Skip to content

Krishna-kg732/cloudPlane

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

cloudplane

(currently being scaffolded)

A multi-cloud control plane that gives you access to AWS, GCP, and Azure infrastructure from a single interface — while keeping everything in your own accounts.

MVP: Currently AWS-only. GCP and Azure coming in future phases.

Tech Stack

Go Kubernetes Terraform AWS


Architecture

┌─────────────────────────────────────────────────────────────────┐
│                         User / Web UI                           │
└─────────────────────────────────────────────────────────────────┘
                               │
                               ▼
┌─────────────────────────────────────────────────────────────────┐
│                      Control Plane API                          │
│              (Projects, Connections, Routing)                   │
└─────────────────────────────────────────────────────────────────┘
         │                    │                    │
         ▼                    ▼                    ▼
┌─────────────┐     ┌─────────────────┐    ┌─────────────────┐
│  Credential │     │    Training     │    │   Inference     │
│   Broker    │     │    Service      │    │    Service      │
│  (OIDC→STS) │     │ (Kubeflow Jobs) │    │  (vLLM, TGI)    │
└─────────────┘     └─────────────────┘    └─────────────────┘
                               │                    │
                               ▼                    ▼
                    ┌─────────────────────────────────────┐
                    │            Orchestrator             │
                    │    (Terraform + K8s Operations)     │
                    └─────────────────────────────────────┘
                                       │
                           ┌───────────┴───────────┐
                           ▼                       ▼
                   ┌─────────────┐         ┌─────────────┐
                   │  Your AWS   │         │  Your GCP   │ (future)
                   │   Account   │         │   Account   │
                   └─────────────┘         └─────────────┘

AWS Services (MVP)

Category Services
Compute EKS, EC2 GPU (p4d, g5), EFA
Storage S3, FSx for Lustre, EFS
Networking VPC, VPC Endpoints, Security Groups
Identity IAM, STS (OIDC)
Container ECR
Monitoring CloudWatch, CloudTrail

Workloads

Type Frameworks
Training PyTorch DDP, TensorFlow, XGBoost, Horovod/DeepSpeed
Inference vLLM, TGI (HuggingFace), Triton

Repository Structure

cloudplane/
├── services/
│   ├── credential-broker/     # OIDC → AWS STS exchange
│   ├── control-plane-api/     # Projects, Connections, Routing
│   ├── training-service/      # Distributed training jobs
│   ├── inference-service/     # LLM inference deployments
│   └── orchestrator/          # Terraform + K8s operations
├── docs/
└── infra/
    └── iam/                   # IAM trust templates

Example :

Quick Start

1. Set Up IAM Trust

aws cloudformation create-stack \
  --stack-name cloudplane-trust \
  --template-url https://cloudplane.io/setup/aws-oidc.yaml \
  --parameters ParameterKey=ProjectID,ParameterValue=your-project \
  --capabilities CAPABILITY_IAM

2. Link Account

curl -X POST https://api.cloudplane.io/v1/connections \
  -d '{"provider": "aws", "role_arn": "arn:aws:iam::...", "region": "us-east-1"}'

3. Submit Training Job

curl -X POST https://api.cloudplane.io/v1/training/jobs \
  -d '{"framework": "pytorch", "workers": 4, "gpus_per_worker": 8}'

4. Deploy Inference

curl -X POST https://api.cloudplane.io/v1/inference/deployments \
  -d '{"model": "meta-llama/Llama-2-70b", "engine": "vllm", "replicas": 2}'

Docs

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors