The SQL Server Evolution: Deep Dive

Download this episode

Download Video

Download captions

Download Captions

Description

Deep-dive into the evolution of SQL Server, from Mission-Critical investments in Transaction Processing and Data Warehouse to Enterprise Business Intelligence and Hybrid Cloud Scenarios.

Day:

2

Level:

200

Session Type:

Overview

Code:

BRK2558

Room:

S104

Embed

Format

Available formats for this video:

Actual format may change based on video formats available and browser capability.

    The Discussion

    • User profile image
      Patoduck

      Hi,

      Nice to hear about more features specially R integrated in SQL 2016 -hold on to your seat because here come a BIG but- .... BUT you have missed an opportunity to really shine. No support to GPU based database and GPU memory use. 16 sockets really makes no sense, you should grab all the power you get for example out of a USD 1000 graphic card like NVIDIAS Titan X  with 12 GB of Ram, and move data IN an OUT no matter if you put support to 16 sockets or 32, 64, 128 A single GPU will kill them all. 

      Really incredible you have skipped this feature, if you are not familiar with this go to www.map-d.com or watch this http://on-demand.gputechconf.com/gtc/2015/video/S5544.html

      Best regards,

      Patricio

    • User profile image
      jla

      Proud to be partner, amazing new features for starting the Revolution of Big Data , connected with IoT and Machine Learning. The must have data dlatform we were waiting for. I'll keep a look on the public release.

      JLA

    • User profile image
      jonreade

      "No support to GPU based database and GPU memory use".

      Well that's because in reality, how many infrastructure managers a) want or b) really need GPU based processing in their data centres? Fine if you're forecasting weather. Or doing some other type of fluid dynamic simulations.

      Most people who use SQL Server ARE NOT doing this sort of work with it. Ditto any of the other big relational platforms. Yes, GPUs are incredible. Personally, I love them. But most organisations do not need them and cannot justify putting them in a server to do something in a microsecond that might take a second on a vanilla CPU. And the bottom line is if you don't need it, and it gives you no discernible advantage, don't buy it.

      Doing R inside a relational platform without ETL (which can take hours) is FAR more advantageous than shaving a few hundred milliseconds off an analytical query that someone might not even look at until the next day. ETL is the operational bain of any data warehouse team's life, not the analytical speed. So, no realistic opportunity lost for the majority of Microsoft's customers - which are commercial organisations that need to run a solid, reliable, fast OLTP platform and occasionally do some useful analytics out of the back end. Ideology is nice, but pragmatism, security and reliability is what sells to business.

      Jon Reade.

    • User profile image
      bvukov

      @Patoduck:They just want to sell a lot of 7K$ enterprise licenses ;)

    Comments closed

    Comments have been closed since this content was published more than 30 days ago, but if you'd like to send us feedback you canĀ Contact Us.