Speaker: Joachim Hammer, Principal PM Lead for Data Platform
Joachim starts with North American Eagle – Chasing history with Microsoft Azure on Channel9.
Lots of example where companies are generating data and this data have to be analyzed. This is where the cloud comes in. How to transform data into information so you can take intelligent action?
Today’s modern apps produce and consume data at a staggering rate, are increasingly complex and experience unpredictable explosive growth.
The Data Platform in the Azure cloud offers lots of tools. Where do the customers start? When you think of data management, it’s a loop: you start with capturing and managing the data. Then you want to transform and analyze this. Finally you want to visualize it and make decisions based on it.
Azure SQL Database is a fully managed database-as-a-service built on SQL with near zero administration. Built for SaaS and Enterprise apps, it can scale up and scale out, it has high availability and geo-replication, the data is secure and private. It is fully compatible with SQL Server 2014.
Azure SQL DW Service is the companion to database, a relational data warehouse-as-a-service fully managed by Microsoft. It’s fairly new, in extended preview. One of the features is that you can pause this, all you pay is for storage costs.
Behind the scenes, you can monitor your database for spikes and you can submit a scale request to up your plan to handle more load.
Scaling considerations: you have a database with spikes. On-premise you have to provision your database with enough hardware to handle this load. Even if the spike happens only once a year, you have to buy a bigger box. In the cloud you can easily scale up and down depending on your current load.
If you have lots of databases and these have different spikes, for this Microsoft is previewing the notion of an elastic pool. Instead of just managing the load on per database basis, take the databases in a pool so the size of the pool is enough to handle the load. All you need is a pool that can handle the spikes. Pool allows you to group together databases with different spike behavior and provision according to your load. In a pool you have the ability to borrow DTUs for databases that have different spikes. Use elastic jobs to run a script across all the databases in the pool. A pool will save you money in some cases. If all databases spike at the same time then pool won’t help you.
Data is kept safe and secure with Control Access for Application and Database, Proactive Monitoring and Data Encryption. Auditing and Threat detection is done with machine learning and suspicious behavior is flagged for the owner to look at. New feature: Always Encrypted, SQL doesn’t have the key. Role level security.
Data masking happens inside SQL and no changes have to be done to the application. There is a difference between static and dynamic data masking. The mask is applied on the fly before the result is returned, original data is not changed, e.g. credit card number, phone number or email address. Custom mask can be configured. Static allows changing the values but this is not offered yet. Recommendation for fields that should be masked.
What if not only relational data. How to provide more flexible data. Unformatted or loosely formatted data, also called semi-structured information. Azure DocumentDB is a fully managed NoSQL document database service build for the cloud for this purpose. Can be queried using SQL, has also JSON integration.