Five products, one production stack
Each works standalone. Together they run a complete application — VPC-networked, monitored, and scaled.
Cloud Compute VMs
CPU-intensive, memory-optimised, dedicated and GPU node families with OS images, snapshots, CDP backups, reserved IPs, VPC and monitoring. For web apps, databases, SQL Server, rendering — and everything else.
Explore Compute →Managed Kubernetes
Managed control plane with static and autoscaling node pools, Ceph-backed persistent volumes, MetalLB service IPs, and kubectl/dashboard access. Run containerised workloads without babysitting masters.
Explore Kubernetes →Database as a Service
Managed MySQL, PostgreSQL and MariaDB with automatic failover, read replicas (including cross-region Delhi ↔ Chennai), scheduled backups, parameter groups and SSL. Your data stays in India.
Explore DBaaS →Load Balancer
Application (L7) load balancing with SSL termination, ACL routing and health checks — or Network (L4) for raw TCP throughput. Round robin, source-IP hash, or least-connection algorithms.
Explore Load Balancer →Auto Scaling
Scale groups built from saved images with elastic policies (CPU, memory, network, IOPS triggers) and cron schedules — wired directly into the load balancer as a backend.
Explore Auto Scaling →How they fit together
A typical production setup: an Auto Scale group of app VMs (or a Kubernetes node pool) behind a Load Balancer, talking to DBaaS over VPC — with TIR model endpoints serving inference alongside. One console, one invoice.
Get an Architecture Review →Which service for which job?
A quick decision guide.
| You need to… | Use | Why |
|---|---|---|
| Run an app, API, or legacy workload on a server | Compute VMs | Full control, snapshots, choice of node families |
| Run containers with orchestration and self-healing | Managed Kubernetes | Managed control plane, autoscaling node pools |
| Run MySQL/PostgreSQL/MariaDB without ops burden | DBaaS | Failover, replicas, backups handled for you |
| Distribute traffic across multiple servers | Load Balancer | L7 SSL termination or L4 TCP, health checks |
| Handle traffic spikes without overprovisioning | Auto Scaling | Policy and schedule-driven scale out/in |
| Train or serve AI models | GPU Cloud + TIR | Frontier GPUs and the managed AI platform |
Everything ships with the E2E baseline
Common to every compute product.
INR billing
GST invoices, no forex exposure, published rates.
Indian data centers
Delhi NCR and Chennai — data never leaves India.
SOC 2 + ISO
Certified infrastructure, VPC isolation, security groups.
Human support
Engineers who pick up — no paid support tiers.
Build your production stack on E2E
Start self-service, or get an engineer to map your architecture.