RE‑ServerDCMetrik
Member of RSVR Invest Group
Contact us
AI-SERVERS

Solutions and projects

We develop tools for GPU infrastructure management, diagnostics and economics. We are also planning ways to use servers while they have spare capacity.

Final stage of developmentAccess by invitation

DCMetrik.com

Server overview and utilisation
Server overview and utilisationActual DCMetrik interface. Fictional demonstration data.

Server monitoring, management and project economics. From individual cost items to revenue forecasts and investment returns.

About the project
Under developmentPrivate

DCMedic — SERA

Server Emergency & Recovery Agent

A tool for diagnosing issues and supporting recovery of AI servers. Incident handling also considers the financial impact.

About the project
Planned

Flexible Compute

A planned scheduling layer that distributes compute jobs across GPU servers with spare capacity.

About the project

The economics of the whole project

We combine hardware expertise with many years of experience in capital markets and process management. We assess both the technical and financial performance of a project.

Acquisition and lifecycle

Hardware price, financing, expansion, replacement and residual value.

Individual operating costs

Power, cooling, space, connectivity, maintenance and management, including incident handling and downtime costs.

Revenue and overall results

Server utilisation, actual and expected revenue, value of internal workloads and returns on invested resources.

Inside DCMetrik

DCMetrik.com

Final stage of development · Access by invitation

Server monitoring, management and project economics. From individual cost items to revenue forecasts and investment returns.

Server overview and utilisation
Server overview and utilisationActual DCMetrik interface. Fictional demonstration data.

Monitoring and detailed accounting

DCMetrik connects server health, utilisation and operating history with project costs and revenue. It covers individual costs for hardware, energy, cooling, connectivity, maintenance and management.

Outlook and returns

Alongside actual results, it provides future revenue estimates and ROA and ROI calculations. It supports assessment of the whole project and individual decisions about repairs, renewal or expansion. Access is by invitation.

Access, customers and server history

The Advanced Account Admin oversees multiple accounts. Each Account Admin manages their entire account and assigns servers to users, who see only their allocated machines. They can set different commissions and costs and reassign servers between users while preserving revenue and cost history.

Heartbeat and alerts

Regular reports help track the availability and health of connected servers. If reporting stops or a problem is detected, users receive alerts according to configured rules and permissions.

Request an invitationFull project page

DCMedic — SERA

Under development · Private

A tool for diagnosing issues and supporting recovery of AI servers. Incident handling also considers the financial impact.

Server Emergency & Recovery Agent

SERA stands for Server Emergency & Recovery Agent, a tool for handling operational issues and supporting server recovery. It is part of the DCMedic project.

It helps administrators understand what is happening on a server, assess available diagnostic information and plan the next steps. Its purpose is to support incident resolution and the return to operation.

It is an internal tool under development. Partner access is available on request by arrangement with the operator.

DCMedic — SERA on DCMetrik

Technical faults have a financial impact

Development takes account of downtime, unused capacity, repair costs and the impact on revenue. Integration with DCMetrik is intended to put individual interventions in the context of the whole project’s economics.

Contact us about this projectFull project page

Flexible Compute

Planned

A planned scheduling layer that distributes compute jobs across GPU servers with spare capacity.

A wrapper for spare GPU capacity

The planned control layer will identify available resources and distribute suitable jobs across servers. Customers’ primary workloads retain priority. The focus is on jobs that can wait or support interruption and resumption.

Use hardware that already exists

Better use of spare capacity can reduce the need to acquire additional servers. We will assess completed work, energy consumption, operating costs and revenue. Public job submission is not yet available.

Contact us about this projectFull project page