Over the last few years, we’ve seen a number of companies that were previously ‘all in’ on cloud moving their data back to on-premises infrastructure. With cloud previously hailed as the silver bullet for driving flexibility and agility, many data leaders have recognised that cloud also has its fair share of drawbacks.
Regional regulations and cost are the driving force behind many of these decisions. Companies are rightly wary about the financial implications of getting governance or data sovereignty wrong. Additionally, the cost of cloud itself is another consideration; cloud can become increasingly expensive as companies scale their usage. Companies need to have a better understanding of their data and workloads before making any decisions on cloud. Ultimately, a data driven approach to planned migrations is highly recommended.
The scaling conundrum
Many organisations have recognised that running certain workloads in the cloud is significantly more expensive than initially anticipated. This is prompting them to re-evaluate where some workloads reside.
The first question data leaders should ask when they’re looking to shift data to the cloud is why? If the answer is “to save money” then they may be better off remaining on-prem, because the main benefit of cloud is flexibility. That’s not to say organisations can’t reap the storage benefits of cloud in on-prem environments. Companies are starting to understand how approaches such as containerisation and virtualisation can be achieved in their private clouds, delivering qualities such as elasticity, workload isolation and improved storage density on-prem. So, we have moved past the cloud-first ethos and are now in a workload-first era.
Decisions around whether a workload is better suited to cloud native deployment in shared public cloud, or an on-prem environment must be driven by good data. Workload analytics enable companies to observe the performance of a workload before making a call one way or the other. Workloads that are more predictable and consume a relatively stable level of resource are often cheaper to run on-prem. Whereas a customer-facing service that’s more variable may be better in the cloud because of its elasticity.
The cost of compliance
There is no denying that data compliance and governance is front of mind for many organisations, especially those operating in highly regulated sectors.
The governance landscape is becoming more complex by the day. For example, regulations like Schrems II have changed the requirements around citizen data and privacy – introducing tighter controls and steeper financial consequences. Against this backdrop, many organisations are opting to play it safe and move their data back on-prem to gain control over where data resides and ensure it doesn’t leave their jurisdiction. While the issue of sovereignty is less of an issue for companies in the US because the major cloud providers are based there, it is a growing concern for those based in EMEA and APAC.
To increase control over data, it’s crucial that companies have cohesive security policies in place across all environments. Ensuring governance is consistently applied ‘always and everywhere’ makes it significantly easier for a company to remain compliant. With one set of globally defined policies in place, enterprises can replicate security standards across all cloud and on-prem environments. This reduces risk, saves time and mitigates the risk of human error.
Optimising your cloud investments
Organisations need the capability to securely move data from cloud to cloud or from on-prem to any cloud, and vice versa. Until now this has been a challenge. But with the emergence of modern data architectures, organisations can drive more value from their data and optimise their cloud costs at the same time. This is a win-win for organisations looking to drive efficiencies in an ever-changing business climate.
Christopher Royles is EMEA Field CTO at Cloudera. He helps organisations innovate through the use of data, working across industries that are regulated and organisations where data privacy is critical. Royles’ focus is on the development of skills and methods for migration to the Enterprise Data Cloud. Royles holds a Ph.D. in Artificial Intelligence from Liverpool University which he subsequently applied to voice recognition and voice dialogue systems. Royles has advised on Government Open Data initiatives as part of the Open Data User Group (ODUG) and sat on the quick wins stream of the UK Government Cloud Program (GCloud). Previously Royles has held roles at Oracle, Pitney Bowes and Vicorp.