Zero overhead cloud management for hypervisors anywhere.
Classical cloud management solutions like OpenStack, SDC, or even Project-FiFo itself require you to colocate the management infrastructure alongside the hypervisors. Colocating the management infrastructure comes with some problems, starting from cost over enforced locality to the knowledge necessary to run cloud management solutions.
With FiFo Cloud we are taking a different approach to the problem. The fault tolerant architecture that lays the foundation of FiFo allows us to move the management infrastructure offside with minimal risk to running applications. This completely removes infrastructure cost, knowledge costs and provides management to span many sites and even different providers.
Since FiFo Cloud goes away with the need for dedicated hardware, it makes the footprint near neglectable.
Other systems require dedicated servers to run taking away gigabytes of memory and making running your cloud in many cases prohibitively expensive. With FiFo Cloud the only a small agent runs on the hypervisor, reducing the usual footprint by orders of magnitude.
This is especially an advantage for smaller installations where each bit of memory counts as it allows you to manage a single hypervisor with the same tools and technology you would use for a 100 and seamlessly scale out as you need to.
FiFo cloud allows us to offer a price that is simple and competitive. While traditional solutions calculate the cost based on memory, CPU sockets or cores and in result incentivize specific builds, FiFo Cloud has a simple per host pricing.
It does not matter if you want to build a system with 2 TB of memory, 500 CPUs or a petabyte of Disk you pay a monthly flat fee per host independent of your design.
As we are a small company and know the constraints that smaller setups work under we offer special pricing for small installation on top of that.
One of the fundamental concepts behind FiFos development was to ensure fault tolerance. Moreover, this does not only mean not to try to break something. Moreover, it is designed to handle errors gracefully, no matter what.
The most fundamental aspect of that is that every Zone, every Jail, and every VM work on their own without outside dependencies. As in the end, your zones and applications are what matter most so this is the failure core that can not, under any circumstances, go down.
To make this possible we use a shared-nothing architecture that makes every zone self-sufficient and leaves complete control in your hands and your hands only.
A welcome side effect is that we have