Infrastructure Consulting
Servers, storage, interconnect — and the software above them. We measure what is there and design what is needed. Vendor-neutral, without preferences: the requirement picks the building blocks.
Services.
This is what we deliver.
Network, servers, storage, HPC. Databases, migrations, job chains.
We know both halves — the hardware and the software above it — and above all we know where they slow each other down. That is what decides whether an installation carries the load or merely runs.
We do not build a standard solution around your requirement. We build the solution that fits it.
We do IT. Enterprise.
Project management, architecture and delivery do not sit in three different companies here. That spares you the handoff where responsibility tends to disappear — and the argument about whose fault it is when things do not run.
It works the other way around too: on a large project currently under way we field five people in distributed roles, the architect among them, while control and responsibility stay with the customer. We do that when that is what the customer wants. We prefer the version where we own the outcome as well.
What we love
Project Management
Scope, dates, budget — and the technical conversation with your people. Handoff points are where projects fail. With us they are in the plan from the start.
Technical Specialists
Installation and cabling. System and network administration. Scripting and development. The people who actually get the thing running — and then show you how to run it.
Where we come from
From the data center. Where you can put your hand on the metal.
Where you can hear data and feel computing as warm air. Where you find machines, not computers — tape robots the size of three sedans. Where twenty cables hang off the back of a server and two of them are power.
A computer network is not a power grid. Power is not virtualized. Power does not run over glass and copper at the same time, knows neither single-mode nor multi-mode, and you do not just splice in another branch. Treat the two alike and you build a data center that computes — but does not carry the load.
Vendor-neutral, but shaped by where we come from: most of us come out of the UNIX world.
More precisely
HPC, named honestly. We build HPC setups: InfiniBand, NVMe over Fabrics, load balancing. Traditional parallel computing for weather models or simulations is not something we have done. So we call it mini-HPC. Sixty 2U servers for the application alone, plus the fabrics and the databases — there is nothing mini about that anymore.
Workflow is not workload. We know the difference. It sounds like splitting hairs until a job chain gets stuck on exactly that confusion.
What we do not do
No operations, no managed service. Your infrastructure stays yours. We build, hand over, leave.
No SAP. Interfaces to it, yes — from a job chain, for instance. The product itself is something others do better.
No Microsoft, no desktop, no VDI. We are the back end. Where the client plugs its network cable into the wall, our side begins — the other one, the one nobody knows and nobody sees. And no Microsoft there either: others do that better. We come out of the UNIX world.
No public cloud. Not so far.