How it works

Methodology support before deployment

• Captures/generates standard YAML deployment descriptors across private and public clouds

• Convention over configuration – uses the intelligence of MCO to minimize the size of descriptors (YAML or

• Description-as-code)

• Eliminates redundant, overlapping, conflicting YAML updates

• Easy to use and understand: cost and time savings

• Eliminates vendor lock-in; open architecture allows to write once and deploy everywhere

Methodology support during Deployment

• Observability by using Natural Language Processing and LLM, without adding unnecessary, confusing dashboards

• Zero-admin workload deployment and operation

• Automates optimization of multicloud operation

• The open architecture eliminates vendor lock-in, allows 3rd party tool/providers ecosystem.

• Robust, universal and futureproof

• Knowledge graph model makes it easy to extend and integrate with any current cloud and non-invented yet clouds

• Multicouds can be constructed with any clouds and without normalization of APIs

• Networking effect across vertical stack is part of design: compute, network, data, applications

• Agentless and microservice based

• Open architecture

• Ecosystem for 3rd party tools and vendors

 Detailed Description

Cloud of Clouds instantiates itself as a microservice or a virtual machine running on a customer's premise or a public cloud. It is a distributed control system that is not in the data path. It does not include its own runtime. The runtime is provided by the clouds connected to the Cloud of Clouds. 

The customer must use the account credentials for the hyperclouds they want to control. These credentials are private to customers and are not intercepted or stored by Cloud of Clouds. 

Once connected to the component hyperclouds or private clouds, Cloud of Clouds uses the APIs of the hyperclouds: think of an equivalent of intelligent device drivers in an operating system. First, it discovers the workloads, attributes, and cloud infrastructure associated with that account. If the customer has used YAML files to configure their accounts, Cloud of Clouds can use them to proceed to the discovery process. The result of the discovery process is stored in a knowledge base with only three types of objects: nodes, edges, and attributes. 

Cloud of Clouds uses this knowledge base to create a distributed theoretical model, generalized mathematically as a field space. This field contains all the data, information, and knowledge needed by Cloud of Clouds to describe, monitor, control, and optimize the multicloud infrastructure and associated workloads. We underscore that the only supported workloads are containers and virtual machines.

Cloud of Clouds assumes customers prefer to work with their monitoring systems. By default, it uses the monitoring facilities of the connected hyperclouds, and it includes in the standard binary distribution Prometheus as an option for an on-premise monitoring system. If the customer uses an in-house monitoring system, an interface for that monitoring system needs to be written.