In the Cloud, or on the Ground? Considerations for Cloud and On-Prem. Environments

December 10, 2018
David Stielstra

How do you evaluate what data should be stored locally on-premise and what should move to the cloud? Managed Services Lead David Stielstra gives his advice on considerations for network performance, security and data privacy, and sustainability for each technical environment.

Managed Services

Deciding on where to store and host your data and applications doesn’t have a cut-and-dried, one-size-fits-all answer.

It’s about having a well-defined objective and then picking the right tool for the job. There are advantages to both cloud and on-premise solutions, and there are appropriate times to leverage each when considering your technical environments. On-premise and cloud aren’t mutually exclusive. In fact, it’s possible to have both with a hybrid solution.

For example, if your organization has had a large cap-ex investment in an on-premise data center, you can still leverage the cloud to help supplement local storage or research and development (R&D) efforts. If discussing and evaluating the move to becoming a more cloud-oriented organization, you can start to experiment with applications and slowly migrate as the value is proven out.

So, how do you evaluate what to move to the cloud and what to keep locally? Let’s talk through some areas to consider:

Network Performance

On-premise servers provide fast access speeds by being on a local network. By keeping data on-premise, organizations get a level of control that cloud-based solutions just can’t provide – complete control of the network. IT teams can configure, monitor, and collect network data directly and configure solutions to match needs. If an application or service requires a high-bandwidth throughput, and more users are accessing services in-house, it may be advisable to keep servers local to maintain strong, fast access speeds from a local high-throughput network.

On the flip side, one of the greatest advantages to a cloud environment is that it’s accessible nearly anywhere there is internet access. If more users are accessing the system from an external network, a cloud solution could work in your favor. Most cloud providers have redundant internet providers and allow users to disperse applications over a greater geographical area than most on-premise environments.

Security and Data Privacy

While no single method is 100% effective at preventing data breaches, each has its own advantages and disadvantages. IT teams that desire full control over their data tend to favor on-premise environments. It’s important to note that control doesn’t necessarily equate to secure data. If choosing on-premise, consider your team’s ability to create the infrastructure, maintain it, improve it, and appropriately react to new developments. The unfortunate reality is that even the most reliable and trustworthy team can fall victim to cyber-attacks.

While teams may be skeptical about information being managed by a cloud provider, this approach isn’t any less secure than the traditional approach. Most cloud providers have a shared responsibility model – the provider is responsible for the hardware, software, networking, and physical facilities and the customer is responsible for the security of the applications or services being run in the cloud. Running applications and storing data in a cloud solution may reduce responsibilities of digital security teams while providing better physical security.  

Total Cost of Ownership

CapEx and OpEx

In terms of cost of ownership, it’s important to acknowledge the differences in capital and operation expenses to each solution. On-premise solutions are a higher capital expense upfront, a higher fixed asset cost that depreciates over time due to the lifespan of the equipment. The more money put towards capital expenditures means less free cash flow for the rest of the business, which can hinder shorter-term operations and requires spending prior to development work being completed.

Cloud solutions, though, offer a OpEx approach in the sense that you pay smaller fees in intervals (depending on subscription conditions) to run day-to-day business operations. Unlike the depreciation of CapEx, OpEx purchases have tax benefits in the given year they are made. This also allows organizations to spend less money upfront and have a quicker return on the expenses.

Upfront and Subscription Fees

Beyond the initial hardware investment, there are indirect costs associated with hosting on-premise servers – electricity, networking, racking, space, labor or resource costs for the IT team for support and maintenance of the servers, along with any replacements, repairs, or additions that will come naturally with the lifecycle of an on-premise environment.

Organizations that choose cloud-based solutions do so, in part, to realize the financial benefit and return on investment. Rather than paying a large licensing fee upfront, customers pay a subscription fee on a recurring (monthly or pre-paid package) basis. This SaaS model typically includes maintenance and support, so organizations are spared on some of the additional expenses mentioned above. Organizations are also less likely to over-provision hardware with a cloud service, as you pay only for what you utilize.


Automation and Deployments

For customers that control the deployment process, these might take longer and need more resources from IT teams. While you’ll have direct local access and won’t be limited by hosting capabilities, it’s costly from upfront investments in software before automation can start, and it requires the customer to be responsible for all product updates, platform support, and overall maintenance. With on-premise, teams generally pay more for the automation infrastructure software upfront, taking longer time before seeing the benefits.

For the cloud, ease of deployments is one of its strongest and most attractive advantages. Deployments in the cloud are short-lived, usually only requiring a couple minutes and little effort to do. Automation as a cloud service offers minimal upfront investments in both software and infrastructure, with tons of subscription services offering pay-as-you-go pricing, and even allowing the ability to automate both cloud and on-premise resources. While vendors take on maintenance for customers, the vendor also has control over when that maintenance occurs which could cause chaos among IT teams that may not be ready or expecting it. On the other hand, the vendor is responsible for resolving the problem if something goes wrong.

Backups and Disaster Recovery

A solid on-premise solution for backup and disaster recovery requires that organizations make multiple copies of backups and keep them in physically separate locations. If you keep backups on-premise, and do them correctly, you’ll need at least three times the amount of storage that you plan to use because you’ll have your data and two copies of that data – one copy to keep and one to physically transport to a different location. Some organizations may have similar concerns about the security of their data, as with their backups, so they may appreciate the responsibility of physically protecting their backups. A disaster in an organization’s on-premise data center can be minimized if the organization has a secondary data center or co-location provider.

An attractive advantage to the cloud is that data can be backed up frequently or as often as a customer desires. It’s common practice for organizations to have multiple copies of data in the cloud, storing them in very different geographical locations where third-party cloud providers host their data warehouses, meaning, you’ll almost always have the latest version of information in the case of a disaster that affects one geographic area. While a cloud provider may have 99.999999999% of durability for your data, it still makes sense to have multiple copies in disperse locations. A hybrid solution may be that an organization may choose to store its frequently accessed information on-premise and store backups in the cloud.


It’s critical to consider business growth and the infrastructure’s ability to scale up operations as the need arises. When organizations choose to host their data centers on-premise, there is a high likelihood of them over-provisioning or purchasing for what they currently need to use. No one wants to make a capital investment only to find out that there isn’t enough capacity for the application – either processing or storage. In fact, there’s a chance that they will continuously pay for capacity that they’ll likely never fully utilize. On the other hand, if you have on-premise servers and come into the bind for more capacity, you’ll need the technical talent and capability to introduce additional servers to the infrastructure – along with the time to purchase and get the newly acquired equipment into service.

With a cloud solution, customers only pay for the capacity they use. If an organization has a lower volume of demand one month, the cloud services can be set up to scale capacity back, and if demand suddenly increases, cloud servers can be configured to automatically scale up to allow for additional capacity.

When it comes to choosing between on-premise and cloud environments, there’s no one-size-fits-all answer. It’s best to understand where the organization can and cannot compromise from a data perspective. Based on what we covered, define those hard requirements for your environment and plan ahead of time. Once a plan is in place, do your research and ensure you partner with a solution provider that has experience with both on-premise solutions but is also certified and has experience in cloud solutions.