“Why are you still running on-prem instead of in the cloud?”
It’s an easy enough question to ask, but you’ll get an infinite number of answers, including ones that disparage cloud computing as a whole. To a lot of people looking in from the outside, sometimes it might seem like the cloud isn’t all it’s cracked up to be. Alternatively, some are too set in their ways to consider migrating away from on-prem servers until they’re dragged kicking and screaming.
I’ll lay it all out as plainly and simply as I can, so even the most doubtful can see the benefits.
Cost is key
One of the key components to running a business is cost. It might not be the only component, but it’s definitely high up there on a list of priorities. So even if we don’t like talking about it, we have to consider it. When faced with two ways of doing something, both of which are legal and ethical and yield the same end result, a hundred percent of business owners are going to choose the one that is most cost effective.
On that note, let’s consider some of the statistics quoted in a 2009 IDC report by Dana Wiklund and Joseph C. Pucciarelli, “Improving IT Project Outcomes by Systematically Managing and Hedging Risk.” The report states that 25% of IT projects fail outright. Meanwhile, an additional 20-25% don’t provide return on investment (ROI), and up to half require material rework.
Planning out these cloud migrations is crucial. You don’t want to be yet another statistic and have your project fall into any of the three categories. So, let’s take a look at your options:
Falling into the category of high-risk when it comes to failure or re-architecture, we have on-prem solutions, specifically:
- load balancers
- enterprise software
All of these are capex. Wouldn’t you rather redirect those dollars somewhere else?
What solves this problem?
Infrastructure as code (essentially, cloud). It lets you fail cheaply instead of investing lots of time and money in something that might not pay off. With it, you can bring up servers, storage, firewalls, databases, middleware, apps, etc. with either a few commands or automation, if you orchestrate it right.
If your architecture fails, you have at your fingertips the ability to bring new services up or down with a couple of commands. Often it’s not an ‘if,’ but a ‘when.’ The cloud lets you fail faster and cheaper than you would otherwise, which in turn lets you arrive at possible solutions more quickly.
It’ll also scale as needed, adjusting to new workloads as they come, which to me personally, is alone enough of a reason to get on board.
Universal cloud credits mean it’s all available instantly, with no more wait times like before.
And lastly, in the fast-paced modern world, do you really want to continue to run things on-prem? Think about big data, blockchain, next-gen machine learning…running these anywhere outside of the cloud would be a nightmare.
Bare metal in the cloud
We already have articles written about preliminary things to consider when moving to the cloud and the networking aspect, so let’s skip ahead to bare metal cloud options with Oracle, and why they’re a solid investment.
Bare metal offers you industry-leading performance with pay-as-you-go pricing, but more importantly, the security of a bare metal machine is unmatched. Unlike every other provider, Oracle doesn’t run code on their machines or provision it with a hypervisor. Instead, they provision via the SDN, which gives them no access to your data. Once you buy the machine, you can bring your own image. They’re also unable to do any knowledge filtering based on your I/O type. This is crucial if you’re in an industry with competitors. Because you own the machine, you have full control of it, and there’s no way to access your metadata from the outside.
One of the most common examples of this happening is with Spotify. They built their music streaming service on AWS, and a few years later, Amazon came out with a similar service. The Amazon service was superior to Spotify, because Amazon spent those few years building their product and analyzing metadata collected from Spotify.
Okay, so why run it in the cloud, though? Can’t I run the bare metal machine on-prem?
You can, but remember the scalability? When you’re running in the cloud, there’s no peak workload. During the time that your business is busiest, you can expand your environment to handle it, which saves a fortune. The only time this wouldn’t make sense is if you’re running 24/7 peak workload, which most businesses aren’t. Using orchestration, you can detect when CPU usage increases, and allow the environment to scale as needed during busy times. When workload decreases, it automatically collapses back down to save costs.
Ready to try out bare metal?
If I haven’t convinced you yet, email Infolob’s senior director Glen Shok at [email protected] and pick his brain for some more answers.
This article was compiled by Carson Collins, Infolob’s resident writer, from Glen Shok’s webinar presentation, “How to best implement cloud and hybrid cloud for maximum efficiency and real cost savings.” You can reach Carson at [email protected]