In two previous posts, I explored the concepts of "Minimum Viable Clouds" and "More than Minimum Clouds."
To recap, a Minimum Viable Cloud must:
- be an SOA
- hide implementation
- be fully automated
- be a utility
- have global fault tolerance
- be Opex
A cloud that is More than the Minimum must:
- promote stateless, distributed compute
- have asynchronous messaging
- have data persistence as a service
Future clouds are those that move past the performance and composition aspects of cloud-native applications into new territory in efficiency and security. There are many possible futures for cloud, and likely several that will be realized. At Luminal, we have a vision of cloud computing that provides significantly more control, efficiency, and security than is currently available. We're working hard on solutions to allow all the patterns below to be available in the near future.
1. Future Clouds are Markets
Minimum Viable Clouds (MVCs) are Opex utilities, but future clouds are markets that allow purchasing decisions to be dynamic and competitive. We've made progress as a civilization at creating efficiency in markets and cloud should be no different. Amazon Web Services (AWS) has implemented the Spot Market for EC2, which is the first step in this direction. There are companies working on cloud resource exchanges who seek to create a "compute is like oil" abstraction. That belies the nature of compute resources, which are often fairly tightly bound to particular compute problems.
2. Future Clouds Move
Applications have users in many locations. Future clouds support dynamic movement of compute resources based on business need and utilization over time. This must be automated and dynamic so that the surface of the application infrastructure autonomically responds to inputs.
3. Future Clouds are Natively Secure
We now know that you cannot add security onto infrastructure as a layer and expect to be protected. As Amazon CTO Werner Vogels put it: “If firewalls were the way to go, we'd still have moats around cities.” The way you get to true fidelity of systems is through other patterns that are native to the system, not with wrappers. Clouds provide the first real opportunity to impart native security to applications.
With this post I want to paint a picture of how cloud technologies can be leveraged to build a new "whole-greater-than-sum-of-parts" way to compute. It's early days at Luminal, and we believe the best way to chart a course forward is through interaction with customers – we are now taking private beta customers. If this post resonates with you, you might be a good candidate. We look forward to discussing more about our approach in the future and exploring it with you.