Load Balancing and Carbon Emission Control for Sustainable Cloud

Abstract

The system named Stratus, intended for big public cloud infrastructure, is described in the abstract. Several power plants, each with different electrical costs and carbon emissions, power this system. For every request, the system attempts to direct traffic to the data center that is closest in terms of geographic distance, least expensive power, and least amount of carbon emissions. This is accomplished by modeling the networking and computing components as graphs and using Voronoi partitioning to identify the data center for routing according to the priorities set by the cloud operator.  The significance of cloud computing services is emphasized in the introduction, along with the difficulties associated with load balancing, latency, operating expenses, and carbon emissions in data centers. The advantages of distributed servers globally in terms of lower latency and operating expenses are mentioned. It also covers the possible long-term effects of carbon emission rules as well as the mounting worry over carbon emissions from data center electricity.