Where to put your data best: in the cloud, or on premises? Actually, that’s not the main issue for those who want to get value from their data. Even the data platform you choose doesn’t play that big a role. But do try to make the right choice, or else…
You cannot extract value from your data without being in control of the data. With that comes a choice: do you opt for data on-premises, or do you embrace the cloud? “In principle, you can do both,” notes Christophe Robyns, Managing Partner at Agilytic. He joins Mathias Coopmans, Cloud and Architecture Lead at SAS, Ziad A Fayad, Data Cloud Specialist Lead at Salesforce and Thijs Paepen, Account Manager at Ctac, round the table. Together, the experts discuss the role of data within companies, its management, and how to ultimately unlock data to obtain value.
Important and irrelevant
The question around on-premises or the cloud is very important in that context, but, at the same time, not at all. “In fact, where your data resides is irrelevant,” says Coopmans. “There are technological and financial reasons that come into play, and also requirements around regulation drive the discussion. Whatever the outcome, technology that works in the cloud can also work on-premises.”
Coopmans doesn’t think the discussion is unimportant, but thinks the real unlocking of data only begins when you start looking at the rest of the technology stack. Proper implementation of that will ensure good results, whether on-premises or in the cloud.
Forced choice
Paepen likes to nuance that a bit. “The choice between on-premises and the cloud indeed depends on the specific needs and constraints of an organization, and both on-premises and the cloud are valid options. That will always be the case. On-premises will remain relevant for specific cases.”
“Companies are not literally being forced to go to the cloud,” he continues. “But in practice, you have little real choice. If you want to use the latest modern capabilities and features, and enjoy all the new goodies, you have to choose the cloud. Moreover, more and more companies are moving their solutions completely to the cloud. Just look at SAP. You don’t have to migrate with it. But try and find an ERP system today that is built exclusively for on-premises environments.”
Exception
“Cloud is becoming the standard,” Fayad concurs. “On-premises will always be possible, but is increasingly becoming the exception.” A hybrid approach is currently common within organizations’ overall IT environment, but is not common as far as the data platform is concerned, he says. “The hybrid situation is usually the result of a gradual migration to cloud. The hybrid data landscape can still be found, but more and more companies are steering toward cloud.”
On-premises will always remain possible, but is increasingly becoming the exception.
Ziad A Fayad, Data Cloud Specialist Lead at Salesforce
According to the experts around the table, the choice of on-premises or cloud is driven by several logical factors, such as regulation, and sometimes less logical factors such as the preference of certain people and departments. The cloud offers the most flexibility and the latest capabilities, but in practice everything gets neatly connected.
‘All the same’
More important, then, is the next layer: that of the data platform. Paepen illustrates, “Companies often no longer ask whether a platform is in the cloud or not. They choose the capabilities of the platform and don’t wonder whether the solution is available on-premises.”
read also
Choosing is not losing in a hybrid cloud
The choice of the right data platform is dictated by needs, although they are not necessarily purely functionality-related. “Actually, all major data platforms can do ninety percent of the same thing,” notes Robyns, and the rest of the table nods in agreement. “Other factors are more decisive. If, as an organization, you are already largely in the Microsoft ecosystem, then a solution within Azure will look attractive. Finding people with expertise is very difficult today. By choosing what is familiar, organizations can reuse the knowledge they already have.”
Zero copy
“Ultimately, all systems are connected to each other anyway,” Fayad reiterates. “With zero-ETL, or zero copy in the vernacular, environments are well integrated with each other. What matters most of all is what you build on top of the data platform and how quickly people can get to work with it.”
Fayad points to the importance of a platform that is close to the user, while Robyns points out that it is actually possible, with APIs or other modern data integration capabilities, to glue just about anything together. The foundation has to be solid, but what you do is priority.
What is the context
The discussion of on-premises, cloud or the ideal data platform is interesting to techies and is part of data strategy. However, Robyns insists that the data foundation should by no means be the starting point of a discussion on AI, analytics or any other plan to extract value from data.
You should never do IT projects for the sake of IT.
Christophe Robyns, Managing Partner at Agilytic
“You should never do IT projects for the sake of IT,” he explains. “In every conversation, you have to start from business objectives. What is needed? What problem do you want to solve?” Everyone else agrees. “It is essential to identify and understand the business question before you start collecting and analyzing data. Those who start projects without a clear understanding of the business context run the risk of developing solutions that do not meet the needs of the business.”
Marginal choice
You can solve a business objective by extracting value from data. To do that, that data has to be somewhere and you need a platform. But the data, servers, cloud solutions and platforms must always serve the problem you want to solve. In that respect, the debate about on-premises or cloud is one at the margins, as Coopmans points out. But what you decide will still help determine how you tackle future problems.
In other words, it doesn’t matter whether you choose the cloud or not. But to be safe, choose the cloud anyway, circumstances permitting.
This is the second editorial in a series of three on the topic of data and analytics. Click on our theme page to see all the articles from the roundtable, the video and our partners.