August 2021 Summaries

The project involved building a serverless application using Cloudflare Workers, SingleStore database, and OpenSky API to display real-time flight data worldwide. The team used Cloudflare Workers' cron trigger feature to automate the process of fetching data from the OpenSky API and uploading it to an S3 bucket. They also created a REST API using Cloudflare Workers that interacted with the SingleStore cluster via the SQL HTTP API, leveraging its connection pooling capabilities. This allowed them to simplify their architecture by removing the need for S3 in the future. The project utilized SingleStore's geospatial capabilities to filter flight data based on country boundaries. Continuous Integration and deployment were set up using Cloudflare Pages and GitHub Actions, allowing the frontend and API to be auto-deployed on every commit to the main branch. The hackathon experience was valuable, with the team learning about SingleStore's capabilities and how to build serverless edge applications. They also highlighted some rough edges in the technology but were pleased with the support from the Cloudflare Workers community.

We moved away from Next.js 11 months ago and haven't looked back since. The reasons for this were that we found the framework's static rendering feature to be overly complex, making it difficult to integrate custom webpack configurations and extensions. Additionally, Next.js' dynamic routing limitations and performance issues during local development made it unsuitable for our application's needs. We experienced severe performance issues with the local development web server, which was a major pain point. To overcome these challenges, we adopted a client-side routing library (react-router 6) and wrote our own custom webpack configuration. The migration allowed us to simplify our codebase, improve local development speed, and open up new possibilities for complex routing refactors. We've found that maintaining our own solution has given us the freedom to move faster and tackle more complex problems.

This guide provides a step-by-step tutorial on setting up a basic SingleStore application using Node.js. It covers the basics of what makes SingleStore unique, such as its real-time and high-throughput query capabilities, making it suitable for both transactional and analytic workloads. The guide then walks through the process of creating a new project with Node.js, installing the necessary dependencies, including MySQL2, setting up a free SingleStore managed cluster, creating a database and table, connecting to the cluster using Node.js, and performing CRUD (create, read, update, delete) operations on the database. The tutorial concludes by providing resources for further learning and support.

Setting up a machine learning model on data has become significantly easier with the integration of MindsDB and SingleStore. By leveraging an additional predictive layer on top of their database, developers can automate model training while utilizing existing datasets, saving time and cost. The process involves setting up a SingleStore account, creating a cluster, importing a dataset into the database, and connecting it to MindsDB to run predictions and queries using simple SQL. This integration enables users to make real-time predictions, such as predicting average BMI of a male smoker, and provides valuable insights from their data, making it an attractive solution for developers looking to leverage machine learning capabilities without extensive coding knowledge.

The text discusses the limitations of traditional data warehouses, which are often plagued by four bottlenecks: limited support for streaming ingest, ETL batch windows, query latencies, and concurrency limitations. SingleStore, a modern data warehouse solution, offers a bold opportunity to achieve 100x performance improvements over traditional data warehouses, running data-intensive analytic workloads, and achieving significant cost reductions. The article highlights three companies that have successfully augmented their data warehouses with SingleStore, including a mobile phone manufacturer, a cybersecurity organization, and a media company. These customers have achieved dramatic improvements in performance, reducing latency and increasing concurrency, while also improving customer experience and bottom-line results. The text concludes by emphasizing the importance of considering data warehouse augmentation as an option for organizations struggling with outdated architectures or slow batch processes.

Kubernetes was originally designed as a stateless platform to manage containers running computation-intensive tasks and web servers. However, it can be extended to support stateful workloads like data tables and e-commerce sessions. Rob Richardson's recent DZone webinar series demonstrates how to build stateful workloads in Kubernetes. The webinar covers the basics of Kubernetes 101, file storage, configuration, secrets, data stores, and singleton services. It also explains how to get all stateful resources using a custom command that combines `kubectl api-resources`, awk, grep, xargs, and sed.

Data infrastructure complexity is rampant in the industry, with modern applications built on top of multiple special-purpose databases. This approach, dubbed "data infrastructure sprawl," has been criticized for its drawbacks, including a steeper learning curve, more points of failure, and increased business costs. However, SingleStore claims to offer a better alternative by providing an all-in-one relational database that can support both transactional and analytical queries at any scale, leveraging in-memory and on-disk storage capabilities. With SingleStore, companies can rationalize their architecture, save money, and ensure that their application won't find a bottleneck at the database layer, making it a popular choice among startups, unicorns, and Fortune 10 companies alike. By using SingleStore, developers can simplify their development velocity, reduce technical debt, and improve overall performance, while also reducing the complexity and costs associated with managing multiple databases.

As Head of Community at SingleStore, my goal is to prioritize developers' unique needs and create a win-win situation for both SingleStore and the developers who use our platform. To achieve this, SingleStore is introducing the SingleStore Ambassador Program, which aims to build a network of users eager to share their experiences with other developers through speaking, blogging, and creating videos. The program will also provide an open channel for feedback on how SingleStore can best support its ambassadors, allowing them to share knowledge and build community. This initiative will allow SingleStore to tap into the diverse range of industries and experience levels represented by its user base, enabling the company to better understand what developers find most valuable about using their platform.