Establishing a multi-cloud environment

A computer program product for establishing a multi-cloud environment, the computer program product comprising one or more computer readable storage mediums having program code embodied therewith, the program code comprising programming instructions for:

• modeling the multi-cloud environment using a game theoretic model by defining the multi-cloud environment to consist of cloud buyers, such that each cloud buyer has a machine learning workflow to run, wherein the machine learning workflow is composed of a data extraction task and a machine learning task, wherein the game theoretic model is a trading network corresponding to a business-to-business network linking the cloud buyers and cloud providers in a marketplace;
• creating a probability distribution over information technology costs for the cloud buyers and the cloud providers, wherein the cloud buyers are purchasers of computing resources, wherein the cloud providers provide on-demand, scalable computing resources;
• calculating an estimate of an expected utility for each party in a single-cloud environment;
• creating a centralized broker for the multi-cloud environment that collects payments from the cloud buyers and makes payments to the cloud providers using the estimated expected utility for each party in the single-cloud environment and the probability distribution over information technology costs for the cloud buyers and the cloud providers as parameters to the game theoretic model;
• establishing the multi-cloud environment by the centralized broker by identifying a set of bi-lateral trades that constitutes an equilibrium in the trading network, wherein the centralized broker utilizes the probability distribution over information technology costs and the estimate of the expected utility for each party in the single-cloud environment to determine a valuation function for each cloud buyer and cloud provider, and wherein the centralized broker routes payments for the set of bi-lateral trades based on a real value mapped by the valuation function;
• and
• automatically routing data traffic for the machine learning workflow between at least two different public cloud platforms via a network gateway according to the set of bi-lateral trades identified by the centralized broker to maximize utility for the cloud buyers and the cloud providers.

A system, comprising:

• a memory for storing a computer program for establishing a multi-cloud environment;
• and
• a processor connected to the memory, wherein the processor is configured to execute program instructions of the computer program comprising:
• modeling the multi-cloud environment using a game theoretic model by defining the multi-cloud environment to consist of cloud buyers, such that each cloud buyer has a machine learning workflow to run, wherein the machine learning workflow is composed of a data extraction task and a machine learning task, wherein the game theoretic model is a trading network corresponding to a business-to-business network linking the cloud buyers and cloud providers in a marketplace;
• creating a probability distribution over information technology costs for the cloud buyers and the cloud providers, wherein the cloud buyers are purchasers of computing resources, wherein the cloud providers provide on-demand, scalable computing resources;
• calculating an estimate of an expected utility for each party in a single-cloud environment;
• creating a centralized broker for the multi-cloud environment that collects payments from the cloud buyers and makes payments to the cloud providers using the estimated expected utility for each party in the single-cloud environment and the probability distribution over information technology costs for the cloud buyers and the cloud providers as parameters to the game theoretic model;
• establishing the multi-cloud environment by the centralized broker by identifying a set of bi-lateral trades that constitutes an equilibrium in the trading network, wherein the centralized broker utilizes the probability distribution over information technology costs and the estimate of the expected utility for each party in the single-cloud environment to determine a valuation function for each cloud buyer and cloud provider, and wherein the centralized broker routes payments for the set of bi-lateral trades based on a real value mapped by the valuation function;
• and
• automatically routing data traffic for the machine learning workflow between at least two different public cloud platforms via a network gateway according to the set of bi-lateral trades identified by the centralized broker to maximize utility for the cloud buyers and the cloud providers.