Proof of Concepts: Making Hard Things Easy with Azure

Back when I managed my employer's enterprise architecture, prior to the days of virtualization and cloud computing, one of my most difficult projects was to build out a sandbox environment for a major proof of concept (POC). Locally hosted servers and all of the required licenses were expensive and scarce.

I could always expect a political wrestling match any time we needed to perform research in an environment that required either a few servers for a long time, or a lot of servers even for a short time. Heaven forbid that I should need a lot of servers for a long time to complete a POC. Hijinks would ensue. But if you're still struggling with scarcity when conducting research and performing POCs, I've got to tell you you’re doing it wrong.

Today, you have a couple options to make this process dramatically faster and easier. The first option is to use fully virtualized environments. This is a good option and has many benefits. But I’d like to focus on our second option, cloud computing with Microsoft Azure, for today’s article. (Of course, Amazon also offers a lot of goodness here, but I’m not an expert in their technologies. I’d love to hear your input though! Drop me an email).

The year 2016 has seen a torrent of new features, compute offerings, regional availability, and services with the Microsoft Ignite conference, earlier this month, highlighting a great many additional announcements from Redmond. Check out all the latest announcements from Ignite here

The first goodie I want to draw your attention to is the set of posters showing how the fictional company Contoso uses cloud technologies to support a large global organization, including security, identity management, networking, infrastructure, data, and applications. These posters are very informative and are written with the Enterprise Architect in mind. They’re immediately usable and useful. Read more about them here.

Azure Images are the Icing on the Cake

Learning resources are nice and all that, but I have a POC to get up and running by the end of the day! That’s where Azure Images come in. And, frankly, I love these things. Azure Images are pre-built and fully functional Azure server VMs available in the Azure Marketplace. Let’s say you’re running your SQL Server estate on Windows and SQL Server version 2008. You need to do regression testing for an upgrade to Windows and SQL Server version 2016. Normally, that would require a big, expensive, and time-consuming build out to accomplish. With an Azure account using the automated Portal features, you can pop up a multi-server estate in just minutes. (Check here for more details.)

Wanting to try out something you don’t even have in-house yet? Then Cortana Intelligence Gallery is for you! Let’s say you want to get involved in data science. Normally, a decision like that takes days or even weeks to process because we have to figure out the basics, like which tools to deploy. Instead, I can now deploy one or more fully configured Windows Data Science Virtual Machines in just a few minutes! There are also fully pre-configured Linux Data Science Virtual Machines and a variety of custom-designed solutions for specific scenarios, like anomaly detection, predictive analytics, churn analysis, machine learning, and much, much more.

Help Round Out My Knowledge

I know that similar things are possible with the other major cloud computing vendors. But I don’t know the details. If you’ve used some of the other popular vendors, fill me in on your experiences and their equivalent capabilities by commenting here.