Okay, so maybe I didn’t quite “lose” my job to AWS, but after reading a few articles published in 2014, it became evident how some IT practitioners could be apprehensive about recommending the delivery of a public cloud infrastructure to their companies. It’s a damned-if-you-do-damned-if-you-don’t decision. Damned if you do, and the benefits of cloud aren’t all they’re chalked up to be; damned if you don’t, and your business opts for the cloud anyhow… Looks like we’ll need fewer on-prem sys admins…
For executives, there seem to be two prevailing obsessions when it comes to the appeal of the cloud: security and cost.
If It’s Good Enough for the CIA…
First, CIO’s read that the CIA uses Amazon and then begin to ask their IT staff, “If Amazon AWS (cloud) is secure enough for the CIA, why can’t we use it?” Now, obviously the CIA isn’t using the cloud to run its top-secret applications or store top-secret data, but the sheer presence of anything-CIA in the Amazon cloud does make people ask “How unsecure could it be if the CIA is using it?”
Secondly, CIO’s will tell their IT staff that they “Are not in the data center business,” and that by leveraging cloud computing they could lower costs.
Problematic to both positions, however, is that cloud computing has of yet neither proven itself more secure (Heartbleed, anyone?) nor more economical than on-prem infrastructures.
So if you’re the guy (or gal) running IT on a day to day basis, and you know the CIA isn’t running its most secure apps in the cloud, and you know that running your businesses applications in the cloud could lower service levelsand will not save money in the long run, how do you delicately tell your boss that their perceptions of the cloud are misguided?
As with all things, the answer to whether or not cloud service adoption is right for you is: it depends. Some contend that the clouds of AWS or Microsoft Azure are perfect for startups not willing to marry and maintain a private cloud – at least not in the beginning. At the opposite end of the spectrum, some aver that AWS - when properly configured – can deliver enterprise applications to mature enterprises just as effectively as on-prem infrastructures.
In both scenarios, intelligent businesses would scrutinize their monthly AWS billing, because at a certain point, ease of deployment equates to ease of spending, and building out a private cloud becomes the unequivocal best option.
Cloud 2.0: Commoditizing Cloud via Price War
So now forget all of that.
In March, Google – for many reasons – instigated a cloud price war, slashing prices by 85% on PaaS and BigQuery. As if on cue, Amazon cut prices in same-segment products accordingly, and cuts to Azure followed soon thereafter.
In this era TechCrunch has named Cloud 2.0, the only constant is the continuing commoditization of cloud services. Though the price changes will vary across instance configurations, a simple and memorable rule of thumb is “down.” What this means for you, Mr. or Ms. Private Cloud owner, is that unless you find a way soon to approach the economics of Cloud 2.0, your job may be at risk.
What You Can Control
At the end of the day, the win lies in the ability to control your own destiny, and when you have control, you can lower costs by maximizing efficiency.
Whether this control exists in the cloud or on premises, IT really cares about one thing: quality of service (QoS), or better put, quality of service for the right price, (QoSftrp) anyone can deliver QoS on an unlimited budget. When you have more control, you have the flexibility to drive down costs while at the same time ensuring delivery of the proper SLA’s to application consumers.
How can you ensure proper service levels and control costs? As I said in my piece about The Desired State of IT, the best approach is to leverage automation in some way. Only a few short years ago, there were no automation tools to help drive infrastructure - at least not any with an understanding of how to drive a balance between application SLA’s and cost.
True automation first requires a good understanding of virtualization. Almost every infrastructure today is virtualized to some extent. By 2015, the global server virtualization rate is projected to be 59 percent. Second, it requires an end-to-end understanding of all of the physical and virtual components that comprise today’s new virtual infrastructure.
Additionally, today’s virtual ecosystems continue to evolve through innovations such as converged infrastructure and hyper-convergence based on “server SAN” capabilities. It takes quite a bit to stay on top of all of this technology when at the end of the day, IT is just trying to deliver SLA’s.
It’s not to say that cloud computing is bad, nor are any of these emerging technologies, but sometimes I don’t think vendors understand what IT deals with on a daily basis in order to keep the tires on the bus and the bus moving forward. IT doesn’t need to change their infrastructure every 3 years in order to “keep up,” nor can they really afford to do so (by the way, cloud providers are faced with these same challenges as well, the delicate balance of innovation, services and costs). Sometimes it makes sense to slow down, take a deep breath, and ask yourself, “What am I really trying to accomplish?” and search for tools or capabilities that allow you to accomplish that task.
Delivering intelligent workload management for your virtual infrastructure to ensure SLA’s (at the right price) doesn’t necessarily require flash/SSD storage, server SAN capabilities, hyper-convergence or, cloud. There are software capabilities available today that help to deliver true quality of service for your applications without a requisite redesign of your data center.
Additionally, these tools can help you to ensure you are not over provisioning your servers or storage – and by adding automation you can avoid spending valuable IT management resources on troubleshooting or architecting new virtual infrastructures. By not over provisioning and not wasting valuable IT time, you can save money, take control, and deliver application level quality of service. So, in some cases, AWS (cloud) may be a good fit for some workloads, in some instances (bursting to the cloud for example) but to look at cloud as the savior to all your IT budget issues isn’t the right answer.