When I talk about public clouds and how mission critical workloads are moving to run there, I am often being asked: “but how would enterprises agree to run in the public cloud, when it means they lose a lot of flexibility and control on the infrastructure, and have less optimal environment for their applications”? When one is running in the cloud, they rely on the virtual environment provided by the cloud provider, and therefore cannot tweak the application and infrastructure to perform optimally, achieve maximum performance or minimal latency, nor build their own security mechanisms. They simply need to rely on whatever the cloud provider has build and exposed to them through the APIs and services available in the cloud. This goes against what we have experienced so far, with larger organizations preferring to have full control on their infrastructure, to optimize, secure and control it by themselves.
Does this mean large organizations will not use the public cloud? Can they give up what we have always learned is a mandatory requirement for them? Reality shows us that this is not the case, and the cloud is full of large organizations’ workloads. So how can we align this with what we have learned? Why would they suddenly agree to accept less?
I believe the answer is in the balance between the various benefits and needs of the user – the fundamental question that is the foundation of every technological revolution: What do users have to give up, vs. what they would gain by doing so? A classical example would be the introduction of mobile phones. You may remember how the voice quality in the early cell phones was not that great, or worse. It was common to hear people say: “let me call you back from a landline”. The wireline telephony has just completed its migration to 64kbps digital transmission, with a crystal clear audio, while the new emerging cellular phones had much worse audio quality, with either analog transmission (1G) or lower bit-rate (12kbps) coded audio (2G) over a digital network. A conversation over the cellular network was not as clear and fluent as it used to be on landline phones, however the benefit of mobility compensated for it and cellular phones took the market by a storm. The consumers were willing to trade-off audio quality for the benefit of being able to make a phone call from practically everywhere. Twenty years later the first iPhone have made an entrance and history repeated itself. Audio quality was poor, worse than other, existing cell phones, and many questioned the viability of this new telephone with such low-quality audio capabilities. However the value of the new “smartphone” with its applications, internet connectivity and app-store has proven once again that if consumers are provided with sufficient benefits, they are willing to forego some of their original “requirements” and embrace the new technology for its new benefits.
I believe a similar trend can be observed with the public cloud. The public cloud provides users with benefits which drives them to waive some of the requirements they set for their traditional IT. They get a much simpler consumption model, with no need to worry about purchasing hardware and software, nor operating it, OPEX vs. CAPEX, a variety of services they can leverage to build their new applications etc. For all of this they are willing to give up their complete control on the infrastructure, optimization of the hardware and software for performance and scale (which is replaced by modular, scale-out design that adds more resources as needed), control over the security of the environment, and probably even lower security mechanisms (e.g. try to build an air-gap isolated recovery system in a public cloud…).
So will the public cloud “limitations” prevent larger organizations from moving workloads to it? Learning from history, it probably will not. If sufficient benefits are offered, people will move away from their existing position and will adopt a new stance, even if this means changing some fundamental assumptions or requirements they have always thought to be mandatory. Just think about the success of the new “phone with poor quality phone calls”.