March 2018 Summaries

Algolia offers developers a unique interactive search experience with features like as-you-type search, instant faceting, mobile geo-search and on the fly spell check. Their Distributed Search Network (DSN) aims to remove network latency's impact on search speed, allowing customers to provide an instantaneous experience to end-users worldwide. Algolia is obsessed with speed as it significantly affects businesses; Amazon found that 100ms in added latency cost them 1% in sales. The company's DSN replicates indices across different regions globally, ensuring search results under 50ms in top markets and an optimal experience for all users. Currently, DSN is accessible to Starter, Growth, Pro, and Enterprise plan customers. Several services like HackerNews, TeeSpring, Product Hunt, and Zendesk leverage Algolia's DSN for faster global search.

Hacker News readers have created a search engine called hn.algolia.com to provide instant access to the entire content of Hacker News, including 1.2 million articles and 5.2 million comments. The creators are seeking feedback on how users search for news on HN and what features they would like to see improved. They have open-sourced their code and thanked Octopart and YC teams for making this experience possible.

Quadrant.io is a platform that simplifies the process of finding and charting economic data from trusted sources by offering an intuitive interface with real-time updates. It aggregates over 400,000 indicators from more than 1,000 public sources, making it easier for users to find relevant information quickly. Algolia's search solution has been instrumental in providing a seamless and efficient user experience by delivering accurate results within milliseconds and allowing searches across different data attributes. Quadrant.io's customers have praised the improved search functionality as more comfortable and intuitive, while its implementation was simple and quick, taking less than a week to complete.

At Algolia, they utilize React and Redux for building complex search UI widgets. They found that setting up conventions is crucial to harnessing the combined potential of these libraries. They use a fractal directory structure and collocate Redux action creators and reducers in the same files, following the Ducks proposal. This allows them to write more modular components with abstracted action creators. To simplify reducer-side code, they created the strongly opinionated createReducer implementation called redux-updeep, which assumes that most actions result in deep updates of the state. They've also developed a middleware called redux-magic-async-middleware to handle asynchronous actions and promises, as well as an eventual-values library to encapsulate and abstract asynchronous value behavior.

In today's complex world, operations engineers and Site Reliability Engineers (SREs) are not only concerned with the quality of their infrastructure but also with cost reduction. To achieve this, they can use various tools and strategies provided by popular cloud providers such as Amazon Web Services (AWS) and Google Cloud Platform (GCP). A key step in cost analysis is to track costs over time, including month-to-month and daily analyses, to identify trends and optimize resources. Both AWS and GCP offer billing dashboards and analytical tools, such as Cost Explorer for AWS and BigQuery for GCP, which enable users to analyze their data and set up billing alerts. Additionally, reserved instances, spot instances, and preemptible instances can help reduce costs by offering discounts on compute hours, while bare metal providers like LeaseWeb and OVH offer the best value for the buck. By choosing the right cloud provider and optimizing resources, operations engineers and SREs can make informed decisions to reduce costs and improve infrastructure efficiency.