December 2018 Summaries

Vespa's December 2018 product updates introduce several key features aimed at enhancing performance and functionality. A notable improvement in Streaming Search reduces latency and increases throughput by limiting data scanning using document types, optimizing query efficiency without reverse indexing. Vespa now supports ONNX model integration, allowing for the import and transformation of AI models into Tensors for ranking, complementing its earlier TensorFlow capabilities. Additionally, Vespa has optimized transaction log pruning with the prepareRestart feature, enhancing node restart efficiency crucial for continuous integration and deployment environments. Lastly, the new grouping functionality for map attribute fields facilitates complex data structuring and faceting, particularly beneficial for applications with structured data such as e-commerce. These updates underscore Vespa's commitment to advancing AI and search capabilities while encouraging community feedback and contributions.