On-premise vs. cloud: Navigating the data center workload dilemma

This audio was created using Microsoft Azure Speech Services

A critical success factor in a data center strategy is proper planning. Data center decisions shouldn’t be reactive but thoughtful of the longer-term implications for the business. Deciding whether to place data center workloads in the cloud, on-premise, or at the edge requires an assessment of needs, space utilization, and other factors such as deployment speed, infrastructure, and cybersecurity. Companies have become better at making these decisions, thanks to lessons learned in recent years.

data center growth

It’s one of the topics that I discussed with Thomas Humphrey, North American Business Development Manager for Schneider Electric Modular Data Centers, and Todd Boucher, Founder of Leading Edge Design Group. If you’re curious, check out Part 1, which delves into why the public cloud might not always deliver the expected results and reveals how market trends have shifted strategies.

This blog post covers Part 2 of the series, which discusses what factors should be considered in reviewing a cloud vs on-premise strategy.  

Joe: What drives organizations to keep compute on premise or move it to the cloud?

Thomas: In data center planning, organizations have to identify their primary business goal by looking at data center capacity, speed of deployment, cost structure, and infrastructure. If there’s free space to use, it could be leveraged for a data center. If they don’t have the space, then how quickly do they need the additional capacity? This is where speed of deployment comes into play. Do you build the data center or use the cloud? There are questions about CapEx vs. OpEx. How do you control costs over time? Also, do you have to beef up infrastructure with fiber routes, for instance? What about cybersecurity? It comes down to assessing the business processes and picking the best solution.

Todd: With on-premise production environments, oversizing has been an issue. Compute capacity isn’t fully utilized. Virtualization helped but only partially solved the issue. Then, the cloud emerged as a way to right-size the environment. But you also have to oversize environments in the cloud for scalability. Some customers tried to move whole production environments to the cloud, leading to cost overruns. And these are the types of things you need to take into consideration.

Joe: Is right-sizing still a challenge, and how do you solve it?

Todd: Forecasting capacity is always challenging, especially given how dynamic business is today. Modularity helps address this with prefabricated data centers that you can quickly deploy where needed. Now, owners can make more thoughtful decisions: “I know I will need an asset here. What should I do with it? What should that asset be? What do I already own? What’s its lifecycle?” If customers have too much infrastructure, they realize they can downsize, become more agile, and become more modular. 

Joe: As customers repatriate some assets, what lessons have they learned from their cloud experience?

Todd: When something is transformational – like the cloud becoming commercially available – the pressure on IT is to deliver more services and more resiliency. But there were a lot of unknowns. When you have a new technology, you don’t know what the experience will be like. And one organization’s experience is different than another’s, even in the same industry. You have varying resources, different business cultures, and unique drivers for the technology. It’s not a homogeneous adoption strategy, and that’s why these lessons are so important.

Thomas: There’s also a huge variability in how companies perceive their IT. For some, like e-commerce providers, it’s core to their business; it is their lifeblood. They rely on it 100%. Other companies see it as a support function for their core business. So, you have different schools of thought. An automaker will move slower to adopt technology than a highly transactional, web-based company. This is where factors like speed of deployment and data center space availability come into play. And they’re better understood now after the initial cloud experience.

 Joe: What impact do hybrid environments that combine on-premise assets with cloud and SaaS applications have on Chief Information Officers (CIOs)?

Todd: The CIO has to work with other business leaders to determine how an application can support and scale with the business. An IT organization has to be really responsive to business capacity needs, which are difficult to forecast. Having competency and workloads in both on-premise and cloud is what makes the organization responsive. They can make quick choices about where to put resources, where to develop and test applications, and then deploy them in the right environment.

Joe: How important is the role of modularity, especially at the edge?

Thomas: When you talk about modularity and scalability, in order to be agile, you have to consider the edge. Some data must be processed close to the source, and some can be pushed to the cloud for higher-level analytics. Modularity delivered by prefab data centers helps bring it all together. You plug in the prefab data center and get increments of multimegawatts of power as you scale. It sits at the edge, providing compute where data centers typically weren’t deployed.

Stay tuned for Part 3

In Part 3 of the blog series of cloud vs. on-premise, we will wrap up the discussion by reviewing specific solution options and weaving them into the business challenges. We will address questions such as: How does an edge computing solution fit into current IT infrastructure? How should sustainability be addressed? What business justifications or use cases do we need to focus on? and more.

In the meantime, visit our website to access additional data center planning resources.

Tags: , , , ,

Add a comment

All fields are required.

This site uses Akismet to reduce spam. Learn how your comment data is processed.