why a network, not a datacenter

When the law changes,
capacity follows.

A cloud provider answers a new data-residency rule with a capital project: land, power, permits, a region announced years out and priced accordingly. There is nothing you can do but wait for it and pay for it. A network answers the same rule with an incentive, and the capacity appears where the requirement is. That is not a feature we shipped. It is the reason we built a network instead of a datacenter.

See the live network → How it works

The unit of expansion

One of these takes three years. The other takes a weekend.

Every storage business grows by adding capacity. What differs is the smallest unit of capacity it is able to add, and who has to authorize it. That single difference decides how fast either model can respond to anything — a law, a customer, a region nobody predicted would matter.

A datacenter

unit of expansion: a building

Site selection, land, power contracts, permits, construction, commissioning. The decision is a capital allocation, so it competes against every other capital project the company has, and it is made by people who need the demand proven before they will fund it.

Timeline measured in years. If the demand is real but small — one country, a few dozen customers — it never clears the bar at all.

A network

unit of expansion: one machine

Someone who already has hardware, power and bandwidth in the right country installs the software and starts earning for the capacity they contribute. Nobody has to approve it. There is no capital committee, because there is no capital being committed.

Timeline measured in the time it takes to rack a machine. Small demand is still demand, because the smallest useful response is one node rather than one building.

Where it actually matters

Four situations a building cannot answer in time.

A residency rule lands with a compliance date

Your regulator decides certain records have to stay inside the country, effective next year. If no hyperscaler has a region there, the answer is to wait for one or to move the workload. On a network the question is different: is there capacity in that country, and if not, is there a reason for operators there to add it.

Your country is not on the region list

Most of the world is not. Cloud regions cluster where the demand is densest, which means a business in a smaller market chooses between latency, a foreign jurisdiction, and running its own hardware. Contributed capacity does not have that geography problem, because the operators already live there.

You need spread, not just a location

Two copies in two buildings owned by one company share a single operator, a single legal entity and a single failure of corporate judgment. Fragments across independent operators in different countries do not. Composing that is a filter setting here; it is not purchasable at all from a conventional provider.

The requirement is real but small

A few dozen customers in one jurisdiction will never justify a building. It comfortably justifies a handful of machines. The economics of the smallest useful response are the whole argument on this page.

The other side of it

What a network gives up in exchange.

A model with only advantages is a sales pitch. This one trades away things a single-owner datacenter has, and a buyer should price them in rather than discover them later.

No single throat to choke. Independent operators mean no one company is contractually on the hook for every machine holding a fragment of your data. What replaces it is arithmetic: enough redundancy that individual operators failing is expected and survivable, with the model published so you can check it rather than asserted.

Capacity responds to incentives, not to promises. Operators appear where it is worth their while. That is what makes the network fast to respond, and it also means we cannot guarantee a specific country will have capacity on a date you choose. What we can do is show you the live map before you commit, and let you compose your provider set from what is actually there.

You operate the endpoint. The gateway is yours to run, patch and monitor. What XNS cannot do sets out both what that buys you and what it costs you.

Look at what is out there before you decide.

The live map shows every operator currently holding capacity, where it sits and what it is measured at. Compose a provider set against it and see how many machines match your requirements before you write a single byte.

Open the live map

Product

Why XNS

Company

Legal

© Copyright - SCP, Corp | Xa Net Services and Affiliates