Pick by bottleneck, not by brochure
The create-node flow asks for a category, family and plan. Here's how the families actually map to workloads.
| Family | Optimised for | Typical workloads | Worth knowing |
|---|---|---|---|
| E1 / E1WC | Lowest practical cost | Dev, test, staging, small services | Root storage sized separately (75GB–2400GB); no default public IP — attach one where supported |
| C3 | CPU throughput | Web/API servers, batch processing, build workers, analytics | Compare M3 if you're swapping — RAM may be the real bottleneck |
| M3 | Memory capacity | In-memory databases, Redis-style caches, real-time big data | Check disk/IOPS too — data workloads are often storage-sensitive |
| SDC3 / SDC4 | Dedicated, predictable compute | Production apps sensitive to noisy neighbours; dedicated DB/middleware | Guaranteed resources; Linux and Windows (WSDC4) variants |
| SQL | Windows + SQL Server | SQL Server Web / Standard workloads | SQL edition and version visible in the plan selector |
| GPU / GDC / GDC3 | GPU acceleration | Deep learning, LLM inference, rendering, HPC, parallel processing | GDC3 = next-gen dedicated GPU plans; shows GPU memory & CUDA version |
| Marketplace | Prebuilt stacks | Ready-made application images | Starts from a product, not a bare OS |
The portal is the source of truth for live plans, OS versions and prices — two plans in the same family can differ significantly.
Day-2 operations, handled from one console
Everything below ships in MyAccount — no third-party tooling required.
Snapshots — on-demand & scheduled
Point-in-time copies of the root disk plus selected attached volumes. Lock snapshots against deletion, schedule recurring restore points, or promote a snapshot to a reusable image.
CDP Backups
Continuous protection with recovery points, visible next/last backup times, schedules, and configurable compression — separate from snapshots, purpose-built for restore.
Reserved IPs & VPC
Reserve static public IPs, attach add-on IPs, IPv6 and floating IPs, and place nodes in VPC private networks with security groups controlling inbound and outbound traffic.
Monitoring & alerts
CPU, memory and network graphs per node, with default health alerts on new nodes and custom alert policies that email your team before users notice.
Saved images
Capture a configured node as a golden image and launch identical nodes later — the fastest path to repeatable deployments across environments.
Security by default
SSH keys, security groups, disk encryption where supported, BitNinja server protection, accidental-deletion protection, plus activity, action and billing audit logs on every node.
From signup to SSH in minutes
The create-node flow at Compute → Nodes, as it actually runs.
CPU, GPU or Marketplace
Pick the launch path, then the OS (Linux, Windows, SQL image) or GPU card, and the OS version.
Family & plan
Choose the category and family — E1, C3, M3, SDC3, SQL, GDC3 — then the exact vCPU/RAM/disk/IOPS plan.
Billing & protection
Hourly or committed billing, root storage size, backup on/off, encryption where supported.
Network & launch
Security groups, SSH keys, VPC, volumes — review the live cost summary and launch. Manage everything after via Actions.
Automate it if you'd rather not click
Every node operation in the console is backed by the MyAccount REST API — create nodes, manage snapshots, attach IPs.
# illustrative — see docs.e2enetworks.com/api/ for real endpoints & auth
curl -X POST "https://api.e2enetworks.com/myaccount/api/v1/nodes/" \
-H "Authorization: Bearer $E2E_API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "prod-api-01",
"plan": "C3-8vCPU-16GB",
"image": "Ubuntu-24.04",
"security_group_id": 1234,
"ssh_key": "team-key"
}'Terraform and full REST API references are published at docs.e2enetworks.com/api.
Beyond AI: what teams run here
- Production web and API servers on C3
- PostgreSQL, MySQL and Redis-style caches on M3
- Windows line-of-business apps and SQL Server
- Rendering and graphics pipelines on GPU nodes
- Batch processing, build farms and analytics jobs
Pair compute with the rest of the stack
- Managed Kubernetes for containerised services
- Managed databases without the ops burden
- Autoscaling groups for traffic spikes
- AI workloads on the TIR platform →
- Dedicated GPU training clusters →
Frequently Asked Questions
Everything you need to know about Cloud Compute VMs.
Cloud Compute VMs
Linux virtual nodes, Windows virtual nodes, Windows with SQL Server (Web and Standard editions), Smart Dedicated Compute variants of both, plus marketplace application images. GPU nodes can also run NVIDIA NGC images. Exact OS versions are listed in the create-node flow.
A snapshot is a point-in-time restore copy of node storage (optionally including attached volumes) — use it before risky changes. A saved image is a reusable base for launching new identical nodes. You can also convert a snapshot into a saved image.
Yes — plan upgrades are available from the node Actions menu (state-dependent), E1-series root storage can be increased, and block storage volumes can be attached, detached and upgraded from the Volumes tab.
Commit for a fixed period (up to 365 days) in exchange for lower rates than hourly billing; committed details, renewal settings and period end dates are visible on the node page. Best for steady production workloads.
No — these are full virtual machines managed from MyAccount. TIR nodes are container-based AI workspaces with JupyterLab and ML images. Same data centers, same INR billing; pick the abstraction that fits.
Launch your first node in the next five minutes
Hourly INR billing, GST invoices, no forex.