The problem
The star that drained the web
Networks are built as a star because running a network is genuinely hard. Routing, addressing, security — the sheer complexity of operating a network pushed everything toward a central upstream that owns the addresses and forwards the packets. Simplicity won; the star topology became the shape of the internet.
The catch: your data flows upstream by design. And information flows are gold — your IP in both senses, intellectual property and internet protocol, drifts toward whoever sits above you. That gravity drained the value of the web into a handful of capital-rich hyperscalers. Not because they built better software — because the topology pointed the money their way.
Why the star won
46-year-old shoulders
IPv4 was introduced in 1980. Almost everything we've built since stands on that protocol. Its header is tiny — it can't list every hop a packet will take — so we bolted on BGP: upstream devices literally shout "I know the way to Bob," your device trusts them implicitly, and off the packet goes. Addresses themselves are issued by the upstream. NAT then hides your device behind the router; your phone has no persistent address of its own. Step onto café WiFi and your counterparty can no longer reach you directly at all.
The result is structural: the cloud is the natural direction — data already flows to the aggregator — while the edge is made artificially hard. IPv6 kept the same star; it was only a marginal improvement, which is exactly why adoption stalled.
What's changing
The ground is shifting
Segment routing (SRv6) is rolling out at carrier scale — not lab trials, live revenue-carrying networks:
Star → mesh: data stops draining to the centre and starts routing to you.
It works by writing programmable "postal instructions" onto the envelope. The data plane can now peer directly, devices become individually addressable, and — crucially — identity moves into the packet's extension header. The where (a location-based address) is decoupled from the who (a persistent, cryptographically verifiable identity). New IP / IPv6+ (Xinghe) pushes that identity to a first-class network citizen; privacy is handled with zero-knowledge — in February 2026 the first international ZKP standard, ISO/IEC 27565, was published.
Individually addressable devices, identity as a header property, and the long-standing demand for direct peer-to-peer addressing all arrive at once. The moat is opening.
The consequence
SaaS is dying
When every device is addressable and routing can steer traffic close to you, there is no real reason left to rent ten services running in someone else's datacenter. Direct addressing near the edge wins even where a central control plane still exists — it lowers latency, raises reliability, cuts cost, and moves faster.
The moat was never the software. It was the topology. The topology is changing.
The first step
Noolog — infrastructure, handed back
Noolog is the first step toward running your own apps — and delegating apps to a trusted network — without ever asking "is my cloud provider compliant?" Compliance is out of the box: a root certificate authority baked into the HSM on every node. Trust is rooted in the hardware and cross-coupled across manufacturers, so no single vendor has to be trusted.
Your node does your work, holds your keys, and earns when it serves others. You stop renting most of your SaaS. Personal freedom, endless human-centric collaboration, the boring work left to machines — running on infrastructure you own. From SaaS to mist.
Built in the open
Open-source software on trusted hardware
The trust runs all the way up. Noolog's interface is written in open-source Slint — statically typed, compiled to native code with no garbage-collected runtime and deterministic execution — so the software steering your node is as verifiable as the HSM rooting it. That's the engineering discipline that gets safety-critical UIs certified. See how it maps onto regulation.