Future Features

Future features

One layer on top

Implementing a network layer on top to create a communication between multiple networks. This could lead towards a decentralised, distributive computation network.

The abstract idea behind this layer is a network, that provides the users with the ability to access computational resources if needed with autmatic load balancing over the whole network.

My personal example is this:

I record a video with my smartphone, uploaded the raw material to a cloud server and now want to cut it. I start my cutting application on that smartphone (or any other device in my reach, like a tablet).

As those devices are connected to my Hydranet ("Being hydrated", pun intended), I can cut the movies directly on my device and render them remotely. By starting the rendering of the new movie with possibly another encoding, my device automatically selecteds the nearest Hydra (getting location information by GPS for example) that provides the service Render Movie. Those nodes could be one of my one machines, or a machine inside my peer group or ring of trust.

The Hydra starts rendering my movie by using its own hardware and/or delegating the work to other hydras inside its ring of trust.

As a result, it could be possible, that the movie, I recorded in my vacation in Berlin could be rendered on a Hydra in Hamburg which delegates the work to other Hydras all over the world.

It should be possible that a Hydra can request the code for a service from another Hydra, to expand its capabilities. As a result, their could be malicious code, which makes the security of code a prime aspect of the implementation.

This layer needs sophisticated routing algorithms to transport services and their data throuh the whole network of interconnected hydras which is the second main aspect of the implementation.

In this layer a Hydra can has a single node, meaning only one single machine, a group of machines forming a network an also a cloud computer like Amazons EC2 which can spawn new node or even complete Hydras up to their needs. The boundaries between single and multi-node Hydras blur. The big difference is the security and the interconnection. A node inside a Hydra knows every other node and its capabilities and trust its process code - A Hydra doesn't. Their is no need to know every single Hydra in the world, especially not those malicious one.

The only thing a Hydra needs to know is:

• What can I do?
• If I can't do something, do I know someone who can (or someone, who knows someone who can...)?

This theoretically enables a Hydra to perform every thinkable task, either on its own nodes, or by delegating the task to (multiple) other Hydras it trusts.