DF-R4Shipping

Composable Rack.

Turn a fixed rack of servers into a fluid pool of GPUs, memory, NVMe, and accelerators — claimed by any host on demand. Utilisation goes up. Refresh cycles move to the endpoint. Vendor lock-in evaporates.

AT A GLANCE

Composable Rack — what you get.

A single Datafabrix DF-R4 rack, delivered as one purchase order, installed as one unit, supported as one system.

Mission
PCIe-any-to-any composability for enterprise fleets
Form factor
42U · full rack · 20 kW nominal · N+1 PSU
Workload profile
Composable enterprise infrastructure · Multi-workload consolidation · Test-and-dev pools · GPU-as-a-service · Accelerator sharing
Target customer
Enterprises modernising to composable, financial services HPC, private-cloud operators, GPU-sharing platform teams
KPIS

The numbers that matter.

The specific metrics our customers ask about first — and where this rack is engineered to win.

Endpoint types supported
GPU · NVMe · DPU · FPGA · accelerator
Host-to-fabric ratio
1:1 up to 4:1 configurable
Composition change time
Seconds, live — no host reboot
Vendor lock-in
Zero · open PCIe endpoints
BILL OF MATERIALS

Every layer of the rack, itemised.

Merchant silicon is called out explicitly. Datafabrix engineering is called out explicitly. No hand-wave categories.

Layer Component Notes
Host servers 6 × host servers with Datafabrix PCIe HBAs Existing hosts or Datafabrix DF-X Xeon Server Cards
Fabric 2 × Datafabrix PCIe Gen4 Fabric Backplane (DF-B4) Redundant PCIe switching for HA
Composable EP GPU shelf · NVMe pool · accelerator shelf Endpoints that any host can claim through the control plane
Control plane Management Server (DF-S13) Zoning, policy, RBAC, fleet-wide composition
JBOF JBOF Platform (DF-S03) · 24 × NVMe Disaggregated flash pool visible to all hosts
DCIM All 8 DCIM modules pre-instrumented Includes composition telemetry: which host owns which endpoint, right now
Networking Standard datacenter networking (customer choice) Rack does not sit in the data-network path
Cooling Air-cooled · standard rack thermal envelope Standard datacenter cooling profile
WHY THIS RACK

The problems this rack addresses.

GPU stranding is a $M problem

A server with a $30K GPU that runs at 20% utilisation is $24K of stranded capital. This rack lets any host claim any GPU on demand — utilisation moves from 20% to 60–80% overnight.

Refresh at endpoint level, not server

Traditional servers force you to refresh compute + I/O + accelerators together. Composable lets you refresh only the endpoints — a new GPU generation slides in, old NVMe stays, hosts unchanged.

Test-and-dev pools become elastic

One rack replaces a dozen separate departmental dev servers, each half-idle. Every user gets the resources they need, when they need them, and gives them back automatically.

Open silicon means no vendor lock

Because the fabric is merchant PCIe and the endpoints are standard PCIe cards, you are never locked into one accelerator vendor. Migrate to next-generation silicon on your timeline, not theirs.

WHERE IT'S DEPLOYED

Four workloads. One rack.

Enterprise GPU-as-a-service

Internal GPU pool shared across data science, ML engineering, and product teams.

Financial services HPC

Elastic risk-analytics infrastructure — burstable to peak load, contracted at quiet times.

Multi-workload consolidation

Retire a dozen legacy servers into a single composable rack with better utilisation and lower TCO.

Accelerator evaluation lab

Swap between GPU, DPU, and FPGA endpoints under the same hosts and workloads to benchmark objectively.

FROM ORDER TO POWER-ON

From order to power-on. In weeks.

Our delivery discipline is the single biggest reason customers pick us over DIY assembly or a locked AI system. Here is what the timeline actually looks like.

01

Discovery & quote

One call to align on workload, site, and timeline. Written quote with BOM, power/cooling profile, and delivery date inside one business week.

02

Build & validate

Rack assembled and validated at our integration facility. Firmware loaded. DCIM modules provisioned. Burn-in tested against the specific workload class.

03

Ship & install

Rack ships fully assembled. On-site installation by our team or a certified partner. Cabled, powered, network-integrated, DCIM streaming to your operators.

04

Power-on & handover

Workload deployed on the rack. Acceptance tests. Handover to your operations team, with our engineering on standby for the first 90 days.

WHAT'S INSIDE THIS RACK

The solutions inside this rack.

Each Datafabrix rack is a curated combination of solutions. Below are the solutions that ship inside this rack. Click any card for the block diagram, connection topology, and per-subsystem specs.

READY FOR YOUR FIRST DF-R4

Let's design your first composable rack.

Tell us your workload, your site, and your timeline. We will respond within one business day.