NebulaGraph Year in Review 2023 Part 1: Enhancing Stability, Speed, and Usability

As we bid farewell to another year filled with its own unique set of triumphs and trials, it's time to pause, rewind, and reflect on the journey we've navigated together. This blog post is dedicated to a comprehensive recap of the past 12 months, a year that was anything but ordinary. From the smallest improvements to the most significant milestones, we'll be revisiting the moments that made us grow. So join us as we journey back through the year in review.

Product Refined for Upgraded Stability and Performance

Throughout the year, NebulaGraph prioritized stability as its primary focus. We introduced the innovative feature, Memory Tracker, to implement fine-grained memory usage control in order to improve NebulaGraph’s kernel stability. Additionally, we undertook several rounds of comprehensive testing, incorporating tools and methods like fuzzer to give NebulaGraph a thorough "health check" to make sure its development meets our expectations. Overall, the latest release, v3.6.0, witnessed stability improvement and performance enhancements in common graph-based queries, such as path and multi-hop queries, with the highest performance improvement reaching six-fold. More details of the latest version are available in Enterprise v3.6.0 Introduction.

In addition to a more stable kernel, we have also enhanced NebulaGraph’s cloud native infra to better serve enterprise customers deploying NebulaGraph on Kubernetes. Check out some of the highlighted features introduced in 2023:

• Provided the NebulaAutoscaler CRD to implement Horizontal Pod Autoscaling (HPA) for the Graph service.
• K8s scheduled backup, and restore tasks could be done within the Operator with CRD
• Supported zone management, a logical rack used to group storage Pods within a cluster, to improve the resilience of the cluster by ensuring an even distribution of data replicas across each zone. Scheduler plugins are provided to maintain an even distribution after scaling in.
• mLTS are supported to ensure secure communication between clients and Nebula Pods.
• Support for extended configurations such as init-container, sidecar-container, etc., which can be used in scenarios like log collection, service-mesh, and more.

Adding LLM related functions to NebulaGraph tools

Several NebulaGraph tools underwent significant functional iterations in 2023, and below are some important updates you should not miss.

• Nebula Studio
  • extended to LLM related functions, such as AI assistant that converts natural language into query statements and KG builder that can realize automatically importing data into the NebulaGraph database. Based on the mentioned functions, more large model application scenarios can be derived.
  • supported displaying schema creation statements for graph space in version v3.4.0, including graph spaces, tags, edge types, and indexes.
  • adapted to the ARM architecture in version v3.7.0 and optimized the visualization of execution plans.
• Nebula Importer was refactored in version v4.0.0, significantly improving import performance and readability of import configuration files.
• Nebula Exchange supported reading HDFS configuration files in version v3.4.0, preprocessing data sources in v3.5.0, batch deletion and updates in v3.6.0, and obtaining template configuration files based on data sources in the latest v3.7.0.
• Clients, including nebula-go, nebula-java and nebula-console, have supported HTTP/2 since version v.3.6.0.

Throughout 2023, NebulaGraph has been proved to be excelling in scalability, stability and performance for enterprise customers. Industries, such as finance, social media, and e-commerce, adopted NebulaGraph due to its ability to effectively handle complex data relationships, providing insightful analysis for better decision-making. The past year has witnessed NebulaGraph's commitment to providing superior solutions and services.

As we step into 2023, we are filled with optimism and excitement for the opportunities it brings. NebulaGraph remains committed to delivering advanced, reliable and scalable graph database solutions, striving to empower our enterprise customers with the best technology to drive their business forward.