A Technical Debtor

Azure is the kernel of the Microsoft cloud platform, allowing others to build killer apps

Azure is a scalable solution for deploying services to the cloud, hosting by a hypervisor optimized for cloud computing. Security is key and is implemented at many levels.

Azure manages the entire global infrastructure, while providing a layer of abstraction. Automated service management to upgrade services without degrading performance or taking service down. OS and service are managed separately. The core of Azure is a fabric controller that treats hardware as a fabric servicing all services – service endpoint is specified and the fabric controller handles the transition.

Developers need to specify service models, including roles and groups, endpoints, interfaces and configuration. This is essentially the architecture of the service.This architecture is stored as an XML file. We need to specify both the code and the model.

The expectation is 24/7 availability. All Azure components are built to be HA, both highly redundant and fault tolerant. (Two drive failures won’t take a box down).

Developer in Visual Studio, then deploy to cloud. Azure leverages current tools to ensure skills transfer. With standard web apps, we use the local web server for debugging. With cloud apps, we use the “cloud on the desktop” in much the same way to simulate and debug locally.

Evolution of Development (roughly one per decade)

1. Monolithic
2. Client-server
3. Web
4. SOA
5. Services

As we move to the cloud, we need to be architecting applications for horizontal scale-out, rather than scale-up.

Services Requirements

1. Interoperability, business processes spanning different environments
2. Identity and security
3. Data management and compliance
4. Services management

.NET Services

1. Service Bus: helps data traverse firewalls
2. Access Control: single, federated identity platform; users control their own identity; open and interoperable (“Geneva”)
3. Workflow Services

SQL Services: built on top of SQL Server. Geospatial data will eventually be included.

1. Database
2. Data Sync
3. Reporting
4. Data Mining
5. ETL
6. Reference Data

Over time, Microsoft sees the cloud computing technologies moving back on-premises.