Cloud Footprint History (wss-cloud-footprint)

Point-in-time captures from 6 sources that do not keep their own history. Each capture stores the bytes as served, with a manifest row recording the URL, fetch time and SHA-256.

History held: 2026-09-01 to 2026-10-13.

Findings

Figures are rebuilt from the derived tables on every run; each one names the entities it is about so it can be acted on without a further query.

**Declared capacity over time** — Line chart of declared interconnection capacity per network across weekly captures.

**Shared, or owned?** — Scatter of metro total declared capacity against the largest network's share of that metro.

**Two ways to build a network** — Scatter of internet exchanges present at against average declared capacity per exchange, bubble area total capacity.

**What a new region still lacks** — Stacked bar per AWS region showing the opening-kit services, services added since, and services still missing.

**The rollout frontier** — Ranked bar chart of AWS services present in the fewest regions, naming the regions for the thinnest.

What is captured

source	publisher	cadence	terms
`aws.infra.ip-ranges`	Amazon Web Services	weekly	public AWS endpoint; see https://aws.amazon.com/service-terms/
`aws.services.regional`	Amazon Web Services	weekly	public AWS endpoint; see https://aws.amazon.com/service-terms/
`azure.servicetags.public`	Microsoft Azure	weekly	Microsoft publishes this openly for firewall configuration
`peeringdb.exchanges.directory`	PeeringDB	weekly	PeeringDB API; see https://www.peeringdb.com/ terms — VERIFY before redistributing derived data commercially
`peeringdb.facilities.geo`	PeeringDB	weekly	PeeringDB API; see https://www.peeringdb.com/ terms — VERIFY before redistributing derived data commercially
`peeringdb.networks.capacity`	PeeringDB	weekly	PeeringDB API; see https://www.peeringdb.com/ terms — VERIFY before redistributing derived data commercially

Using it

Derived tables are CSV under data/ in the repository; the bytes they came from are under raw/, and SOURCES.md lists every source URL.

Every observation carries source_id and raw_ref. The matching manifest row holds that raw_ref with the URL, the fetch timestamp and a SHA-256 of exactly what came back, so any number here traces to the bytes it came from.

Cite this dataset by its repository and the date you took it from: https://github.com/q3dresearch/wss-cloud-footprint, accessed 2026-10-13.