July 2021 Summaries

Mislav Vuletic's blog post explores interpersonal relationships in the Bible using a graph database to analyze a dataset compiled by Brady Stephenson, which details biblical figures and their connections. The dataset, while incomplete up to the second book of Chronicles, provides insights into familial and tribal relationships, and Memgraph is used to answer questions such as the longest patrilineal descent from Adam and the family tree of Adam excluding Israel (Jacob). The study reveals the size of the Twelve Tribes of Israel through bloodlines and marriage references, highlighting Judah and Levi as the largest tribes. The article also provides a step-by-step guide for readers interested in exploring their datasets, including data preparation and integrity testing, using Cypher queries to model relationships within Memgraph. Despite some inconsistencies in the dataset, such as duplicate entries and relationship errors, the exploration serves as an example of how graph databases can be used to derive meaningful insights from complex historical texts.

In a detailed exploration of debugging within the Memgraph system, the author recounts a journey of tackling overflow issues related to query timeouts and stack overflows. Initially, the team encountered integer overflow problems with their timeout mechanism, which was linked to the TSC-based timer that proved unreliable on certain CPUs, specifically AMD Ryzen 7. This led them to adopt POSIX timers, which provided a more stable solution without performance penalties. Concurrently, the integration of the jemalloc library presented new challenges, as its inclusion led to segmentation faults due to stack overflows, partly because of the minimal stack size used by helper threads created with the SIGEV_THREAD flag in POSIX timers. The solution involved switching to SIGEV_THREAD_ID, which allows directing signals to specific threads, thereby avoiding stack overflow issues and maintaining compatibility with jemalloc. This debugging process, while complex, is highlighted as an opportunity for deeper understanding and learning within programming and systems engineering.

Memgraph 1.6 introduces significant advancements, particularly in data streaming capabilities by integrating with Kafka, enabling the execution of openCypher queries based on Kafka messages. This version also enhances transaction management with new isolation levels, READ COMMITTED and READ UNCOMMITTED, allowing more flexible control over data transactions. Additional updates include the ability to manage streaming connections, disable the LOAD CSV clause via a configuration flag, generate snapshots through queries, and address random query timeouts on specific hardware. The release is accompanied by refreshed and improved documentation, providing comprehensive guidance on these new features and encouraging users to explore and report any issues via their Discord server.

Memgraph recently held its first hackathon to gain insight into the user experience of their products, as many of their engineers had not previously used Memgraph for personal projects. The event allowed participants to step into the roles of users and fostered collaboration by forming diverse teams through random pairing, with the goal of creating innovative app examples utilizing Memgraph. Participants were provided with data sets but had the freedom to choose their projects, resulting in a range of applications such as sentiment analysis on Reddit comments, a Slack bot, a user insights tool, a music recommendation engine, and a railway system monitoring app. The event culminated in presentations and awards across five categories, with the team "405 Found" winning the best overall app for their streaming application. The success of the hackathon has encouraged Memgraph to plan future events, and the projects will soon be available on their website for other developers to explore and build upon.

Memgraph's 2020 summer internship was a fully remote program designed to engage students in hands-on projects that aligned with the company's ongoing engineering initiatives. Despite the challenges of transitioning to a remote format due to COVID-19, eight selected interns worked on various impactful projects, such as developing a Rust library for the Bolt protocol, enhancing the Memgraph Lab UI, and applying graph algorithms, under the guidance of experienced mentors. The program emphasized practical experience, beginning with workshops and leading to independent project work, supplemented by regular check-ins and support. The internship concluded with a presentation ceremony where students showcased their accomplishments, received certificates, and celebrated their achievements. As a testament to the program's success, several interns were offered permanent positions at Memgraph, and preparations for the next internship edition were already underway.

Memgraph Cloud 2.0 is announced as a significant upgrade with enhanced developer experience and new features based on user feedback, aimed at simplifying the deployment and management of graph databases. The cloud platform now supports a wider range of instance sizes, starting from 1 GB RAM, making it more accessible and affordable, with prices starting at approximately $16 per month. A new pause functionality allows users to significantly reduce operating costs by paying only for storage when their instance is not in use. Memgraph Cloud also offers seamless integration with popular programming languages through detailed connection instructions and Memgraph Lab, a visual client available both in-browser and as a desktop application. Additionally, the platform provides snapshot features for data backup and instance upgrades, ensuring secure data management. Users can experience these features without upfront payment, thanks to a 14-day free trial that offers a 2GB RAM project for free, encouraging exploration of graph analytics.