TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Considerations on the Path to Cloud Acceptance

Thinking about three key aspects of the cloud can accelerate your adoption of the technology.

• By William McKnight
• April 1, 2019

No matter which "stages of grief" model you buy into, the idea has importance for individuals and companies alike. We're only human (or, ahem, a company), so we will go through similar stages when our big moves forward become painful. The key is how long it takes to go from, say, denial to acceptance.

For Further Reading: The Cloud Can Be a Massive Data Source for Analytics Data Management Best Practices for Cloud and Hybrid Architectures Cloud Shift Is Real, Happening, Imminent -- and Behind Schedule?

I see this happening at organizations with respect to the cloud. Organizations can tend to act as if they are at an early stage of grief (such as denial) when a quorum -- or sometimes just one key individual -- continues to hold out on the inevitable cloud push. Once an organization accepts the cloud and overtly acknowledges its importance, the organization's cloud plans become more solid, better architected, higher performing, and more secure. There is more "grist for the mill" and you can get on with a sensible cloud program.

Three Considerations for an Acceptable Cloud Program

Compliance. Your industry, government, or company direction could include restrictions on data location and certification. Clouds should be analyzed for a fit from this perspective.

Bandwidth. We can talk all we want about the benefits of the cloud, but one critical factor resides squarely on our organization's shoulders: bandwidth. The cloud can get a good or bad reputation based on this factor. Although some find latency intolerable and object to the high cost for moving data in and out of the cloud, others with high-speed or private networks can enjoy better performance than on premises! With cloud acceptance, you naturally stop drinking your data out of a tiny straw.

Costs. Yes, it's OPEX not CAPEX and yes, it's less up-front outlay, but you can easily outspend the on-premises equivalent if your cloud environment is poorly architected. You may well need to locate data, or portions of data, in multiple regions to save on transfer costs.

The economics of the cloud vary depending on the model (SaaS, IaaS, PaaS), the software itself, and the chosen instance specifications. There are costs incurred just to have data in the cloud and use that data (even to back it up). For example, if you're not compressing and using lower cost storage tiering for your backups, your overall cost could be 20 percent higher.

I've found immature companies with respect to the cloud are usually doing a few things that escalate costs. Applications moving to the cloud need to be architected to work with the cloud cost model in the most efficient way.

The Future Is Cloud-Based

I believe the model of the future is not for each organization to have its own data center(s). Most companies will be in the cloud in a big way within the next five years. Issues such as compliance, bandwidth, and cost are becoming less of a concern as providers and client organizations have brought the platforms up to standards. Use of the cloud allows organizations to be more agile and meet increasing customer demands. The cloud opens up a world of service possibilities.

Accepting the cloud will not create a perfect implementation, but it will greatly accelerate your path forward.