Mastering the Cloud.. with Windows Azure Curricula Resources

windowsAzureLogo

A number of Universities have now added Cloud technologies into the curricula and at Microsoft we are assisting with the development of specific cloud computing curricula to support the teaching, learning and research of cloud technologies. As predicted by Analysts such as Gartner, IDC and Forrester cloud computing will take a substantial piece of the market from traditional deployment models.

Today there are four major cloud computing platforms available Amazon EC2, Windows Azure, Google App Engine and Force.com. There are a number of difference between each of these platforms Amazon and Microsoft have succeeded in building general-purpose cloud environment so meet the needs of SaaS, PaaS and IaaS and Google and SalesForce have specific SaaS offerings based on internal technology and use programming languages like Python and APEX.

Microsoft has developed curricula and certification to allow you as educators to make aware, train, educate and certify students on the use of cloud technologies. The adoption of the cloud within industry will grow demand for application developers and cloud administrators. Graduates will need these specific skills and experience  for roles such as cloud software engineers. 

Therefore the opportunity for student graduates with experience and understanding of cloud technologies is a huge.

What are these skills?  

In terms of skill sets both Amazon EC2 and Windows Azure can host applications written in different programming languages. However skills in Java and .NET are the most favourable as both are primary development platforms for Amazon EC2 and Windows Azure.

What is the Required Knowledge?

Mandatory Skills

A programmer looking to dive into Windows Azure application development should have a working knowledge of Microsoft .NET technologies. Particular skills are:

    • .NET Framework 4.0

    • ADO.NET Data Services

    • LINQ

    • Windows Communication Foundation (WCF)

    • ASP.NET MVC 3.0

    • Multi-threading

Developers should be familiar with RDMS concepts and MS SQL 2008. Additionally, a solid understanding of HTTP protocol and REST concept is very desirable as it helps to assess the implications of network topology (load balancers, proxy servers, CDNs) on RESTful web services. Knowledge of Service Oriented Architecture (SOA) design principles is essential as cloud applications strongly rely on services. A great starting point is the Microsoft Cloud Curricula resources.

Cloud Concepts

The first logical step is to become familiar with cloud-related concepts and to adopt the principles of cloud application development. There is a lot of information on the Microsoft Faculty Connection site.

Cloud Development Skills

From a software developer's perspective, cloud can be treated as a way to get on-demand access to two types of scalable resources: compute (CPU) and storage which are available via services provided by the cloud platform. There are also great resources and presentations at Channel9 describing platform infrastructure and application lifecycle which will help you learn how Windows Azure hosting environment works in detail.

Azure Fundamentals 

Cloud application runs in a bare Windows 2008 operating system

Don't assume that Windows Azure hosting environment has any preinstalled software; it's a bare operating system. Any functionality, usually supported by preinstalled software, should be instead implemented within the application hosted on Windows Azure.

The application instance can be recycled by the platform at any point of time

Everything stored on a local disk drive memory will be deleted once the instance is recycled. To preserve the data and make it available to other instances, use Windows Azure Storage services.

Cloud application runs in a concurrent environment

Services provided by Windows Azure platform are designed to operate in concurrent environments with the use of "try and correct" pattern. The application should follow this pattern and properly handle cases in which access to a service is declined by repeating the operation later. Another aspect to keep in mind in that a web application under Windows Azure always runs behind load balancer.

Getting Started with Windows Azure Development

I would recommend looking at the resources and curricula available at the Microsoft Faculty Connection site additionally if your not familiar with Visual Studio 2010 we have a number of resources available.

Development Environment

The Windows Azure development environment requires Windows Vista SP2, Windows 7 or Windows 2008 operating system. The following software should be installed via DreamSpark.com

You can then use the Windows Azure Virtual Lab resources to start you Azure Cloud experience.

Web, Worker and VM Roles

There are three types of application roles supported by Windows Azure:

  • Web role

  • Worker role

  • Virtual Machine (VM) role

Further information on these roles can be found in Introducing the Windows Azure Platform (Whitepaper)

Storage Services

There are three types of storage supported by Windows Azure platform:

  • Blob storage

  • Table storage

  • Queue storage

The services are exposed via REST API and available outside of Windows Azure hosting environment as well, so one can create an application for a mobile device which interacts directly with the storage. Windows Azure SDK comes with a managed library providing access to storage services via an object model. Further information on these types can be found in Introducing the Windows Azure Platform (Whitepaper) and at Steve Planks Blog https://blogs.msdn.com/b/plankytronixx/

There are several important things to know when working with Azure storage:

  • Storage objects are addressed by URL, so certain restrictions are applied to the object name.

  • There are three types of blobs: single blob, block blob and page blob. Each of them has a minimum and a maximum size.

  • Table storage is not a relational database. There are no relationships, indexes and constraints. It's more like an Excel spreadsheet highly scalable in the number of rows.

  • A table always includes two properties (PartitionKey and RowKey) forming a primary key, the total length of the key can't exceed 1024 symbols. Only 256 symbols of the primary key can be used to address the record.

  • Table always includes Timestamp field used to resolve conflicts

  • Table has limits of 1M per entity (row) and 64K per property (field)

  • Sorting is not supported by Table storage, so it's always done on the client side.

  • Maximum number of records returned by a query against Table is limited to 1,000 entities per request. A continuation token should be used to retrieve subsequent data.

Azure Storage Performance

Azure storage performance https://azurescope.cloudapp.net/BenchmarkTestCases/ which gives a good idea of storage throughput under different scenarios.

Azure SQL

Azure SQL is a cloud version of regular MS SQL database. It looks like a complete replacement of regular MS SQL database with minor restrictions on T-SQL syntax. However, the fundamental restriction of Azure SQL is size limit of 50Gb per database. Another key advantage is that the Azure SQL databases is available outside of Windows Azure hosting environment. Additionally it is possible to setup access restrictions based on an IP address.

AppFabric

Windows Azure AppFabric is a set of middleware services designed to facilitate development of enterprise applications on top of Windows Azure. Learning AppFabric SDK is not required to develop Windows Azure applications, however one service highly demanded by developers to take note of: Cache service which provides fast access to in-memory data storage.

Deployment

You have arrived at the stage when your first Windows Azure application is implemented and tested in the development environment. To deploy the application to the Windows Azure environment, you need to sign up for the service. There is a free trial available for 90 days

Once the account is set, you can create a storage account and a hosted service. Each hosted service supports two environments: staging and production. Deploying the application is quite simple and requires uploading a package and a configuration file. The application in the staging environment is available under a private URL for testing purposes. Once the staging environment is tested, it can be switched to production in a single click.