Cloud provider discount & payment / Google Cloud
GCP Compute Engine VMs
Per-second virtual machines: machine families, regions, and discounts
Compute Engine is the virtual machine layer under Google Cloud — the same category of product as AWS EC2 or Alibaba ECS, with three differences that actually move the bill. Billing is per second with a one-minute minimum, so spiky workloads are not billed for a full hour. Machine families are cleanly tiered: e2 for general work, n2 for performance, c2d for CPU throughput per unit of memory. And egress defaults to the Premium network, which routes well but costs more per GB than the Standard tier. We open the GCP project and act as payment intermediary: no personal credit card required, no identity documents submitted. Once the account and billing are live you order directly in the Google Cloud Console.
Specs & pricing
| Machine type | vCPU / memory | Network egress | Common disks | Good for |
|---|---|---|---|---|
| e2-standard-2 | 2 / 8 GB | 10 Gbps default | pd-balanced | Web apps and APIs |
| n2-standard-4 | 4 / 16 GB | 10 Gbps default | pd-ssd / local-ssd | Production and databases |
| n2-highcpu-8 | 8 / 8 GB | 10 Gbps default | pd-balanced | Compiles, transcodes, batch |
| c2d-highcpu-8 | 8 / 16 GB | 10 Gbps default | local-ssd | CPU-dense stream processing |
What Compute Engine is: same category as EC2, different in three ways
It is Google virtual machines: operating systems, containers, and databases all run on it. What actually moves the bill is billing granularity, network tier, and automation — instance templates, managed instance groups, and startup scripts are all built in, so templated deployment and scheduled scaling need no extra tooling.
Choosing a machine family: e2, n2 and c2d cover most cases
| Family | Typical ratio | Positioning | Watch out for |
|---|---|---|---|
| e2-standard | 2–64 vCPU, 1:4 to 1:1 | The value default | Move sustained load to n2 |
| n2-standard | 2–96 vCPU, 1:4 to 1:1 | Production general purpose | Higher price buys better per-core speed |
| c2d-highcpu | Up to about 4 vCPU / GiB | Transcode, compile, streaming | Memory is tight |
| t2d (AMD) | 2–96 vCPU | General purpose, good value | Verify platform-specific dependencies |
Billing: per-second, SUD, CUD, and Spot
- On-demand, per second: billed to a one-minute minimum; best for development, priciest over long periods
- SUD (sustained use discounts): automatic, nothing to apply for; ramps monthly and typically lands in a 40–50% range around month six
- CUD (committed use discounts): a one- or three-year term, either resource-based (shape locked) or flexible (shape can change)
- Easily missed line items: egress, static IPs, boot disks, snapshots, and load balancer processing are each billed separately
Cutting a GCP bill: exhaust the discounts first
- Check the effective SUD rate: compare it against the list rate in billing — there is nothing to switch on
- Convert steady workloads to CUD: if it will run a full year, resource-based CUD on a fixed shape saves more
- Turn on budget alerts: Budgets plus email, and FinOps Hub will flag idle resources
- Rightsize: an 8 vCPU machine pinned at 20% CPU belongs on e2-standard
Regions: Hong Kong (asia-east2) and Taiwan (asia-east1)
| Region | Location | Good for | Notes |
|---|---|---|---|
| asia-east2 | Hong Kong, China | South China, Hong Kong, Macau | Closest; fewer resource types than US West |
| asia-east1 | Taiwan | Taiwan users, or a second APAC footprint | An underrated, compliance-friendly option |
| asia-northeast1 | Tokyo, Japan | Japan and Northeast Asia | The default for the Japan market |
| asia-southeast1 | Singapore | Southeast Asia | Usually beats US West for Indonesia, Vietnam, Thailand |
Side by side: GCP against AWS and Alibaba Cloud
| Dimension | GCP | AWS EC2 | Alibaba ECS |
|---|---|---|---|
| Billing granularity | Per second, 1-minute minimum | Per second, 1-minute minimum | Hourly or monthly/annual |
| Value general purpose | e2 family | m6i / t3 families | t6 / g families |
| Compute-dense | c2d | c7i / c6i | High-frequency g8i |
| Sustained discount | SUD, automatic | Savings Plans, opt-in | Savings plans or reserved instances |
How to open the account and start with us
- Describe the workload: CPU and memory needs, plus the target region (Hong Kong, Taiwan, Tokyo, Singapore, or US West)
- We handle provisioning and payment: no personal credit card bound, no identity documents submitted
- You operate in the console once live: enable APIs, build instance templates, set budget alerts
- Run on-demand first, commit later: watch two or three months of real usage before buying a commitment
Four common mistakes
- Assuming shutdown is free: SUD requires continuous runtime, and nightly restarts never reach the deeper bands
- Running production on Spot: the preemption notice can be up to 60 seconds, so stateful services do not belong there
- Ignoring Premium egress: outbound traffic is billed per GB and often the biggest line on image- or video-heavy bills
- Assuming cross-region traffic is free: traffic between GCE instances in one region is free; crossing a region is not
Typical workloads
Open a Google Cloud account with us
We handle official provisioning and payment verification — no personal card, no identity documents. Order directly in the vendor console once live.
More Google Cloud products
Related reading
AWS Shield Advanced: what is free, what is billed
Shield Advanced covering your WAF costs is true, with four boundaries: only protected resources, only standard features, and subject to a request ceiling. Bot Control, CAPTCHA and web ACLs above 1,500 WCUs are still billed separately.
Cloud CostAWS private pricing (EDP): threshold, risk, shortcut
Private pricing is the highest-threshold layer of the AWS discount system: a cross-service discount rather than a single-service commitment. Two traps come first - a shortfall payment if you miss the commitment, and mandatory enterprise support. And AWS states plainly that buying through a partner qualifies too.
Cloud CostMulti-cloud cost visibility: unify categories first
Each provider runs its own billing system, so there is no ready-made unified view. Most people reach for a tool first, which reverses the order: unify the cost categories and the tagging scheme first. Four accounting mismatches will make the numbers disagree.