An open network for compute infrastructure can help meet the growing demand for compute, while keeping markets competitive. A draft of the Concept Paper is available. Please feel free to add comments on the Google Doc itself.
Tanvi Lall [[email protected]] and Swaroop [[email protected]]
Learn more about OCC here.
We live in a world of increasing compute demand. The global cloud compute market is estimated at $500 billion and will be $1.5 trillion by 2030. Currently, much of the compute needs for businesses and startups are met by a handful of large cloud providers. Many fields and applications require significant computing power including AI, scientific research, finance among others and the need for expanded capacity is only growing. The objective of this paper is to brainstorm the creation of an open and decentralised computing network and take a DPI approach to unlocking access to compute-at-scale.
This network of compute would bring many independent providers together on one availability marketplace and improve discoverability. Users of this infrastructure would have access to computing power and related services from any provider. The network would create an extension to the existing computing paradigm and promote decentralised data storage that helps with resilience and overall improved security. This approach can be adopted globally to make markets that offer resilient, interoperable and adaptable environments for innovation.
Scaling compute requires capital and specialised knowledge. Thankfully, due to the advancement of various hardware and software innovations, compute has largely followed Moore's law. The amount of capital and knowledge required to start a cloud has reduced dramatically, and will continue to do so in the near future. Hence, we believe that it is now feasible to create such a distributed network.
Engage with this paper with an open mind, welcoming imaginative exploration over a critical evaluation stance. Likewise, take note of similar decentralised trends emerging in GPU.net and Salad.com. This paper is dedicated to supporting private innovation, particularly benefiting micro and smaller players. While government support is acknowledged, the primary focus is on nurturing a dynamic private sector environment, and discussing the essentials for their success - policy support, technology and infrastructure support, and financial tools among others.
We are looking for enthusiastic volunteers who can own pieces of this design puzzle. Please reach out to [email protected] if you have expertise and/or interest to:
- Establish a strong business case for an OCC network: (a) Understand pain points that existing data centre providers and software service providers of this stack face. (b) Identify use cases and applications that could be early adopters of this network. (c) Identify incentive schemes that lower the barrier for providers and customers to join this network.
- Design a technical architecture that gives providers and customers a positive NPS experience.
Our hypothesis is two fold:
- There is a need for a network, where consumers can discover heterogeneous compute providers, allowing provisioning and usage via standards-based interfaces, thereby democratising access to cloud computing infrastructure.
- There is a need for a financial incentive structure built around encouraging new people to (1) come invest in this compute and (2) to use the compute.
- Concept paper
- Overview of the project in powerpoint format
- Q1 2024 update