Nicke Andersson

Windows Azure: How I deployed an existing Silverlight application in 30 minutes

Almost two weeks ago I posted about my first experiment with Amazon EC2, and how I was able to get an existing application (our Sproodle SaaS offering) up and running on EC2 in a couple of hours. That was the same day as Microsoft announced Windows Azure, and of course I wanted to try the same application to run on Azure as well. Two days later I finally got my Azure account, and to my delight I was able to get our application up and running on Azure in about 30 minutes. Which in my mind seriously rocked. In this post I will explain the steps I went through to deploy an existing application to Azure, and talk a bit about how Azure compares to EC2.

Oh, and I’m pretty sure this was the first real (real as in real application, not a Hello World sample) Windows Azure deployment from Finland (on the 29th of October) :).

Deploying an existing application to EC2

There are a bunch of Hello World-examples for how to create a small app and deploy it to Azure, but when I first did this deployment I couldn’t find a sample that showed how to take an existing application and deploy it. However, the mechanism behind deploying to Azure is pretty simple, and essentially what you need is a deployment project. In Visual Studio you can’t add just the deployment project to an existing solution, you always get a new web application as well. This may not seem that useful at first glance, but it’s actually not that hard to change it to deploy whatever project you want. Below you can find out how. I’m using a sample app for this article, since the project I actually used when I did this at PDC is a pretty large and complex beast.

Step 1. In your existing Silverlight application solution, add a new Cloud Service project

My sample application solution contains two projects. The first is the Silverlight application, which in this case just writes “I like clouds! on a blue background. The other project is a web application that shows the Silverlight app. No changes has been made to this project, it’s just straight out of the box.

Now, to add cloud goodness we have to cheat a bit. Currently (as far as I know) Visual Studio won’t let you add a Cloud Service project (which is needed for deploying to Azure) and have that deploy an existing web application. It always wants to create a new application for you. So what you can do is to add a new “Web Cloud Service” to your solution. Let’s call it “CloudService”.

 

 

 

Step 2. Get rid of the generated WebRole-project

When you add a new “Web Cloud Service”-project Visual Studio actually creates two projects for you. One is the deployment project (“CloudService” in this example), and the other is a web application (“CloudService_WebRole”). As far as I can tell the second one is just a regular web application.

The deployment project is set up to deploy the CloudService_WebRole-project, and I haven’t found a way to change which project it tries to deploy through the Visual Studio UI. So we need to wire it up to deploy our existing  “SilverlightCloudApplication.Web” instead.

 

 

 

Step 3. Use notepad to edit the project file.

This is the scary part. If you’ve poked around in the Visual Studio project files before you know that they are just XML files, and note even very complex ones at that. It turned out that all we need to do in order to make “CloudService” deploy any web project we want is just change which project its deployment point refer to.

So what we first need to figure out is the ID for our “SilverlightCloudApplication.Web”-project. First close Visual Studio. Then open the file “SilverlightCloudApplication.Web.csproj” in notepad (or your favorite text editor). Look for an XML element called “ProjectGuid”, and copy the contents of that element. In my sample project here it was “{61CADBB5-24ED-4F11-848B-7754877F7A06}”, but it will be something different for you project.

Next open the project file for the “CloudService”-project. Look for an XML element called “ItemGroup”, and another called “ProjectReference” below that. You need to edit three things in there. First, change the “Include”-attribute of the “ProjectReference”-element to be the path to the project of the web application you want to deploy. Next, change the contents of the “Name”-element to be the name of your web application. And finally, change the contents of the “Project”-element to be the guid you copied from the web application project file earlier.

Below you can see a screenshot that gives you an idea of what the results of your edits should look like.

Save the file, and restart Visual Studio. When you open the solution you will see that the CloudService now refers to your web application (in this case “SilverlightCloudApplication.Web"). At this point you can safely delete the “CloudService_WebRole”-project, as it is no longer needed.

 

Now you should be able to build and publish your application to Azure. Happy cloud computing.

 

 

 

Comparing Azure to Amazon EC2

A lot of people have asked me which is better, Azure or Amazon EC2? First I want to point out that this is a bit of an apples to oranges-comparison. Sure, they both are part of that fuzzy bag of stuff called “cloud services”, but from a (service) perspective they are quite different. What Amazon does is give you virtual servers that you can use to run whatever you want on, and what Azure does is give you the ability to deploy web applications to the cloud. The difference is primarily that with Amazon you still have to operate and maintain the environment your application runs in, and with Azure Microsoft is taking care of all of that for you. And they seem to do it in a very impressive way. Their environment even takes care of staging for you, so that you can test your new deployment before you switch out the currently live one.

So which one do I prefer? I really like Windows Azure. I also like Amazon EC2. I think there’s room for them both. Currently I think Azure is the better choice if what you want to do is run a web application in the cloud. Most software developers do not want to have to operate the environment, for many reasons (cost being one of the most popular ones).

If you need full control over the environment your application runs in, you probably cant use Azure (at least not in its current state). For example, I have to admit that I ended up cheating a bit when I deployed Sproodle to Azure. Because it has a database and a bunch of WCF services on top of that database I couldn’t deploy those to Azure, since it doesn’t supply you with a SQL Server. I would have had to re-architect the services so that they can run in Azure on top of SQL Services (which is a completely different animal from SQL Server). My kung fu may be strong, but not strong enough to do that in half an hour. :) So what I did was keep the services in the EC2-environment I’d set up a couple of days earlier, and set up my Silverlight app to call them from there instead. So I ended up being double-clouded, the front end running on Azure and the back-end on EC2. Not ideal, but hey, it worked! :)

The only way to predict the future is to invent it

Ok, so that was a slight paraphrasing of Alan Kay.

I just listened to Miha Kralj talk about how IT will change over the next 10 years, and why we should care about this. It was a refreshingly un-TechEd:y session, focusing more on global issues shaping our industry, and on trends that will be as disruptive as the move from mainframes was a few decades ago. I found the session to be very insightful, and I enjoyed it thoroughly. It’s great that the TechEd organizers see that it is important to make us developer types aware of what’s going on around as, and what the bigger picture is.

I won’t repeat the whole talk here, but there were a few topics that resounded with me which I want to comment a bit on.

1. How will the mySpace and facebook-generation accept the current rules of the corporate environments our generation and the one before it have created? Obviously they will not. This is something I’ve talked to several of our customers about, and I suspect it’s something that will keep us IT people very busy in the next few years. The next generation (which is actually now entering the workforce) will have demands we can’t meet, and we will have to find a way to supply it. The only problem I see is that we’ll have to meet it fast, and shipping things fast isn’t something our industry has traditionally been that good at. As an industry we need to improve greatly in order to keep our future users happy.
2. Utility computing and IT as a service. Ok, I’m sure everyone agrees that obviously this is a major disruptor. Hardware is becoming less and less interesting (count the number of hardware vendors exhibiting at TechEd if you want proof of that), and being green requires that the number of data and hosting centers in the world is reduced (and located in places where it makes sense).
3. How we interact with computers will and needs to change. One of my favorite quotes from the presentation was “Mouse and keyboards are so nineties”. I couldn’t agree more. Having used an iPhone for a few months now I wouldn’t want to go back to a non-touch enabled phone. It’s just so much faster, easier and more intuitive to use. And Windows 7, with its touch-enabled interface is something I’m really looking forward to. Let’s hope the hardware vendors get with the program quickly enough, so that there actually are good devices out there when version 7 ships.
4. Social computing is here to stay. It helps fulfill our basic human needs for interaction with others, and to be perceived by others in a way that we find fulfilling. And it is entering the workplace. The issues raised in point one of this list will make sure of that.

All of these (and the rest of the stuff in the presentation) are signs that our industry is maturing. In certain areas (hardware and hosting being one of them), gone will be the snake oil-peddlers and the tinkerers and the garage operations. Overall we will have to accept that we and those who came before us built our current computer-oriented world, but we weren’t born into it. The generations after us, the “digital natives”, will be the ones that define the rules, and the rest of us, the “digital immigrants”, will be left by the wayside, unless we make an effort both to enable our future world and to make sure we stay part of it. Otherwise this time the natives will kick out the immigrants.

Oh, and on a not-so-connected point, Miha is one of that rare breed of speakers who actually knows how to present well and how to construct the information you are presenting. Great and simple slides, balancing verbal, visual and aural communication. Please all of you who do presentations, read the book Beyond Bullet Points, and have a look at PresentationZen. You need it, seriously.

SqlCE and Entity Framework maybe broken on x64

I’ve been running on a 64 bit version of Windows for the first time the past week, and while most things have worked fine I had one problem that I had a hard time figuring out. An existing project, that worked fine before the upgrade, refused to compile. I also upgraded it to SP1 of Visual Studio 2008, so there were two possible sources for problems.

In my project I have a local SQL Compact (CE) 3.5 database

I’m using the Entity Framework in my application.

The application compiled fine, but when I tried to to run it I got the following error:

"The specified store provider cannot be found in the configuration, or is not valid."

I first thought this was related to my newly installed Vista and to moving to Visual Studio 2008 SP1. But then I stumbled across the following article, where someone else was having the same problem:

http://forums.microsoft.com/msdn/ShowPost.aspx?PostID=3422883&SiteID=1

While the post seemed to be dealing with the beta bits of SP1, it seems what it boils down to is that the provider for SQL CE is broken (missing) on x64. So what I need to do is to compile for x86 (32 bits) instead. “Runs natively in the 64-bit environment”, indeed! (borrowed from the SQL CE product page at microsoft.com)

Most of you probably know how to change this, but in case you don’t I’m including the instructions here. So in order to compile for x86 specifically (instead of “Any CPU” which is the default) you need to add a new configuration for that. The easiest way is to click on the dropdown in the VS menu:

and choose “Configuration Manager…”. In the next dialog click the dropdown in the Platform-column.

In the popup you can create a new platform (choose x86). That should be it.

Going to TechEd EMEA 2008

Tomorrow I’m flying to Barcelona to take part in TechEd EMEA 2008 for some more technical insights and some networking. I just barely got my PDC-luggage emptied and did the laundry and almost got rid of the jetlag, so I can’t really say I’m too excited about the travelling part of this trip. But TechEd is always fun and interesting, so in the end I’m happy I’m going. Barcelona is a great city, and I’m happy to get some more sun before the Finnish winter sets in properly (ok, as properly as it does in Helsinki).

I’ve got a couple of blog posts brewing that I hope to get done during the week, and of course I’ll try to post about any new and interesting stuff I come across.

Curiosity killed the cat’s computer

Ok, so installing the pre beta-version of Windows 7 that we got at PDC on my work computer maybe wasn’t the brightest idea ever. There were simply too many things that were half-baked, and prevented me from using the machine. From the past tense used in this paragraph you’ve probably by now deduced that I went back to Vista.

At least the following things happened during my brush with pre-beta Windows 7:

• Random BSD:s
• Hanging when resuming from standby or hibernation
• Windows Explorer crashing every now and then
• For the first time ever, a crash (ending in a bluescreen) that made it impossible to recover a Powerpoint document I was working on
• Problems installing applications
• Files becoming broken when copied over my home network (or copies failing for no obvious reason)

I’m not 100% sure all of these were just Windows 7-related, but the evidence is compelling.

Now, this doesn’t mean that I don’t like Windows 7. I do, and I can’t wait to get to use it.There were a lot of nice things in there. The improved Windows Explorer, for one (especially the library-feature). The presentation mode selector. The window docking. Better configuration experience (although there’s way too many icons in the Control Panel now that they’ve been relocated there). The non-polluted icon tray. Better network configuration tools. It improves the computing experience in several small ways, just like Vista did over XP.

Anyway, one should never give up, so I’ll be back when they release the first proper beta version of Windows 7. Which should be around Christmas this year.

First post from Windows 7

My Vista had gotten a serious case of bitrot, and since I got pretty hyped up at PDC 2008 I decided to take the plunge and install Windows 7 instead (the pre-beta build that was handed out at PDC).

I have to say that based on my first impressions everything seems to have gone smoothly. I haven’t installed all my applications yet, but everything I’ve tried so far seems to work. The computer in question is my HP 2701p Tablet PC, and I’ve not even had any driver trouble yet (despite this being an HP). When something goes wrong I’ll blog about it…