The view from the cloud

I am at 30,000 ft going about 200 miles per hour and I look out my window and sIMG_0998ee this.

Many of you have seen clouds before from both above and below, but have you thought about those opposing views when it comes to the IT version of the cloud?

From the ground a solid cloud cover can look flat with few ripples and honestly looks a lot like a blanket be it white, grey or black but it can block out the sun and bring trouble. From above the cloud deck the clouds look different even storm clouds seem thin and light like wispy mountains of marshmallow fluff. Ok maybe there is some poetic license taken there but you get the drift. From above the clouds have peaks and wisps. The gaps between the clouds are more visible because the contrasting color of the ground, sometimes the ground isn’t even visible at all.  When traveling through the clouds on the way to cruising altitude there is typically turbulence as you hit the air pockets and temperature differences that help clouds form in the first place.

IMG_0999  I can hear you all now saying what does this have to do with IT and why am I reading your crazy ramblings again.

Here’s the thing from the ground we view the cloud as a far away concept, we think if we put things up in the cloud we may not be able to touch them but they are also part of a bigger picture and the cloud can scale to meet the need and we are just part of a big flat bank of processes and compute. While service providers and cloud brokers certainly want you to think that I want you to take a moment and think about the view of those workloads from the cloud perspective.

It’s like the concept of not being able to see the forest through the trees but in reverse. You can see the cloudbanks but might not make out the individual clouds that make it up. Which is good but those clouds show peaks and valleys and yes even holes. That means that the service providers are seeing some workloads take a significant amount of cycles or suddenly spike.

In other words the cloud ops folks are dealing with the same issues that you have seen in your own datacenters but on a way larger scale and with less insight into the workloads they are running. Ideally the have elasticity built into their environments and the extensibility to provide bursting to allow for your marketing or ERP apps to suddenly draw more power, compute and memory. But in the end it’s a matter of control, both for you and for the cloud provider.

The turbulence is another factor to consider, while most take offs and landing are now done fly by wire, private pilots and those of smaller planes still look for the best break in the clouds to avoid the bumps. The same is true when selecting a cloud provider you have to look at technologies used and use cases to ensure a smooth transition of workloads from your DC to the cloud. Moving workloads between clouds is the same there can be bumps in the air because of the two cloud providers having different offerings or even technology lock-in due to proprietary solutions.

So I will pose a different question; what if as an IT organization you retain control of your “cloud” even in a hybrid environment where you have your private datacenters tied to a public infrastructure. Think VMware’s Hybrid Cloud Services (vCHS) and what if you fronted that with your service catalog using something like VMware’s Cloud Automation Center (vCAC). Sure we have all talked this thing to death, but then in stead of just providing these two locations for your cloud customers, you include AWS for deploying Dev environments, and maybe even a hosted private cloud offering where you pay OPEX dollars for the equipment rather than make huge CAPEX investments to get your datacenter up to “cloud ready” status?  There can even be a case made for some workloads to be in Azure, but wouldn’t you feel better if you could look at your business needs and determine that fit based off of costs to the business and specific workload needs?

There are more options than just the traditional workload or even the new hybrid offerings. It’s about workload and risk mitigation. In the end you have to look at the cloud from both sides to figure out where the breaks in the bank are to reduce the turbulence of your move to the cloud.