AnaROS turns a rack full of heterogeneous CPUs, GPUs, fabric, and storage into a single governed system — and extends the same operating contract across racks, clouds, and providers as the pipeline crosses them. Four lenses, one storyline — Visibility, Traceability, Governance, and Workload Placement — built on multi-tier telemetry that goes from L4 application workflow down to L1 fabric, and follows the pipeline wherever it lands.
A single AI call traverses your application code → runtime → GPU → fabric → storage in milliseconds. When p99 collapses, the symptom you see at the top — a slow agent, a stuck retrieval, a timeout — usually has its root cause four layers down: a NIC buffer overflow, a NUMA miss, a noisy neighbor in another tenant. The dashboards you have today don't connect: K8s knows the pods, CUDA knows the kernels, the network team owns the spine — none of them see across boundaries.
Four cooperating layers — L4 workflow, L3 pipeline X-ray, L2 logical fabric, L1 physical rack — visible from one operator surface. Trace any workflow top-to-bottom in a single pane, and the symptom and the root cause finally show up on the same screen. Scroll to drill into each layer.
Agentic flows, MoE applications, inspection pipelines — expressed as a controlled workflow graph with steps, branches, loops, and SLOs. This is the surface CIOs and product owners reason about.
AnaROS decomposes each workflow into Ingest, Preparation, Execution, and Persistence across network, compute, memory and storage — and routes each stage to the right fit GPU class that holds its SLO. Heavy CNN/ViT inference doesn't mean every stage needs an H200.
The Anavec logical fabric resolves source/destination, bandwidth, and zero-copy paths between GPUs, the CPU sled, the staging tier, and storage. The resource map tells you what's allocated, what's draining, and what's free — at all times.
Standard rack envelope. Off-the-shelf modules. Programmable PCIe fabric. CPU and GPU refresh on independent clocks. The POFC tree resolves every pipeline edge to a physical switch port.
This workflow lives on anavec1(GPU) for ingress and preparation, hops across fabric to anavec2(GPU) for GPU execution, then hops back to anavec1 for storage. Pipeline X-Ray sees all of it — intra-node IPC, inter-node fabric, leaf switches, every stage.
Ingress and Preparation both live on anavec1 — but they're separate processes in separate containers, talking through a kernel pipe. Most observability stacks lose visibility here. The X-Ray traces it: pipe-buffer occupancy, context switches, container CPU contention.
Preparation on anavec1 hands off to Execution on anavec2 — across a real fabric edge. Without POFC, this hop goes dark. With POFC, the X-Ray pins it to a specific switch port pair on leaf1-0 and surfaces drift, loss, and ECMP right there.
Every pipeline edge reconciled with its physical path through leaf/spine. ECMP divergence becomes legible. Ask "where does this workflow live on this fabric?" — and get a topology answer, not a Grafana stack.
Six telemetry domains (Node, Kernel, Process, Net L2/L3, TCP, HTTP/L7, gRPC/H2) crossed with six collector layers (SDI/SSH, Prometheus, cAdvisor, eBPF/Beyla, SDK Py/C++, OTel/OTLP). Each phase adds signals — and unlocks new RateGate capabilities.
Open the switch and you see POFC signals: resolved endpoints, reconstructions, drifts, violations, latest tick. Verdicts and pending mitigations live next to the switch they apply to, not in a separate ticketing system.
In this example, GPU SM 1% looks like a GPU problem, but the X-Ray correlates intra-node, inter-node, and fabric signals across all stages — and lands on Persistence (suspicion 98%). Storage IOPS contention back-pressures execution upstream all the way to ingress.
| PIPELINE | SRC → DST | HOPS | CLASS |
|---|---|---|---|
| embedding-amsf-pipeline | anavec3 → anavec2 | leaf2-0 | scale-out |
| embedding-baseline | anavec3 → anavec2 | leaf2-0 | scale-out |
| langgraph-semantic-moe | anavec3 → anavec1 | leaf2-0 → spine1-0 → leaf1-0 | scale-up |
| langgraph-semantic-moe | anavec1 → anavec2 | leaf1-0 → spine1-0 → leaf2-0 | scale-up |
| b10-resnet-classify | anavec1 → anavec2 | leaf1-0 → spine2-0 → leaf2-0 | scale-up |
| PIPELINE | EDGE | PORTS | BOUND |
|---|---|---|---|
| langgraph-semantic-moe | preparation → execution | leaf1-0:Eth80 → leaf2-0:Eth80 | ✓ bound |
| langgraph-semantic-moe | post_proc → persistence | leaf2-0:Eth88 → leaf1-0:Eth80 | ✓ bound |
Visibility → Traceability → Governance → Placement, all on the same governed pipeline. The same AnaROS console runs on your homelab today.
