How is the recommendation worked out?
We pair every server plan that meets your CPU and memory with every object storage and CDN option, and price each pairing for one month: the plan itself, traffic beyond what it includes (on the big clouds, a per-GB egress fee), storage, requests and CDN traffic. The page shows up to three: the cheapest near your visitors, the one with the fewest providers to pay, and a farther one when it's clearly cheaper. The table lists the rest. With a CDN, we assume 80% of requests are answered from its cache; the other 20% go back to your server or storage (the origin) to fetch the file, and that traffic is billed by the origin's own rules.
Why are the big clouds so much more expensive?
Mostly traffic. AWS, Google Cloud and Azure charge for every GB sent to the internet, about $0.08 to $0.12, while most VPS plans include 1 to 20 TB a month. At a few TB a month, a big cloud's traffic bill is often higher than the server itself.
What are "static files", and what should I enter?
Files that are the same for every visitor: images, downloads, video, scripts. Object storage or a CDN can send them instead of your server. As a rough guide: a blog is mostly static (around 80%), an API or proxy not at all (0%), a download site almost entirely (95%).
When is object storage plus a CDN worth it?
When static files make up most of your traffic and go beyond what your server plan includes. Cloudflare R2 charges nothing for egress, and Backblaze B2 is free to send through Cloudflare or Bunny. A few TB of downloads can then cost a few dollars instead of hundreds.
Why are uploads free but downloads billed?
Providers pay their upstream networks for traffic leaving their network, while incoming traffic brings customers in. So sending data to your server is free; what your server sends to visitors (egress) is what you pay for.
What are Google Cloud's Premium and Standard tiers?
Premium carries traffic on Google's own network to the point nearest the visitor before it joins the public internet, and is priced by destination. Standard hands traffic to the public internet near the data center and costs less. Azure's "Microsoft network" and "Internet" routing options work the same way.
Why does traffic to mainland China cost more?
International links into China are few and expensive. Google Cloud prices Premium tier traffic to China separately, at around twice the normal rate, and routes into China are often slow whatever you pay. Asia-Pacific data centers are usually closest; the latency test on this page lets you compare them.
Why doesn't it compare performance?
This filters by price and specs; it isn't a benchmark. Shared-CPU plans (marked in the table) can slow down when other customers on the same machine are busy. If that matters to you, turn them off under More options.
Is it worth buying several small servers for their traffic?
Sometimes several small plans together include more traffic than one big one. The planner doesn't look for that automatically: change the number of servers under More options and compare.
How accurate and current are the prices?
Every price was checked by hand against the provider's official pricing page. The date is in the footer, and each table row links to its source. All prices exclude VAT and other taxes, which your bill may add depending on where you are. Providers price in US dollars or euros; other currencies are converted at a fixed European Central Bank reference rate (its date is in the footer), so your actual charge may differ slightly. Spotted a mistake? Open an issue on GitHub.
How is latency measured, and what about routes into mainland China?
Your browser sends a few tiny requests to each data center on the page and times them, using the same test points and method as Me2Cloud: one warm-up, then the middle of three results. That's your own network's real latency, not a lab figure. Routes into mainland China vary a lot by provider and carrier, and price tells you nothing about them. If your visitors are in China, open this page from there (or ask one of them to) and compare the Hong Kong and Tokyo numbers.
