Once again, the pundits are lauding the new year as the Year of the Public Cloud. This seems the equivalent of the emperor’s new clothes. The Year of VDI having gone out of fashion, it is all about EUC now, you know.
Before we delve into the “whos, whats, and whyfors” of our annual sojourn into crystal balldom, I feel a quick review is in order of what the main drivers are for moving to a public cloud, whether infrastructure, platform, or SaaS.
Public Cloud Drivers?
The main and most obvious driver is the desire to reduce costs to the business: to do more with less and to find ways to improve the bottom line. At first glance, moving from a “buy now, amortize over five years, with support costs annually” model to a “pay-as-you-use” model similar to that for utility usage may seem better value. However, the fact is that although initial setup costs may seem low, ongoing usage costs quickly ramp up.
This leads us to consider investment. Decent on-premises infrastructure is costly. Upkeep and improvement always seem to be more expensive than initial estimates indicated. I have my theories about this, but they are another article entirely. Suffice to say that a little more openness during bid processes and more clarity on costs would seriously mitigate the issue. The lowest bid is never the cheapest option.
Scalability is an oft-vaunted benefit of public cloud. Companies like AWS, Google, and Azure have made huge investments in infrastructure. They have data centers across the globe, which gives them enormous resilience. This sort of investment comes at a massive cost to in-house IT departments. Only the truly gargantuan can afford that sort of investment. As a byproduct of this massive scalability, products like Hadoop and associated analytic products like Tableau, Qlic Sense, and Snowflake Elastic Data Warehouse seem to be custom-made for public cloud. The only issues here, to my mind, are those of data security and safe harbor.
This brings me to the last benefit I am going to look at: security. When Azure, AWS, GCS, and those other public cloud providers first appeared on the scene, one of the biggest arguments against moving to a public cloud was security. It was an unknown entity: a black box, so to speak. To some, this is worrying. They like to know what is going on behind the scenes. However, the fact is, public clouds have to have rock-solid and fluid defense systems. They have multiple tenants on their infrastructure, so they need solid walls to prevent data leakage between tenants. They are continually being attacked, so they need solid perimeter defenses. A public cloud provider has to have these things baked in from inception, not added as bolt-ons. Security is in their DNA.
Public Cloud Issues!
Now, not all is rosy in this garden. There have been horror stories of what happens when things go wrong. My article here highlights that to great effect. Companies that are purely cloud-based have gone to the wall due to process failures.
Other issues include “legacy” applications. Those that the majority of companies still use are not optimized for public cloud–based environments. Third-platform applications are not truly ready for the mainstream; it is these apps that will move public cloud from the Steam Age to the Nuclear Age.
Public Cloud Barriers
To me, though, the biggest barrier to public cloud adoption is still trust. It is your data, your company’s crown jewels, that you are putting on the line. Historically, in our industrial past, research and design, patents, and blueprints for manufacturing processes were the crown jewels of businesses. These were and still are heavily guarded. Today, it is information that guides us. Our decision-making processes are influenced by ready access to up-to-date and accurate information. If you fool yourself into thinking that your competitors are not attempting to gain access to that data to steal a march, then you are misguided. That said, public cloud is the industrialization of computing. Just as some companies still have their own electricity-generation capabilities, water-filtration plants, and gas pumps, some companies will still have on-premises infrastructure.
So, if it seems that I am saying that the rise of public cloud is inevitable, yes, I am. But is that year 2016? No. We are still a long way from there, no matter what the pundits say. There are still vast numbers of companies that have not even virtualized yet. That said, I believe that those companies will most likely miss out on that step and move directly to the cloud. SMBs are moving there in droves, using products like Office 365 and not even realizing that they are using the cloud—only knowing that their data is safe online if their laptop fails and that Riley from Accounts can access the same document in real-time while working from home. This is the future. At the enterprise level, IaaS and PaaS are just stepping stones. SaaS is the endpoint. However, at the enterprise level we are as near to that endpoint as we are to flying to Alpha Centauri and back again in a lifetime. Why? Simple: because companies are unique. They have their own processes, and they have their own policies. These are very difficult to condense into a single program, and they are why internal IT departments came about in the first place. Customization is hard. Customization is expensive. But customization is necessary, and this, not trust, is the real reason why neither 2016 nor even 2017 will be the Year of the Public Cloud in the enterprise arena.
Tom Howarth is an IT Veteran of over 20 years experience and is the owner of PlanetVM.Net Ltd, Tom is a moderator of the VMware Communities forum. He is a contributing author on VMware vSphere(TM) and Virtual Infrastructure Security: Securing ESX and the Virtual Environment, and the forthcoming vSphere a Quick Guide. He regularly does huge virtualization projects for enterprises in the U.K. and elsewhere in EMEA. Tom was Elected vExpert for 2008 and each subsequent year thereafter.