Blog Post

Microsoft Azure Walks a Thin Blue Line

With Azure, Microsoft (s msft) is trying to strike a balance between giving customers the ease of a platform as a service and the customization that power users need to build tailored applications — both in-house and in the public Azure cloud. In the wake of the Redmond giant’s developer conference, where it detailed more of its plans, it became clear that Azure is striving to be a general purpose cloud offering for enterprises that doesn’t make developers sweat the small stuff or compromise on bigger things.

If we compare it to infrastructure-as-a-service providers such as Amazon’s (s amzn) Ec2 or Rackspace’s CloudServers (s rax) products, Azure attempts to handle more of the actual management and provisioning of virtual machines for a user. The biggest issue the target customer faces isn’t the hardware cost but the expense of managing an application on the hardware, Amitabh Srivastava, senior vice president at Microsoft with responsibility for Windows Azure, told me. So the goal was to allow Azure to run so developers don’t think about the underlying hardware as they might on a pure IaaS product.

This is where the idea of Azure as Microsoft’s OS for the cloud comes in. Azure is a platform-as-a-services play that seeks to leverage what Microsoft has learned through its OS dominance. First, it’s playing nice. Microsoft ensures that developers can use a wide variety of  programming languages to build on Azure such as PHP, Eclipse and Java, which is pretty unique among platforms. Earlier this year, I spoke with Microsoft about its plans for Azure and came away with the clear sense that the company’s programs and .Net would really shine on Azure, even though other programming languages would also work. Now I get the sense that Microsoft is working hard to emphasize how suitable Azure is for programs built using a variety of languages, even those that have no ties to Redmond.

Second, Microsoft is hoping Azure can play on many machines. If we view the data center as akin to a machine, Microsoft wants folks to be able to create applications that can run on a Microsoft Azure cloud or internally on an in-house cloud. Newly launched AppFabric is the solution for that. AppFabric is software that folks can run in their own data center to create an internal cloud that can talk to and share information with the public Azure cloud if the client wants. Rackspace and IBM (s ibm) are both attempting to build these types of bridges between internal and their external clouds.

Finally, Microsoft is trying to keep its own offerings separate. Instead of bundling Microsoft products with Azure, Srivastava outlined a software-as-a-services strategy that will offer customers Microsoft Exchange on Azure or SharePoint on Azure as a service. This may explain why some offerings such as SharePoint Services and Dynamics CRM Services, announced last year, are now divorced and missing from the Azure announcement this year.  However, competing software can be offered on Azure as well.

Larry Augustin, CEO of Sugar CRM, has chosen to offer customers the ability to host Sugar CRM on Azure, and says the platform was easy to build on, even though SugarCRM is built using PHP. He also mentioned that Microsoft’s database offerings were more complete than the newly launched Amazon Relational Database Service, and offered me a possible model for future clouds. In his vision, customers will be able to choose which clouds their software as a service are hosted on. This is the polar opposite of the vision espoused by SugarCRM’s biggest rival, (s crm), which hosts its software on its own cloud. But the power of the modular and open approach is hard to deny, and Microsoft is smart to embrace that.

17 Responses to “Microsoft Azure Walks a Thin Blue Line”

  1. Right now i’m fuming. I’m in Namibia (Africa for those who might not know). I signed up for Azure being a technocrat who likes all things new and powerful. And true to their word, Microsoft sent me a link to one of their development tools to connect to and use their hosted SQL servers.

    And lo and behold, i cannot download it because i’m not in the United States. Thats really a big blow, i’m not sure i’m gonna come back, will probably stick to developing my web apps the tried and tested old fashioned way, and upload them to a variety of smart hosting companies that allow me do what ever i want (data wise that is).

  2. Mishan Kontroll

    “Microsoft ensures that developers can use a wide variety of programming languages to build on Azure such as PHP, Eclipse and Java, which is pretty unique among platforms.”

    Eclipse isn’t a programming language. It’s a development environment supporting several languages. And is this all that unique? Doesn’t EC2 let you use a wide variety of tools? Don’t run-of-the-mill hosting providers?

  3. Azure is to expensive for entry level projects. MS should of looked at what google did with GAE pricing.

    They should have an entry level pricing which is no more than a virtual dedicated server to run.

    The price of Azure is too much for developers boot strapping in my opinion. Yes on BIzspark I get some help but after that runs out i need to find the money to run it when I could run it for cheaper elsewhere.

    I love the concept of it but it’s over priced.

    • I have to agree! As a mobile developer MS needs to attract people to this technology. It almost makes me fondly remember that at one point not too long ago, they could have been the savior of mobile. Microsoft never quite delivered on Windows Mobile and Symbian et al. never allowed for the types of products that people such as myself wanted to build.

      If Microsoft wants some mojo, start at the tools and build up. I was just looking for online wireframing tools and project management technology for mobile efforts. Being agnostic with interesting build export options could be a very good point to stem the tide. Similar to Flash, they could allow for builds from wireframes and even produce iPhone builds of certain types. Interested? We built a customized build process that allows us to push a button and spit out an iPhone app. Apple will never allow this!

      It could also be a base for publishing scenarios. One of the reasons we built this was to allow a magazine or a catalog publisher to customize builds for apps that contained media such as photos and video and music. Apple again will not allow stuff like this. But being able to wireframe a magazine, tie it to the databases controlling the content, dynamically bridging web views and then exporting a build across Android, iPhone and Windows mobile would be quite cool. This is literally possible for certain classes of applications.

      Such products would also be interesting for onscreen prototypes and such. Imagine building a wireframe and exporting it cross-platform to a bunch of mobile devices? Microsoft could essentially become the glue, and get hold above Palm and Nokia. This will not be the case without workflow tools or very good technologies. Additionally, this could be tied to WebKit views, and a packaging system could be devised to hold the views and media. That would require that MS make Mobile Explorer WebKit compliant, which they definitely should do!

      • This conversation really has my interest for the moment, because this is completely possible and a pretty good idea. Think HyperCrad for mobile apps. That is how somebody described what we built It doesn’t build everything, but it builds a certain class of objects including wire-framed walkthroughs. Microsoft has all of the database tools to allow for the creation of mobile products and services.

        Take the project I am looking at now, since this one nearly bankrupted me! Grrr. Mobile education requires that I be able to build certain types of products repetitively. Boioks, readers, quizzes, professional development video training tools. If I had a templated system that allows me to do this, as a teacher I could literally plug in my content and build for three mobile platforms. As a UI/UX professional, I could create a wireframe, and build a walk through application across multiple platforms. That is exactly what we use this for now. Walkthroughs…

        The process requires databases on all sides for views, data and content, apps on all the platforms, and a tool that understands how to bridge the three spaces. Such tools would equivalently be useful for Nokia on Maemo. I would anticipate this to be much more difficult through.

      • If you automate things the cost come down, but like in all complex system there is a site effect. The cost of entry comes down.
        No incumbent in any market likes that. Better keep playing with players you know then to let unknown new players enter the field.

        As more people are working on a product as more linear the solution becomes, and as more complicated.

        Which clearly favors the existing large players, because of the resources they can throw at a problem. Even to “copy” it.
        Except when a small group of people comes up with a solution which is outside the linear path. Hence the success and failure of start ups.
        If you are outside the linear path don’t expect the big players to jump on an idea, but on the other hand you might be able to knock them over.
        My “guess” is 80% of all the code written for apps in the application store could be generated.

      • If the costs don’t come down it benefits Google who is making sure the costs come down. Do nothing. Die.

        We built the iPhone side of the tech, but Apple will not allow certain types of apps with embedded content — music. There is an incumbent that could be knocked over, and will be. The dynamics favor external, dynamic intelligence rather than an application that organizes liner file systems. But, we can ill afford to develop the Android side of the equation, and not all of the view controls are available there yet.

        My experience is that either I am crazy and this is silly or that people are not quite there app wise. I waiver between the two! It would be good to generalize app development across platforms. This is where Adobe has an opportunity. An online system that controls the process from build through development would be incredibly cool. It would be like Coda with wireframes!

    • When you say entry level, what you really mean is for hobby projects. With BizSpark you can get free tools for 3 years and Azure free for 8 months. After that time you should be either making enough to pay $100 month for Azure or shut up shop. Azure is not targeted at people who just want to play around indefinitely and make no money.