@2lon+16aKMMeV - Please tell me what "cloud" is, I'd love to hear what you think it is I don't understand. Keep in mind that not only do I have nearly three decades of data center Operations experience from bottom to the top, deployed or managed more devices/machines than you've had meals, but also that I've heard nearly 1000 different definitions of what "Cloud" means ever since it became a buzzword 15 years ago and no two people have the same exact definition. The customer doesn't care if it's hosted at ATOS or QTS or Involta or AWS (not happening soon) or Azure - to the end-user perspective this is "in the Cloud". Period. Full Stop.
@3opt+16aKMMeV - You could make a counter-argument that the true "hack" is all of the evolutionary changes made to applications to make them more functional for Millennial Photo Sharing and Meme Generation and that the current model used is pretty much the way things have been done since Client/Server architecture (you could argue this back to Time-Sharing on mainframes if you want) became a thing. It's all the same thing in the end - compute and storage resources are not local to the end-user - aside from occasional rare latency issues it's irrelevant that the application is in a room on-site or 500 miles away from the customer.
With that said, I agree for the most part that it would be nice to be able to incorporate some of the evolved technologies into the Sunrise ecology - that's what I'm hoping Microsoft will be able to assist with. There are hurdles to something as monolithic, modular, and individually customized as Sunrise being converted into a mobile-phone app delivery service, both technical and regulatory.
The techniques you mention aren't unique to Azure. Microservices are part of the programming paradigm, and would be up to Development to see if shifting more in that direction would be more trouble than its worth. Continuous deployment is also more of a function of the software Development teams and a strategy rather than a "technology" - but I assure you that most customers would not be happy with constant change. We have enough trouble trying to keep them somewhere near Major Versions let alone CUs and hotfixes - although everything is light-years better on the Hosted side, because maintenance downtimes and application of updates is mandated (subject to scheduling agreements). Consider also that most customers have a fair degree of customization and a continuous deployment model assumes a homogeneous end-user base - the risk of any of the updates causing issues with a percentage of the customer base would be frankly pretty scary. Finally, "containers" - the one area I see this being useful would be on the client-delivery side, as Citrix/XenApp is pretty much ready to go with full Docker integration as well as Kubernetes. Azure supports both, and specifically has the "Azure Kubernetes Service". Simplifying and containerizing just this major component would greatly benefit the overall management of the systems involved.
There are other modern "cloud" technologies at Azure that are much more immediate and agreeable from the business end. The ability to spin-up/spin-down metered environments on a schedule, running low-priority environments on lower-tier resources, dynamic allocation of resources, dashboard access to all of the individual customers, the hands-off Disaster Recovery design, and a pile of other things all contribute to a much more stable, resilient, and (most importantly) cost-effective way of doing things. It's all about lowering the TCO, people.
The Azure stuff has been pretty much outside of my scope since they started tinkering with it about four or five years ago, but I've tried to keep pace with what this could do for us.