Today: A Bloated Stack Forced to Work in a Distributed Way
The world no longer builds applications suited to single machines, data-centres, or even cloud providers but relies on a compute stack that is a bloated monument to backwards compatibility.
Reliant on layered abstractions to make multiple machines behave as one, we continue to use middleware to treat the symptoms, instead of looking deeper at the root cause.
Every application has become distributed, but relies on middleware to communicate.
Tomorrow: An OS Designed to be Distributed First and Real-time
Using first principles we asked to ourselves "If we accept that all applications of the future will be real-time, hyperscale and distributed, what changes would we need to make?"
The answer was surprisingly simple - "start again". We'd bring the logic for distributed applications closer to the metal for reliability, and control, and we'd re-engineer the processes to be distributed-first instead of relying on middleware.
So, we set about building an operating system to support the future instead of conforming to the past.
Under the hood of HadeanOS
Hadean sits directly on top of bare metal. It removes all the leaky abstractions that reduce control, and introduce unreliablity.
Augments processes with intrinsic distributed primitives. These are inspired by rigorous models such as CSPs and Occam which have seen success in the likes of Erlang.
Processes cannot observe one another or overallocate resources. This significantly reduces the crashes caused by isolation bugs and unpredictable resource allocation.
Monte Carlo Simulation
Run complex Value at Risk calculations across hundreds of machines with ease and get answers in minutes.Read more...
Perform genetic alignment on one, ten, or ten thousand genomes with a constant exection time regardless of volume.Read more...
Build worlds of unprecendent scale and fidelity without compromising on NPC AI whilst users are inactive.Read more...