The Cloud Formerly Known As Azure
You sir, owe me a new keyboard.
If you are thinking about deploying .NET applications on a platform cloud and whacking them against an SQL Server database embedded in that platform cloud, Microsoft's The Cloud Formerly Known As Azure is not your only option. Amazon Web Services has fired up its own analog to Azure. The two new services, Amazon RDS for SQL …
Are you sure?
I suspect Amazon are saying give me all of your data and I will help you reduce your long term costs by moving you to these cheaper, non-MS database options. We'll even provide helpful migration serivces to get rid of those expensive licencing requirements... That's a much more business friendly option than "migrate all of your software to our platform to save money".
MS jumped into "the cloud" because they saw companies were providing business models that matched what there customers were demanding (i.e. scalable usage models, per-user costs that could grow or shrink with the business, high availability for smaller organisations) rather than to directly fight off open source.
Do you really think the "cloud" is anything other than client-server?
If you do then try not listening to the marketing hype. Cloud is still client-server the server is just not yours. I still have yet to see something done in the cloud that cannot be done with what you call "client-server".
But maybe I am just ignorant.
But Azure has always been shit. SQL Server was (maybe still is) nobbled on it where as on AWS it's fully functional (with more besides).
As for web applications - I'd much rather be able to RDP on and have access to all the server stuff (as you'd been able to do with AWS as long as I've been using it) than have to use the horrible, limited silverlight mess that Azure relies on. Not to mention that you have to redeploy the WHOLE thing just to make a single (persistent) change.
Basically - Amazon aren't the underdog here so they don't need to "take on" Microsoft. If anything it's MS who are playing catch up.
Azure 1.3 does have the ability to use Remote Desktop Services to access Windows Azure roles.
If you use a VM role you basically have complete control of a virtual server so if you can't upgrade part of your Application then that’s your fault not Azure's. As for making changes to applications built using web or worker roles surely that depends on how you structured the application. If you build all functionality into a single worker role then it’s not surprising you have to redeploy the whole thing to make a single change. If you split functionality up a bit into separate components then it supports on line upgrades of each component separately .
The scalability is coming from the way you build your applications from service based components connected via asynchronous messaging. It has nothing to do with Windows as an OS. Azure isn't even the same OS as Windows. Sure, It's able to host VM's that can run Windows Server or Linux, but mostly you would be running components on the .NET Framework and the Common Language Runtime, you won't use Windows OS at all.
Building your application on Unix or Linux won't be any more scalable if you don't use the appropriate design patterns.
A PKI certificate with a bad date can cause any system to stop working, I've seen it happen on BEA Weblogic runing on HP UX. This is an availability and sevicability issue not a scalability issue.
Actually, Azure's web and worker roles are vanilla Windows Server VMs plus some services to manage the infrastructure for you. Don't believe me? RDP in and look around the filesystem.
It's been like that since at least when I started developing to Azure 1.2. No RDP at the time, but with the right code you discover many things :)
Amazon is dominating the cloud computing market. One place they cannot compete with Microsoft is in developing their own sophisticated software services, like MS Dynamics. And of course they are tying Azure services in with Win8.
Azure itself is pretty cool OS technology. Culter's team spent a year just studying the formal requirements for geographically distributed data centers. They merged technology from NT, Hypervisor and (I'm guessing) the SQLOS layer. The scalability tests I've seen are surprising, way better than Linux and even better than NT at multicore parallel performance.
What makes me skeptical about Microsoft is their executive management. I still would not be surprised if they fail. Their lasting impact may be the disporia of former Microsoft systems engineers, who are scattered throughout the industry today (for example, running EC2 at amazon). Systems engineering rules. You don't learn it in college, and it's not part of the open source culture, so if you want to built something big and complex, you gotta find those industrial experts.
Biting the hand that feeds IT © 1998–2019