Slow Down !! Cloud is Not for Everyone

Embracing a Cloud model motivated by a feisty and fierce marketing campaign is like an “Eskimo buying a refrigerator.” The sales team successfully instills an elevated fear of missing out and rushes us toward decision making. Unfortunately, such programs, taken with minimal due diligence, have backfired in the last few years, with the Industry witnessing multiple reversals and a slowing trend involving Enterprise Cloud initiatives.

Consider the case of Zynga, a gaming platform, which scurried its move to Cloud Infrastructure to save costs. With the launch of Farmville, a new game, it had exponential growth with its user base spiking from 10M to 25M in a short span of 5 months. However, Zynga was at the crossroads of “Scale Fast or Fail Fast.” Its cloud movement made sense, and it brought in the much-needed agility on high performing, scalable infrastructure. But continuing there did come with a hefty price tag. While the comfort and care of a hospital’s Intensive Care Units provide the best of care round the clock, it’s not essentially where you want to spend the rest of your life. Not only is it costlier, but it completely cripples your mobility. This truth dawned on Zynga as monthly cloud bills started kicking in. Further, Zynga’s ability to fully customize its infrastructure for creating a cross gaming platform was severely stifled. In essence, Zynga was paying a much higher bill, yet lost control of its strategy. A reverse cloud movement started with a hybrid of cloud and on-prem. With this model, Zynga leverages best of both worlds; it retains its ability to scale at will (Cloud) and has a firm grip on its ability to innovate and provide a unified gaming experience (On-Prem) at an optimal cost.

Similarly, Sony had to re-evaluate its cloud strategy when an unplanned outage brought its gaming services to its knees. In a cut-throat competition, such prolonged outages can result in irreplaceable damage to the brand and irreversible loss of customers. Sony followed Zynga by adopting a Hybrid Cloud culture.

Such outages are not a one-off scenario. This trend continues unabated even to this date. Both these cases highlight that enterprises shouldn’t yield to the lure of the Cloud Model, but should carefully weigh in on the long-term strategy and pragmatically evaluate their risk appetite before making a multi-million dollar decision.

You can’t water with a leaky bucket

Reconsider a cloud move if your tech stack has a significant stable legacy application and you are not in a rush to rewrite, refactor or retire them in the near future. Any such movement without a thought-through strategy will result in cost overruns and create architectural dents.

Consider the case of NASA. It had negotiated the compute price but underestimated the compounding cost of data leaving the cloud. Such oversight led them to a 50% cost overrun running in millions. At such precarious stages, the board will have a couple of options. Either reverse cloud movement (that will create an impression of failure) or include out-of-turn legacy modernization efforts to minimize the damage. In either case, instead of CxO controlling the reigns of IT strategy, now you become a pawn in the chessboard controlled by the whims and fancies of the cloud provider. This is not an isolated scenario, most of the mindless cloud migration happen due to the seduction and oversimplification of cost-saving, but let’s understand” “Upfront Cost saving is not a guarantee of Overall cost reduction.” It’s the same as “Renting vs Owning a car”, based on the need. Both has its set of merits and cost implications in the long term.

Most often It’s not the main course but Desserts that bloat your Bill

In the cloud, it’s not only the cost of compute and memory, but the cost of lock-in. Assume you have an on-prem license of a database enterprise edition that couldn’t be ported to the cloud (incompatibility or contractual complications or much higher cloud licenses) and you opt to move into a native DB offered by your chosen cloud provider. What might appear as straight-cut migration efforts is basically a much deeper trap of locking you in with your cloud vendor. As the first step, you need to train your workforce; then slowly, you will be mandated to rewrite or replace all the homegrown and/or SAS features of your product to be compatible with the new service. These efforts are something that was never part of your earlier plan but now has become a critical necessity to keep the lights on. Say after a certain period when you realize the cloud service is not a great fit and you decide to shift back or move-on to a better alternate there comes the insidious lock-in effect. They make such onward movement particularly difficult – you need to burn significant dollars to migrate out.

Remember, the entry paths are always wider, well-lit, and scenic, but exit paths are narrow and dark urging us to figure out the gates with bruises on the knees.

In summary, when cloud undoubtedly does bring in a new dimension on running our businesses, but embarking on this journey without meticulous due diligence and a long-term strategy is falling for a Honeytrap. Don’t fall for the fallacies. Neither every application needs to be rewritten as microservices-based architecture, nor every architecture has to be rendered serverless. So how do we decide? I recollect a response from a Fortune 100 Healthcare CIO to a question on how his enterprise was planning to leverage the emerging technologies. Without blinking, he succinctly replied that he is willing to invest in any technology that will help him to safely discharge a patient after surgery in a day less than the current average stay at the hospital. It’s true, time and again, that choices of the new tech will batter us, whether it’s Block chain, Cloud, AR/VR, etc., but the final decision needs to be driven by a sound business strategy. Such a clear vision auto-installs appropriate guard rails and not the other way around. Next time when you are lured towards the growing clout of the cloud, remember that the cloud is not for everyone.