Azure Data Lake Analytics is a pay per use big data analytics service where you can write and submit scripts in U-SQL. This is a platform-as-a-service offering which requires less management of infrastructure and you focus on building your applications.

Some Key capabilities that I favored and found useful

• U-SQL
  Simple and powerful while leveraging my existing skills of SQL, C# and Visual Studio IDE. Therefore, making me highly productive and deliver results.
• Cost-effective
  I pay for the processing time of my script jobs. Overall I spend hardly much since my job executions are only a few minutes on average and not very frequent.
• Role-based access security
  Able to grant various levels of permissions and assign roles to various types of users. This is leveraging Azure Active Directory. For example, enabling read-only users to report on data with tools such as Power BI.

For a more detailed overview and more of the key capabilities, read https://docs.microsoft.com/en-us/azure/data-lake-analytics/data-lake-analytics-overview

The following article nicely explains how to create Azure Data Lake analytics, but this blog will talk through how I set things up and weigh in on my own thoughts and considerations.

https://docs.microsoft.com/en-us/azure/data-lake-analytics/data-lake-analytics-get-started-portal

Go to Azure Portal and add a new Azure Data Lake Analytics service

Data Lake Store

I selected an existing data lake store within the same region/location.

Pricing

Pricing Details: https://azure.microsoft.com/en-us/pricing/details/data-lake-analytics/

Pay-as-you-go is $2 USD/hr per Analytics Unit. For development purposes, when working with data in roughly 100,000s of rows as a rough indicator, this pricing was more than well with me. I am not running jobs against the data frequently throughout the day either. I hardly noticed any significant costs. Monthly plans will provide at least 50% discount.

An analytical unit is about 2 CPU cores and 6 GB of RAM. For more details on an analytical unit read Understanding the ADL Analytics Unit

For my simple U-SQL jobs, I usually run jobs with 5 Aus that run about 1 min to give a ballpark sizing.

After deployment, here are some of the key settings and functions

I like to discuss just a few of these:

Add User Wizard
Allows for role-based access security of data analytics and files

• Select Azure AD user
• Role: Owner, Contributor, Reader, Data Lake Analytics Developer
• Catalog permissions for running and managing jobs
• Files and folder permissions

New Job
Running U-SQL scripts in the browser. Nice and convenient for simple scripts. The other option is using Visual Studio with Data Lake tools

When you submit the job, you can get really nifty execution run-time monitoring

Job Management

View the history and status of job submissions

Compute Hours

View the compete hours so that you can check against your costs. Here shows is very little indication of computing usage.

Data Explorer

View the folder and files of the data lake store and the databases and tables in Data Lake Analytics

For Microsoft developers who want explore developing big data solutions and don’t aspire to be an expert, I would recommend trying Azure Data Lake Analytics. It is easy to pick up with existing knowledge of SQL and C#, easy to setup in Azure and, most importantly, able to deliver an end to end analytics solution without too much grind. This is in contrast to the Hadoop platform. For those who aspire to be an expert and want to use all the bells and whistles, then Hadoop is the way to go.