Cloud Egress Cost Calculator

A cloud egress cost calculator prices the data-transfer portion of a cloud bill, which is routinely the least understood and most surprising line item on it. Compute and storage are easy to reason about because you provision them deliberately. Data transfer is different: it is charged per gigabyte on traffic your architecture generates as a side effect, often without anyone choosing it, and the rates differ by an order of magnitude depending on which boundary the bytes cross.

The calculator prices four distinct paths. Internet egress from origin is the most expensive per gigabyte and covers traffic leaving your cloud directly to users. CDN delivery covers the share offloaded to an edge network, typically at a substantially lower rate, and the calculator removes that share from origin egress so the offload benefit is modelled properly rather than double-counted. Cross-region transfer covers replication and multi-region architectures, where data moves between geographic regions. Cross-availability-zone transfer covers traffic between zones inside a single region — the line teams forget most often, because it is generated invisibly by replicated databases, chatty microservices, and load balancers distributing across zones, and several providers charge it in both directions. Summing all four gives the monthly total, and the calculator also computes what the same traffic would cost with no CDN at all, so the offload saving is explicit.

Two lessons usually emerge from running real numbers. First, CDN offload is the single largest lever on user-facing transfer cost, because it moves the bulk of your bytes from the expensive origin-egress rate to a much cheaper edge rate — and raising cache-hit ratio through better cache headers and longer TTLs is nearly free money. Second, internal traffic can rival or exceed external traffic on the bill without appearing anywhere in your product metrics. A multi-AZ database cluster replicating continuously, or services in different zones exchanging large payloads, can quietly generate tens of terabytes a month. Zone-aware routing, colocating chatty services, compressing inter-service payloads, and reconsidering whether every workload truly needs multi-region replication are the usual fixes. The source research recommends comparing cross-AZ, cross-region, CDN, internet, and backup paths with live provider tables, and that remains the right approach — rates vary by provider, destination, and volume tier, so enter your own. This prices transfer only, excluding requests, storage, and CDN feature charges.