Units & conversions

Most bad cost estimates start as unit mistakes. Convert explicitly before you model pricing tiers.

All calculators Methodology Estimation checklist

Quick conversions

Mbps �?MB/s: divide by 8 (megabits to megabytes).
GB �?GiB: GB is decimal (10^9 bytes). GiB is binary (2^30 bytes). 1 GiB �?1.074 GB.
Per 10k vs per 1M: normalize pricing units first (a 100× error is common).
Hours/month: for estimates, 730 is a typical month; adjust for schedules.

Throughput �?monthly GB

A safe workflow for transfer estimates:

Convert Mbps to MB/s (divide by 8).
Multiply by a utilization factor (average / peak).
Multiply by seconds/day and hours/day (or 730h/month for always-on).
Convert MB to GB (divide by 1000 if your pricing is per GB).

Example: 50 Mbps average, 12 hours/day �?(50/8) MB/s × 12 × 3600 × 30 / 1000 �?810 GB/month.

How to get your inputs

Requests: load balancer metrics, API gateway metrics, or application dashboards (RPS).
Transfer: CDN origin egress, load balancer processed bytes, or VPC flow logs summaries.
Storage: allocated vs used (billing often charges allocated, monitoring often shows used).
Time: scheduled uptime (work hours) vs always-on (730h/month).

Common mistakes (and how to avoid them)

Mixing bits and bytes: “Mb�?is bits; “MB�?is bytes. Double-check the capital B.
Using GiB numbers with GB pricing: convert before multiplying by $/GB.
Assuming 100% utilization: average throughput is often far lower than peak.
Forgetting per-request unit size: “per 10k�?vs “per 1M�?changes everything.
Using the wrong month length: scheduled workloads rarely run 730 hours.

Validate after you deploy

Compare estimated monthly GB to actual “bytes transferred�?in your provider metrics.
In billing/Cost Explorer, group by usage type and confirm the unit (GB, requests, hours).
Re-check after traffic changes: launches, batch jobs, and incident retries can shift averages fast.

Apply conversions to cost drivers

Convert throughput to monthly GB before using transfer or CDN calculators.
Normalize request units (per 1k, per 10k, per 1M) before multiplying by volume.
Convert GiB to GB if pricing is decimal, then use the storage calculators.
Use a consistent hours/month assumption across all compute estimates.

Scenario planning

Baseline: average throughput and request rates with normal utilization.
Peak: higher throughput and request volume during launch or incident weeks.
Scheduled workloads: compare 730 hours vs your real schedule.
Unit drift: keep a checklist for GB vs GiB and per-1k vs per-1M.

Validate after changes

Confirm units in billing exports match your calculator inputs.
Re-check conversions after switching providers or regions.
Validate monthly totals against a real week of metrics.

1 unit tools

Cloud Units Converter

Convert GB↔GiB, Mbps↔MB/s, and estimate monthly transfer from throughput.

Sources

AWS data transfer pricing (GB units vary by service): AWS EC2 On-Demand pricing
Example of AWS documentation using decimal GB units in storage contexts: Amazon S3 User Guide

Assumptions: pricing pages use the provider’s billing units (often decimal GB). Convert your monitoring numbers to the same unit before multiplying by rates.