Finding a Safe Place for Your Data and Software
Your organization runs on data and software. But this whole IT environment needs to live somewhere. Preferably a safe place that no unwanted people can access.
What options do you have? How should you choose where to host your data and your software?
In this article, we’ll explore these topics in-depth, hopefully giving you that bit of additional information that you need to choose a safe place for your IT environment.
Where can you host your software/data?
The traditional way is to host it on your own servers, which is called on-premise hosting.
It’s private by nature because the whole infrastructure is dedicated only to your company. The software literally lives on your own machines, along with the data and all of your intellectual property. Servers don’t need to actually be located at your headquarters, they’ll probably be in a dedicated data center.
The “new” (it’s not that new and pretty much standard by now) way to manage your IT resources is cloud hosting.
It’s public by default because it’s provided by a company like Amazon or Microsoft, whose insane server power is shared by all of their customers. But it can be private because cloud providers offer the option to get a share of their servers dedicated only to your company.
Finally, you can also mix the different options, and then you get hybrid hosting. There are a lot of ways to organize a hybrid solution, with different combinations of hardware and software. Choosing one cloud provider doesn’t mean you can only use that one, you can also combine different services from multiple providers.
How much control do you need?
When it comes to hosting your software and data, available server options generally fall into these categories:
- Control the hardware, control the software
- Control the hardware, outsource the software
- Outsource the hardware, control the software
- Outsource the hardware, outsource the software
Control the hardware and software
If you need to control and customize the performance of your physical servers, as well as the software that runs them, the go-to choice is on-premise hosting.
Control the hardware, outsource the software
What if you need to control the hardware, but you want the same workload management experience that’s offered by big cloud providers? There are ways to run, for example, AWS services on your own on-premise servers. The offerings in this area vary based on the provider.
Outsource the hardware, control the software
Your server workloads are pretty typical, you don’t need custom hardware for your IT environment – but you want to use, for example, FileCloud to share and manage your organization’s data. You can easily run FileCloud on AWS, as well as other services that you might need.
Outsource the hardware and software
This is probably the most popular solution at the moment for non-enterprise companies. You just spin up a server instance at your favorite cloud provider and manage it using the software tools they provide. Use it to host your data, your ERP system, or your SaaS, without worrying about the server infrastructure.
Comparing hosting options – On-Prem vs Cloud vs Hybrid
So far we know that on-premise hosting is private (dedicated only to your company), with your IT environment living on your own physical servers.
But when should you use on-premise hosting? Modern tech companies usually start with the cloud, and move on to on-prem.
Take the case of Instagram, they migrated to Facebook’s infrastructure after FB bought them in 2012.
(but then they also branched out to different data centers around the world to ensure that all of their users have a good experience, so they’re definitely not on-prem only)
Companies and enterprises that have been around for decades tend to go from on-prem to adding a bit of cloud, or migrating to the cloud completely.
Like when AdvancedMD moved to the cloud. AdvancedMD is a healthcare-related provider of digital services that’s been around since 1999, which makes this a great example. The most common argument for on-premise hosting is that it’s the most secure option for highly sensitive data. AdvancedMD runs on healthcare data, which is extremely sensitive, and yet nothing tragic happened when they migrated to the cloud.
As AdvancedMD proves, the issue of security is not that important anymore. Both on-premise and cloud hosting can safely store sensitive data.
So the choice between on-prem and cloud is more about control and/or customization.
For the highest amount of control, and the ability to literally customize every part of your infrastructure, on-prem is the right option. Long-term cost management is easier, however, it takes a large initial cost to build your on-prem hosting from the ground up.
On-prem is also a good option when you have high demands:
- You’re constantly moving large amounts of data in and out of your servers (cloud providers can charge fees for moving data outside of your cloud),
- You need the lowest latency possible.
One problem with on-prem is that it’s harder to scale, but you can use a cloud provider to mitigate this issue.
You’ve probably heard this, but – there is no cloud, it’s always somebody’s server. It’s a popular saying, but it carries a hidden warning about your data being on somebody else’s server.
How big is the risk that cloud providers will mismanage your data, or give someone else access to it? Unless you’re handing out access credentials to your cloud to everyone you meet, the risk is actually very small.
There is no way cloud would’ve become the new standard for hosting if it were risky. Providers know this, and they’ve put extreme amounts of money into making sure that your resources are safe with them.
Another popular issue that people bring up when talking about the cloud is compliance with standards. But it turns out that cloud providers are surprisingly compliant with cross-industry IT standards, so this issue depends on your unique case.
There is a different, much more real, risk associated with the cloud – cost management.
Sure, at the start you pay much less compared to an on-premise solution. As you keep going, it’s super easy to spin up new services from a cloud provider, especially if you have a huge IT budget.
This is a benefit because you can scale up extremely easily. It’s also a problem because you might end up paying for a lot of unnecessary services.
So if you don’t want to overspend, you need to be very careful about managing your cloud infrastructure.
Choosing cloud isn’t a problem of compliance nor security, but rather a problem of your unique workloads. As we learned above, on-premise can be better when you need to move huge amounts of data regularly, or you need minimal latency.
For example, if your servers are just supposed to do the standard job of serving a website to people online, the cloud is the logical solution. But if you’re building a complex web application that performs difficult computations on large amounts of data, you’ll probably be better off with an on-prem, or a hybrid solution.
And so we arrive at the most common option, hybrid hosting.
The complex demands of enterprise IT environments make it almost impossible to just pick one hosting option and roll with it for eternity.
There are too many considerations:
- Integrating with legacy software,
- Speed vs reliability,
- Location of data,
… and so on, and different parts of a typical IT environment require varying approaches. For example, a cloud provider might work for your in-house data store, but you still need on-prem servers to run particular applications or legacy software.
Hybrid hosting is a way to address all of this complexity because you can combine multiple options to create the infrastructure that meets your requirements to the letter.
All in all, there is no silver bullet when it comes to hosting your data and software. The safest place for your IT environment might be at a cloud provider, or on your own on-premise servers. Or both.
It depends on what you need, and it turns out that security and compliance are not the biggest issues when you’re thinking about migrating to the cloud. It’s more about the type of data workloads that you have, and the requirements that result from this.
Hope this article was helpful, thank you for reading!