The Best of Both Worlds: Bare Metal’s Role in Fintech
- by Ned Pope
Key takeaways
- Many fintech platform teams find themselves overpaying for burst-capable cloud resources that support workloads that don’t actually require elasticity.
- Hivelocity’s bare-metal infrastructure as a service (IaaS) solution provides dedicated hardware that’s ideal for steady-state workloads while enabling the cloudlike experience today’s fintech teams rely on.
- Many teams today are opting for a hybrid strategy where they optimize workload placement across cloud and bare-metal environments to maximize efficiency and performance.
Many fintech companies examine their cloud bill and discover that most of it was consumed by their CI farms. Why? Because they’re paying premium on-demand rates for workloads that don’t need to burst. And with so much time and effort invested in workflow automation, they’re hesitant to move them.
The teams know a workload is a workload and that, in theory, they can run them elsewhere, but investment in infrastructure as code (IaC) and API-driven provisioning has resulted in the fast, secure deployments that the business demands. Moving the workloads, they assume, means removing these types of capabilities, which will have a huge productivity impact on engineers.
The Bare Metal Alternative
In truth, a bare metal IaaS provider can offer the same experience as the cloud—including IaC and API-enabled provisioning capabilities—but only if the vendor is built for it and exposes essential capabilities to automate infrastructure. Having bare metal exposed through an IaC provider and a REST API with deployments happening on the same timescale as cloud provisioning means getting the best of both worlds. You keep your workflow and optimize your cloud spend.
Hivelocity delivers the cloud-like bare-metal experience your team needs to make it happen. Working with us, your team can quickly provision dedicated servers through an IaC provider and an OpenAPI-driven REST interface, or through the myVelocity portal. Existing IaC modules in pipelines and runbooks keep working. From the engineer’s perspective, the deployment experience still looks like cloud. The underlying hardware is dedicated and single-tenant, with pricing based on steady-state demand rather than around the possibility of extreme peaks.
The workflow stays the same, but the cost base changes.
The Best Home for Each Workload
Predictable workloads work best on dedicated hardware. CI farms, batch jobs, ledger services, and settlement engines are great examples of workloads without much need for burst. Capacity needs for these use cases vary, but not drastically. Additional hosts can be provisioned when needed.
Workloads that see high variability in demand are best suited for the cloud.
These could be things such as marketing events, high-processing ad-hoc analytics, or anything else that benefits from on-demand scale. In the cloud, these workloads can access additional resources as needed to keep everything running smoothly.
The Hybrid Approach
Many teams find that a hybrid strategy that uses both bare metal and cloud is best for them. IaC makes it possible in a highly automated way: in-cloud or out-of-cloud, with dedicated bare-metal hardware for the steady state and virtualized cloud for genuine bursts. The IaC layer treats both dedicated and cloud targets as provisioning resources.
Of course, bare metal doesn’t provide infinite burst capability. But in most cases, the burst is semi-predictable and can be supported via the server’s spare capacity. Platform engineering teams should size dedicated bare-metal capacity against expected steady-state load with deliberate headroom for small spikes and, if needed, the ability to route extreme spikes back to the cloud.
All of this makes capacity planning an even more critical function. And if you’re pursuing a hybrid, IaC-enabled workload strategy, it’s the only thing your team truly needs to do differently. The cloud-like capabilities remain, and the business stops paying hyperscale economics for workloads that never needed hyperscale characteristics in the first place.
Contact Hivelocity today to get started with your fintech workload placement optimization.
FAQ
Q: What fintech workloads are the best candidates for bare-metal infrastructure?
A: Dedicated bare-metal infrastructure is best for workloads with predictable demand and stable resource requirements. Common fintech examples of steady-state workloads include CI/CD farms, batch processing systems, ledger services, settlement engines, and core banking platforms.
Q: Can platform engineering teams keep using IaC on Hivelocity?
A: Yes—platform teams keep the experience they’re used to with Hivelocity. Our IaC providers, APIs, and automated provisioning workflows allow them to continue using IaC, CI/CD pipelines, runbooks, and automation.
Q: Does my organization need to take everything out of the cloud?
A: No, because cloud is a great home for many critical fintech workloads. Dedicated bare-metal infrastructure is ideal for steady-state workloads. But cloud is best used to support variable demand, traffic spikes, or burst requirements that single-server resources cannot meet.
