Infrastructure migrations

At Soverity we have extensive experience with bare metal, European compute vendors, and the hyperscaler clouds. We’ve operated hybrid cloud/bare metal setups and multi-vendor fleets. We know their strengths and weaknesses. We can help you optimize your infrastructure to increase performance, lower cost, and reduce risk.

We are engineers. What we deliver is not a 150-page report detached from reality, but full implementation. We write code, we configure systems, and we measure. We are able to give good advice, because we are the people who implement. We know how infrastructure works in the real world, because we’ve been operating infrastructure for many years. We’ve seen what works and what doesn’t. For example, a highly available distributed system can mitigate the risk of hardware failure, but a system that is complex to configure also increases the risk of operator error. We know how to navigate this subtle trade-off, because we’ve operated software on both sides of this spectrum. Our advice is grounded in decades of real-world experience.

Cloud to bare metal

While the hyperscaler clouds are impressive feats of engineering that solve real problems, they are not the right solution for every problem. Engineering is about trade-offs. At scale, cloud can be an order of magnitude more expensive than bare metal, especially with metered network egress, a pricing model unique to cloud. When it comes to demanding workloads, a bare metal machine with the latest generation CPU and high-bandwidth low-latency NVMe storage runs circles around the cloud’s vCPUs and virtualized block devices. Bare metal is not without its own challenges though: unlike virtual devices in the cloud, physical hardware can and will fail. It is possible to build a reliable system out of fallible components, but this does require designing your infrastructure around it. We’ve been operating infrastructure with tight security and uptime requirements on bare metal for years; we can help you do the same.

Foreign to European

Depending on a foreign vendor for critical infrastructure, especially one that is subject to unpredictable legal requirements, is a serious risk. By now it is clear that data residency requirements alone don’t protect your data when the systems that manage that data are subject to foreign jurisdiction. The US CLOUD Act is invasive. American hyperscalers, once seen as the safe default choice, are now becoming a supply chain liability. We have extensive experience with European vendors, with their strenghts and weaknesses, and we know how to fill the gaps in their offerings. We can help you choose and migrate to sovereign vendors.

Hybrid setups

Production systems are not green-field projects that have the luxury of building on new infrastructure from the start. Real-world systems have legacy requirements. Big-bang migrations that move everything at once never end well for large systems. Even when you have a clear picture of where to move, the reality is that you’ll be running in multiple places for a long time. How to move is in itself a hard problem. Within a single vendor’s network, machines can easily reach each other privately over VPCs and VLANs. In a multi-vendor setup, traffic will need to flow over the public Internet, which requires setting up encrypted tunnels, and designing with egress fees in mind. We have operated production systems with strict uptime and security requirements across multi-vendor and hybrid cloud/bare metal setups. When we work with you, we don’t just look at the end goal in isolation, we also help you get there. We design and help you execute the entire migration process, because often the migration itself is the the most difficult part.