
"public clouds are a means, not an end"
Amen to that ! Thank God there is still at least one person capable of speaking with the voice of sanity in this cloud-boggled world.
Now, if only manglement could realize . . .
Dell EMC will release a four-node Azure stack. This potentially places Microsoft's cloud-in-a-can within reach of the SMB. The subscription model thing could get in the way. Microsoft recently released pricing information for Azure Stack, loudly and proudly championing the subscription model. This is different from what many …
For buying the hardware and running it on our dime we are to get slightly lower prices for our on-premises instances. Anyone willing to give odds on just how much lower?
However much lower it is, I'm guessing it'll be 10% higher than "public" cloud as soon as they feel they can get way with it.
The Azure Stack is NOT hybrid cloud. It is a standardized pod that allows developers to use Azure Services in on-prem instances. Too many people think it's a way of running a Hyper-V stack that allows them to push stuff into Azure. It is not. And Dell's offering is nothing special. Anyone can build it using off the shelf components.
Do you work for Dell? I work for none, thanks. I just happen to have spent some time looking into Azure Stack and I fail to see what the big differentiator is supposed to be?
According to MS you can get validated on premises hardware from HPE, Dell, Lenovo, and Cisco. But when you look at it, it's just a pair of 10G switches, four to eight servers with certain amount of cores and RAM, and internal drives for MS Storage Spaces.
The magic though is in the services stack and platforms offered by MS to manage VMs and containers. It's really just MS Azure PaaS and that can run on any hardware.
So could you please stop with the ad hominem and instead in some way address the question? If I want to run Azure PaaS = Azure Stack, what specific features do I get if I chose Dell or HPE or Cisco or Lenovo that I do not get with the others? Because so far, it doesn't look like it matters at all which vendor you pick.
I suspect that feature parity will be equal across vendors, and the decision will come down to a combination of cost, the company with whom you have the best working relationship, and perhaps lastly, which vendor you've chosen historically. In some cases this may be moot, as Azure Stack is being sold as a 'hands off' black box (from a hardware perspective), with the argument that this frees up time to spend on developing tools delivering business value.
I agree that on-premise IT will always be with us. It's a cost-effective way to go when your utilization is very high and consistent. I think Trevor is massively underestimating the cost, complexity and risk of building and maintaining some kind of cloud-neutral software layer that works for most of your workloads. This approach can certainly work for open-source based, generic workloads, but for example big Oracle database workloads or complex, interconnected Microsoft applications, I don't think so.
And you have to look at it from senior management's perspective: they see the IT department as a competitor to Oracle/Google/Microsoft/AWS cloud services. Also the complicated, high-maintenance infrastructure to attempt to build to somehow/maybe commoditize clouds sometime in the distant future, is money they could spend on business-impacting applications built on PaaS or SaaS services in the public cloud.
Going forward, local infrastructure IT is going to have to compete with the public cloud providers on how they add to enterprise/SMB application agility. Infrastructure is really just table stakes at this point.