget in front of your peers. Remember to tell your friends and the people you work with; make sure you register as soon as you can.
5.30pm – 6:00pm Registration and networking
6:00 – 6:15pm Round Table discussion and Nuggets - ALL
6:15pm - 7:15pm - An Introduction to Column Store Indexes and Batch Mode with Chris Adkin
7:15 - 7:40pm - Break
7:40pm – 8:40pm - Super Scaling SQL Server 2014 Parallel Insert with Chris Adkin
8:40pm - 9:00pm - Feedback and Close
Finally: Drinks on the Quayside :)
This session is a primer on column store indexes and SQL Server batch mode, it delves into the background behind the performance gains achieved by batch mode and includes demos on things that can be done to further enhance batch mode performance such as pre-sorting the data prior column store index creation, delta stores in action, segment elimination and the difference that batch mode makes. Back ground information on the 'Anatomy' of a column store index, how row mode works and why it is inefficient and a brief overview of CPU architectures is also provided. The full deck behind for this presentation ( which may be subject to change ) can be found here: http://www.slideshare.net/chris1adkin/an-introduction-to-column-store-indexes-and-batch-mode.
About Chris:Chris is a freelance SQL DBA / developer who has been using SQL Server since 2000, his passion is for scaling SQL Server and understanding how the engine behaves at depth when pushed to its limits.
This session will take a look at how parallel select into can be scaled to the nth degree in SQL Server, such that all available hardware resources are utilised as fully as possible. Windows Performance Analysis toolkit will be used to ascertain where the pressure points are in order to target efforts in tuning this new feature to the full: including such things as hash partitioning in order to enhance parallel scan performance, reducing pressure on the PFS bit maps, reducing pressure on how the PFS is allocated in memory and the final performnance frontier, reducing the presure on the log cache access spin lock, the thing which ultimately ham strings all workloads that need to be ACID compliant.