Content area
As the current landscape for electric vehicles changes, options for remote charging are expanding to keep up. In the United States alone, sales of electric vehicles grew 85% from 2020 until hitting 450,000 units by the end of 2021. While these growing sales are encouraging, commercial charging stations have a long way to go before they are as ubiquitous as gasoline stations are today. The peer-to-peer energy auction helps fill the gap in underserved areas by allowing private homeowners to share their charging facilities with other electric vehicle drivers. The auction framework wraps existing charging outlets with a Cloud-connected microcontroller. These Edge devices communicate with a Cloud message broker for both reporting and session control purposes. Users, both buyer and seller, may interact with the framework through a web-based user interface. The design of this framework provides many challenges, including how to handle persistent storage. The technology used for data storage is key in determining the performance of the auction application and the smoothness of the user experience.
The two technologies considered are a SQL, server-based, structured and a NoSQL, serverless, unstructured database. NoSQL has gained momentum in the last 20 years triggered by the needs of Web 2.0 companies requiring user-generated content, ease of use and interoperability revolving around big data and real-time access. The natural division of management and storage layers allows for a robust serverless implementation: the Cloud service listens for requests and processes using shared and abstracted compute resources. Serverless allows a payment model where each transaction, beyond the free tier of 20,000 writes/50,000 reads per day, is billed rather than paying continuously for a deployed compute instance.
In this study, we investigate Google’s Firestore and Cloud SQL MySQL solutions. We pit the newer serverless, non-relational, NoSQL, document-model database against the traditional SQL, table-based, relational server. Both solutions are evaluated for query performance, cost, flexibility and scalability. Through benchmarking, analysis and a deep-dive of system architecture, we answer whether Firestore can support the energy auction persistent storage needs despite the superior query capabilities of MySQL’s SQL engine.