Each provider (a user) installs their own OpenStack cluster, which includes a nova controller. That cluster runs a virtual appliance who's job is to watch a given Bitcoin address for payments. If it sees a payment it either starts an instance or resets a counter to keep an instance running. Payments by users are taken by the pool operator. The pool then starts micro drips to the instance, in most cases half a cent or so.
The pool I'm going to run will be at stackmonkey.com. If a provider shuts off their controller, something crashes, or there's a rig failure, only a small percentage of the total number of instances will be affected. Payments won't matter because the instance costs are so low, and the drip can be turned off when the appliance stops responding to the pool controller.
OpenStack makes this handy because it handles all the heavy lifting for starting and stoping instances, doing network setup, images, instance sizes, etc.
Initial audience for this low trust pool will be hackers, developers, testers, tor nodes, individuals, etc. Very low cost, highly ephemeral instances.
Thanks for the whitepaper and video. So, at high level the resource provider is accepting Bitcoin instead of dollars to provide a virtual instance to a customer. Does this sums up what you are trying to do?
I have hard time imagining who might be the customer that will be willing to run their virtual instance on low trust ephemeral pool and for what purpose. What are some use cases for your solution? Is there such a low trust ephemeral pool already in existence that accepts dollars? Why would someone use this solution over just launching a VPS or EC2 instance?
Each provider (a user) installs their own OpenStack cluster, which includes a nova controller. That cluster runs a virtual appliance who's job is to watch a given Bitcoin address for payments. If it sees a payment it either starts an instance or resets a counter to keep an instance running. Payments by users are taken by the pool operator. The pool then starts micro drips to the instance, in most cases half a cent or so.
The pool I'm going to run will be at stackmonkey.com. If a provider shuts off their controller, something crashes, or there's a rig failure, only a small percentage of the total number of instances will be affected. Payments won't matter because the instance costs are so low, and the drip can be turned off when the appliance stops responding to the pool controller.
OpenStack makes this handy because it handles all the heavy lifting for starting and stoping instances, doing network setup, images, instance sizes, etc.
Initial audience for this low trust pool will be hackers, developers, testers, tor nodes, individuals, etc. Very low cost, highly ephemeral instances.
There's a video here: https://vimeo.com/86717476