Offer me a Windows Azure dedicated cloud, either hosted on my premises or with third-party provider.
You know Azure is going to be down at some point... it'd be great to have a way to have 100% replicated on-premise Azure backup where data-replication happens and can handle up to 6 hours of downtime if it happens. I don't want to build a on-premise data center that's different than Azure cloud as backup...
Appliance based private cloud that will run within an enterprise security perimeter so we could own our data. This alleviates the data security worries for enterprises and this infrastructure could be scaled out to Azure public cloud on need basis so we use public Azure only for predictable spikes during marketing campaigns or unpredictable spikes in a short time frame. Believe me this could really DRIVE enterprise adoption for Azure.
Earlier this year, we announced the Windows Azure Platform Appliance, which will allow customers to run the Windows Azure environment in their own data centers or via a third-party hosting provider. Look for more announcements about the Platform Appliance in 2011.
The "Windows Azure Server 2010" idea from Rob below is what I need. My business develops sites for start-ups. We need a way to host on our hardware while they are small, but easily switch to the cloud if and when they need it. Current Azure pricing is prohibitive when we can simply host a traditional ASP.Net site on our own hardware.
Kenik Hassel commented
Check out StorSimple - their appliance allows tiering of on premise data, deduplication and ultimately, storing low-ranked data on Azure. Configuration allows full concurrency betweenon prem and in Azure, as well.
Matt Houser commented
I'd like to see a Windows Azure software installation that i can install on a hosted VPS that simulates the Azure environment without all the scalability. This will allow me to develop against Azure and run it using a cheap host, but when if my app needs it, I can deploy it to the real Azure environment without code changes.
This will help smaller developers and startups who are fighting against the existing cost of Azure deployment.
Frederick Thompson commented
yep - even speeding up the rate of deployment to live servers for testing would be an unreal reason to allow dedicated clouds to be available. currently there are BIG issues with the fact that the development enviro != the platform environment and this would help fix this.
Seems like this is now implemented by this microsoft's announcement today. http://www.microsoft.com/windowsazure/appliance/
Tim Huffam commented
Hi Mike, as the other guys say, the big benefits for this are:
- allowing end consumers to be able to have their data/services hosted with a partner they trust/prefer (this is particularly important to the finance, health and insurance sectors).
- allowing app owners to start small and local, while having the ability to scale and/or move to the Microsoft datacenters with minimal effort/redevelopment.
I'd also like to add that connectivity is a significant issue/hurdle for many cloud/remote services - with regards to reliability and bandwidth. Here in New Zealand both of these are high risks for any remote services - particularly those offshore. Local datacenters overcome this.
folks, I have created a new request "An appliance based private cloud for enterprises that dont trust public cloud" that is kinda similar to this, except it has a how-to along with the request. Feel free to comment on that as well.
Nothing more to add
we want to use similar way to access data on azure and on premise.
Wouter van Eck commented
I don't understand why you would want to have a "cloud" locally. Besides if you want to have Azure local then nothing stops you today. Windows Server comes with message queuing, there is SQL Server as a product and you can already download the AppFabric as a product. What else do you want? besides headaches?
I agree with vidalsasoon, hosting the cloud on your own data center totally defeats the purpose of the cloud. I could see how you might want this type of architecture on your datacenter to allow you to scale up easily when demand increases, but even then scaling up would mean investing in new hardware which is what metered services like this are designed to avoid.
Doesn't hosting your own "cloud" defeat the purpose? there will be no smooth scalability if hosted from your home...
Robert Shurtleff commented
This thread is also an enabler for a multi-tier development stack. My current customer is not happy with having to put the entire dev stack (and thus pay for it) on the production cloud. They really want to be able to self-host, on-premise host, a "mini-cloud" on their own hardware through the SDLC up to production. They they would deploy to actual cloud for production.
And, using the VS emulator is just not really an option for anything above an individual developer.
Their common dev stack:
- Dev Integration Testing
David Burrell commented
I agree with Shane, some companies will not allow their data to be stored outside of their datacenters. I know the point about Azure is Microsoft running the datacenter, but companies should have the ability to install Azure services on their own servers, for development and for production.
In our case, we have customers that are VERY sensitive to thier data being anywhere but in our datacenters. Being able to host Windows Azure On-Premise would give us our own cloud but keep our customers happy knowing that thier data is only being handled by the people they are in contract with. I would really love to move forward with Windows Azure but the fact that it's hosted by Microsoft is stopping the whole process with upper management.
Federico Lois commented
Mike, investing on a Cloud enabled application is certainly a cost that you have to pay upfront for something that you still dont know if you are confortable enough (or even make sense) to out-source "on the cloud" yet. On the other hand, having an "Azure" environment on premise with my limited IT capacity that can be federated and/or migrated to the Microsoft Datacenters because I either need global reach or its scalability would be a selling point to even the most computer illiterate decision maker.
So it is both a technical issue and a business issue.
- Spike handling,
- Consistent development environment,
- Not enough on-premise hardware in house
- I don't want to have that much hardware in-house and still be able to perform some of the work in-house with my own spare cycles/hardware.
- I don't think I can convince my CEO/CIO to invest in something they can only run in the providers data center and not in-house. (certainly some of the tech that you are providing is useful not only in cloud environments).
Thanks Rob and Mamby, appreciate your comments here. So for you guys, this is primarily about ensuring the local development environment provides a consistent experience with what will happen in the cloud?
Other motivations for suggesting this? Anyone's motivation for suggesting it to actually run (in Ron's words) "big multi server complicated fabric thing" on their own premises as opposed to running it in one of Microsoft's datacenters?
Thanks for helping me undertand your needs here - Mike
This doesnt have to be a big multi server complicated fabric thing. A single server, production ready version of dev fabric called "Windows Azure Server 2010" would be fine.
We'd have a single consistent architecture for developing new apps with the flexibility to deploy them locally or in the cloud.