Reaping What You Sow From Cloud Computing in Variable Industries
Mother Nature and highly advanced cloud computing go hand in hand in one of the world’s oldest industries: agriculture. In the world of Ag, when at the mercy of the elements, insects, weather and the other variables in farming, the elastic nature of cloud computing helps bear fruit by combining cloud computing accessible anywhere at any time with advances in the technologies of GPS guidance, IoT, drone and satellite imagery, and geo-special information. Usher in Precision Agriculture to transform the unpredictable to a predictable, organized, and repeatable science.
There can be no doubt Ag requires looking across vast geographical areas and whether the data captured is performed via satellite, drone or other technologies, the end result is typically vast amounts of data. Take, for example, the challenge of acquiring and processing enhanced earth-observation data and achieving large-scale soil moisture mapping. When a single feed can result in millions of data points, even the best datacenter capacity managers are challenged to determine sizing and cost. The challenge can be met face on with cloud storage where acquisition and storage is facilitated and made cost effective.
Building a solution for an Ag cloud-based decision system from the expense consideration has taken on a different approach from the traditional capital expenditure 
How to process all that data? Determining where to run the common cloud-based decision systems for Precision Ag requires focusing on the heavily weighed points of large data volumes, disparate data types from disparate sources, geo-spatial unstructured data and seasonal demands. Take GCP, with its vast repository and powerful Google Earth Engine perfect for mapping analytics. This may make GCP a good choice, but for many with Microsoft based Azure AD and 365, the decision introduces a second cloud to manage. Adding to the complexity, many entities use large scale ERPs, such as SAP; chances are the ERPs run on a private cloud or on-prem, thereby introducing yet an additional compute source. With the growing number of compute sources, bringing data together becomes challenging and, likely, expensive. Cloud storage is cost-efficient, but moving data out of clouds to be processed, correlated or shared often increases cost significantly.
Building a solution for an Ag cloud-based decision system from the expense consideration has taken on a different approach from the traditional capital expenditure. The approach helps mitigate the risks associated with incorrectly sizing compute and storage. However, not forecasting the operational expenses can lead to a different sort of financial challenge. It is imperative to fully understand the distribution of cost across time – where as a capital expense is typically born up-front, the OpEx costs appear small and manageable up front but has the effect of increasing indefinitely unless tight controls are put in place. With the almost infinite ability to grow on public clouds, a business will grow its cloud footprint because the typical boundaries are simply not there. It is not surprising to see an OpEx-based project exceed the cost of a capital project because of that mindset – humans will occupy the capacity available.
For Ag and other industries where unpredictability is common, the risk-reducing advantages associated with capacity management and cheap storage, coupled with the ability to amalgamate disparate data and their sources, definitely allows cloud computing to outweigh the limited capital expenditure route of on-prem systems. Yet, each implementation must account for the full ecosystem and how it will interact with downstream, ancillary, and transactional systems to reap the benefits of cloud while accounting for the bottom line.
