Developer Community for OpenStack & OpenInfra at the Edge
Run OpenStack‑native, Kubernetes‑first workloads on a decentralized edge cloud. Flat‑price nodes. Zero egress fees. Local data. Built for YOLO/OpenCV, SLURM HPC, Docker, GitOps, and microservices.
Why Open Source Devs Choose Cachengo®?
What the OpenStack/OpenInfra crowd values—and how we deliver.
Upstream‑friendly
OpenStack & OpenInfra alignment, standard APIs, CNCF‑native tooling, and OSS licenses where it counts.
OpenStack & OpenInfra alignment, standard APIs, CNCF‑native tooling, and OSS licenses where it counts.
Observability
OpenTelemetry, Prometheus/Grafana, Loki/Tempo—export locally or upstream on your terms.
OpenTelemetry, Prometheus/Grafana, Loki/Tempo—export locally or upstream on your terms.
Security at the edge
SBOMs, image signing, secret stores, WireGuard/IPSec site mesh, least‑privilege defaults.
SBOMs, image signing, secret stores, WireGuard/IPSec site mesh, least‑privilege defaults.
GitOps + IaC
Flux/Argo, Terraform, Ansible. Reproducible edge clusters with policy‑as‑code.
Flux/Argo, Terraform, Ansible. Reproducible edge clusters with policy‑as‑code.
Real data gravity
Compute near cameras, PLCs, and users—less WAN, lower latency, and fewer retries.
Compute near cameras, PLCs, and users—less WAN, lower latency, and fewer retries.
Predictable economics
Rent-a-Node™ → flat $/node/month. No egress tolls. Model‑agnostic deployment.
Rent-a-Node™ → flat $/node/month. No egress tolls. Model‑agnostic deployment.
Stacks & Integrations
OpenStack at the Edge
- Nova/Cinder/Neutron where appropriate; lightweight profiles for constrained sites.
- Keystone‑compatible identity and multi‑tenant quotas mapped to nodes.
- Glance & object storage local to site to avoid WAN drag.
Kubernetes‑First
- Managed K8s on each node group, CRI‑O/containerd + CNI of your choice.
- Helm/OCI registries local‑first; image signing + SBOM attestation.
- GitOps pipelines with canary/blue‑green, drift detection, and policy gates.
Containers & HPC
- Docker/Podman for dev inner loop; SLURM for batch/HPC orchestration at the edge.
- Multi‑queue scheduling across ARM CPUs and NPUs; optional GPUs/Gaudi tiers.
- Data locality hints for schedulers to minimize backhaul.
Networking & Security
- Site‑to‑site mesh with WireGuard; policy routing by workload or namespace.
- mTLS service mesh option; secret stores, KMS integration, Vault compatible.
- Remote attestation and node health proofs.
Workloads We Love
YOLO / OpenCV
Per‑camera object detection at the site. INT8/INT4 paths on RK3588 NPUs. Optional GPU tiers.
Per‑camera object detection at the site. INT8/INT4 paths on RK3588 NPUs. Optional GPU tiers.
Doc AI / RAG
Local vector DB + chunking so prompts stay tiny; keep PII/IP on‑site.
Local vector DB + chunking so prompts stay tiny; keep PII/IP on‑site.
Agents
Tool‑calling on device (OCR, code exec, search) to cut token loops and latency.
Tool‑calling on device (OCR, code exec, search) to cut token loops and latency.
Microservices
Telemetry‑heavy apps that need sub‑regional latency and predictable $/month.
Telemetry‑heavy apps that need sub‑regional latency and predictable $/month.
HPC / SLURM
Batch/array jobs near data producers with site‑level backoff and retries.
Batch/array jobs near data producers with site‑level backoff and retries.
Streaming Analytics
Kafka‑lite patterns with local retention; WAN only for rollups.
Kafka‑lite patterns with local retention; WAN only for rollups.
Estimate per-site monthly cost for camera analytics. Token pricing doesn’t apply; clouds charge per-instance/hour + egress. Cachengo processes on-site with flat $/node and typically zero egress.
Preset (resolution/FPS)
Cameras
Hours / month
Cloud GPU $/hr
Streams per cloud instance
[fullhd]*2
Streams per Cachengo® node
[fullhd]*3
Cloud egress GB/camera/mo
Cloud egress $/GB
Their Price:
([cameras]/[instance]*[gpu]*[hours])+([cameras]*[egressgb]*[egressdollar])
$
Cachengo (Flat Node) Price:
([cameras]/[node]*[gpu]*[hours])+([cameras]*[egressgb]*[egressdollar])
$
Positioning: The Decentralized Edge Cloud for OpenStack & OpenInfra
Why Cachengo?
- Decentralized placement → put compute where data is born.
- Economic predictability → flat node pricing, zero egress fees.
- Open by default → align with OpenInfra projects; no lock‑in.
How It Fits
- Run OpenStack services tailored for edge footprints; integrate with central control planes.
- Kubernetes as the substrate for app delivery; SLURM for HPC lanes.
- Operators list capacity; tenants pick placement via marketplace.
Community & Programs
Developer Channels
- GitHub org with examples: YOLO on RK3588 NPU, SLURM‑on‑edge templates, Helm charts.
- Discussions: OpenStack edge profiles, CSI/CNI choices, policy‑as‑code recipes.
- Monthly office hours: performance tuning, placement, and cost modeling.
Operator Program
- Publish nodes to the marketplace, enforce SLOs, share revenue with tenants.
- Validated images, signing, and attestation; health telemetry tooling.
- Edge‑site runbooks and incident playbooks.
Join the Developer Community
Tell us about your stack and we’ll size nodes, recommend placement, and invite you to our developer channels.
While traditional distributed public cloud architectures have provided the scalability that companies need, they are limited in three key areas – security, data availability/resiliency, and operational complexity.
Cahengo’s decentralized Rent-a-Node™ cloud infrastructure offers a radical improvement over traditional models. Powered by Symbiote® bare-metal servers, this unique approach equips companies with the tools to own their data, reduce risk and operational expenses, and scale efficiently and with greater autonomy.
| Strategic Advantage | Decentralized Benefit |
|---|---|
|
Data Sovereignty |
Locally stored and processed data aligns with corporate security goals and minimizes breach risks. |
|
Fault Tolerance |
No single point of failure—individual Symbiote® bare metal servers can fail without disruption to data accessibility.
|
|
Edge-Optimized Efficiency & Scalability |
Real-time processing near the data source reduces latency and bandwidth consumption, with the ability to easily scale to thousands of Symbiote® bare metal servers. |
|
Deploy Cloud Resources Quickly and Minimize OPEX |
Deploy new compute and storage resources in existing real-estate resources, with industry-best power, cooling, and space efficiency. |
For further questions regarding the decentralized cloud, please contact us.
Solutions
Products
Company