The cloud is here to stay and has, over the past couple of decades become the dominant location for data and processing. But on-premises systems still have a role to play, and there are no signs of that changing anytime soon. Some technology vendors like Box are all in with the cloud; there are no on-premises options. For many small and mid-sized firms, there is no remaining logic in running on-premises servers and storage arrays; the business case is overwhelming for them to use cloud services.
In the early days, there was concern that Cloud computing did not secure data well enough for enterprise needs. Today that argument is redundant for all but the most secret and locked data. But what was once seen as low cost, indeed for many, its most significant selling point was low cost is now coming under scrutiny. Similarly, the historic and inevitable limitation of bandwidth, in other words, no matter how powerful your cloud provider, your service is only as good as your connection to the cloud, is also coming under renewed scrutiny. Larger enterprises and public sector clients are starting to ask some hard questions about the suitability of the cloud for their computing requirements, and some are starting to do the previously unthinkable, move their ‘stuff’ back on-premises.
It’s early days as we like to say, but the two big drivers for this nascent reversal are storage costs and the use growing use of AI. A decade ago, if you had a hundred million files sitting around that you could not get rid of but were seldom accessed, it seemed logical to move them from on-premises to a lower-cost cloud environment. Now that many are dealing with not millions but billions of files, that math is looking shaky. To put it another way, if you have a few Terabytes of data then the cloud option is often a no-brainer – if you have petabytes, you may need to think again as on-premises storage may well be cheaper.
Popular discussions around AI typically focus on things like Deep Learning, which uses vast amounts of data and processing power. Deep Learning is clever and critical for some use cases, but it’s overkill for many. Moreover, for very specialized use cases such as in healthcare and pharma, though Deep Learning may be the only route available, using the cloud may not be the optimal route. Rather if you have deep pockets to fund Deep Learning, you may be better off with a more localized and dedicated data center setup. Some might call this a private cloud; we prefer the term on-premises.
Add to this Cloud promises are not always kept; for example, both Google and Oracles London data centers recently crashed due to unseasonably hot weather in London. Add to this well-reported, if sensationalized at times, security breaches and there is little wonder that some firms are keen to keep their most sensitive data close at hand, and yes, on-premises.
Interestingly, in many, if not all of our briefings, we ask technology vendors how much of their business is in the cloud and how much is on-premises. For some, the answer is 100% cloud, but you would be surprised how many planned to be cloud-only by 2022 but still have large (and in some cases growing) numbers of clients opting for or demanding on-premises deployments.
To use and abuse that awful metaphor, Cloud may well be the 800lb gorilla, but its opponent is in the tenth round and starting to stage a fightback. Ultimately we may end up with less of a knockout punch and more of a split decision. Cloud is here to stay, but on-premises is not going away anytime soon as it has a vital role and may see a serious uptick of interest over the next few years.
If you are a buyer or user of Information Management & Automation technology and wondering whether to move to the cloud or not, we have a free – yup, totally free – confidential advisory service for you. A service that means all you have to do is to find a time on our calendar.
If you are a technology seller, then, like many others, you may want to reassess your current business, pricing, or go-to-market strategies to adapt to these turbulent times. Again feel free to reach out and chat with us as we may be able to help by providing some critical but constructive advice and support.