During the last decade, the decoupling of interfaces from implementation, scalable hosting models, service orientation, subscription-based computing, and increased social collaboration became the goals of distributed systems. Now, Internet-hosted distributed applications with connectivity to internal applications—often referred as Software plus Services (S+S)—are gaining popularity. Organizations are leveraging datacenters hosted by third parties to alleviate concerns about hardware, software, reliability, and scalability. These are just some of the new architecture trends that help you build interoperable applications that scale, reduce capital expenditure, and improve reliability. Cloud computing offers many of these benefits.

A cloud computing platform enables applications to be hosted in an Internet-accessible virtual environment that supplies the necessary hardware, software, network, and storage capacities and provides for security and reliability, removing much of the burden of purchasing and maintaining hardware and software in-house. In the cloud, you can develop, deploy, and manage applications as you have in the past and integrate these services to your on-premise applications. You pay only for the time, resources, and capacity you use while scaling up to accommodate the changing business needs.

In this article, I will examine the typical cloud platform architecture and some common architectural patterns, along with their implementation on the Windows Azure offering from Microsoft.

Figure 1 illustrates the typical cloud computing platform architecture.

In this model, each layer abstracts the layer below it, exposing interfaces that layers above build upon. There is no hard dependency between layers and each layer provides composable or Plug and Play architecture with services from other layers. Each layer provides horizontal scalability as needed.

As you can see, a cloud platform is composed of a number of subsystems. Let's look at each one next.

A hosting platform The hosting platform provides the physical, virtual, and software assets. These assets include physical machines, operating systems, network systems, storage systems, power management, and virtualization software. Bare metal and other operational resources are abstracted as virtual resources to the layers above.

Cloud infrastructure services The most important function of this layer is to abstract the hosting platform as a set of virtual resources and manage those resources based on scalability and availability needs. Fundamentally, this layer provides three kinds of abstract resources: compute, storage, and network, and exposes a set of APIs to access and manage these resource abstractions. Thus you gain access to the underlying physical resources without knowing the details of the underlying hardware and software and can control these systems efficiently through configuration. Services offered by this subsystem are often known as Infrastructure as a Service (IaaS).

Cloud platform services Developing and managing software for cloud computing is complex. It becomes really complex when you integrate on-premise software with hosted services. Platform services provide a set of capabilities exposed as services to help with such integration. For example, in the Azure Services Platform, Microsoft .NET Service Bus helps with discovery and access while the Microsoft .NET Access Control Service helps role- and rule-based claims transformation and mapping. Availability of platform services may differentiate one cloud provider from another. Services provided by this layer are referred to as Platform as a Service (PaaS).

Cloud applications This layer houses applications that are built for cloud computing. These applications expose Web interfaces and Web Services for end users, enabling multitenant hosting models. Some functions include connecting disparate systems and leveraging cloud storage infrastructure to store documents. These services fall under the umbrella of Software as a Service (SaaS).

Security services Security services ensure token provisioning, identity federation, and claims transformation. These services are built on the open standards, WS-Security, WS-Trust, WS-Federation, SAML protocols, and OpenID, for greater interoperability.

Management services Management interfaces cut across all the layers described above. The hosting platform leverages management interfaces and agents for automated scalability and availability administration. Even though the cloud is hosted and managed in a datacenter, customers may need functions that allow them to easily control their application and post deployment configurations, get analytics about service usage, and connect their enterprise management systems.

Tools Tools help you build, test, and deploy applications into the cloud. These tools may be extensions to existing tools (Visual Studio Tools for Windows Azure, for instance) or hosted tools from a specific cloud provider.

Users and providers of cloud computing There are three categories of users in cloud computing: cloud platform providers, cloud consumers, and end users. Cloud platform providers provide the hosting platform and cloud infrastructure services. Cloud consumers utilize the cloud platform and develop applications and services to be consumed by end users. Cloud consumers configure applications for scalability, availability, and security needs. End users leverage the services offered by cloud consumers. These users could be human, organization, or machine and may be hosted anywhere.

In this context, Azure provides a cloud platform while cloud consumers (or ISVs building integrated cloud solutions or enterprises) leverage this platform to build applications. For example the Live Mesh data synchronization platform leverages the Azure Services Platform and Windows Azure to develop and host S+S services for end users.