Highlights Day 2 – Microsoft Build 2016

Some of you might recognize this guy, Jeffrey Snover. We had a blast talking about Windows Server 2016 and application models as well as PowerShell DSC and what we are doing with that at Addlevel. More fun stuff to come, keep your eyes and ears open! We also handed over our “PowerTie” as a thank you for the video we got during the Windows Server 2016 Summit presentation.

build1

Jeffery and Richard

This guy, Scott Hanselman, had a great 20 min session about open source. Why should you get onboard with this? Well, it´s all about collaboration and learn from others. We at Addlevel are on the right track using GitHub as a group and the forking/merging etc. on everything code. Be aware though, don´t use it as OneDrive or Dropbox, it´s not just saving files online – remember, collaboration with friends!

build2

We watched another incredible session presented by Hanselman and Scott Hunter about ASP .NET Core 1.0. The biggest difference between Core and the “normal” asp.net version is that the core version is fully modular and can run on windows, Linux and Mac O (if you didn’t already know this, both of these are fully open sourced). Really cool stuff in this session where the final demonstration was a load balanced docker image of asp.net core running on amazon, azure and docker cloud.

build3

Jonas met up with Igor from the AzureStack team, and got a lot of insider information about what´s going on in the container space and how the Azure Container Service will run on the stack, also Microsoft´s take on buying the stack directly from bigger hardware vendors such as Dell or HP and automatically scaling out your private cloud. You don’t want the “Vendor -> Microsoft -> Consultant” blame game here, what´s important is how you design and operate/orchestrate your workloads on the stack!

Lori Lamkim – Director of Program Management, VS Team Services och Brian Keller – Principal Group Program Manger, VS Team Services held an inspiring session on how Microsoft adopted DevOps internaly.

Key takeaways on how to be agile, fast and deliver quality code can be summarized by the following points:

  • Deliver small code changes often instead of doing it in big sprints
  • Keep the teams small and crossfunctional (about 8-12 individuals)
  • Keep the team together in rooms to foster team spirit
  • Stories and Tasks should be handled by the Teams, not by the stakeholders
  • Kanban is a great tool to visualize the work in progress
  • Every team should take care of their own bugs.
  • Do not use stabalizing periods at the end of a sprint. Fix the code directly
  • Use Feature Flags to enable features spanning multiple sprints when they are ready for production.

In the afternoon we were inspired by the session on Azure Service Fabric for Developers. Service Fabric is a distributed systems platform for building and managing applications at scale, where you deploy Microservices into Service Fabric clusters on prem or in the cloud. Amazingly enough on both Linux and Amazon Web Services, as well as Azure and Windows

With Service Fabric, developers and admins don’t need to solve infrastructure problems (or as stated in the session – ”you just need to code!”) and can instead focus on developing workloads. Service Fabric manages provisioning, deploying, upgrading/patching as well as the monitoring and deletion of deployed applications.

One among many advantages with service fabric is the way of managing state in its state architecture. In traditional models, stateful data is kept in external databases, with a stateless compute layer. With Service Fabric, you can build stateful compute layer, making it easy to build highly scalable, efficient and resilient applications.

Depending of your needs you can scale different parts of the application. Microsoft uses the term ”Hyperscale” when describing microservices, which undeniably sounds like an apt name.

 

Posted in Blog.