SQLBits 2014

Is your data big enough for a PDW?

You have tuned your SQL server to it's limits, you have maxed the RAM on your server, and still your data warehouse is too slow. Is a Parallel Data Warehouse the next step?
When working with a data warehouse on
SQL Server, it is always best to start by looking at your code before buying
hardware. But you’ve looked at your code, and optimised to the hilt.

This wasn’t enough.

You bought higher and higher end hardware.

This wasn’t enough.

Now you’re not sure what the next step is. Is it to buy the absolute fastest
SQL Server money can buy? Is it to look to NoSQL, or caching solutions, or
scale-out messaging using Service Broker to distribute your load?

These may all be valid solutions.

But there is one more solution.
Automatic scale-out using the Parallel Data Warehouse. With Massively Parallel
Processing, balanced IO and the ability to add nodes at will, the PDW
represents a good solution for certain workloads. Is yours one of them?


In this session, Mark will discuss what the different techniques are that you can use to scale out, dive into the specific details of how MPP scales out, and from his project experience on the PDW and on other scale out projects, guide you through deciding if a PDW is the right solution for you.