The past few years have seen a dramatic increase in the popularity and adoption of XML in Internet technology. With this revolution, an effective way for storing and querying XML documents is crucial. In this paper, we propose the INLAB2 architecture for fast native XML storage and accurate query retrieval. INLAB2 comprises of five main components, namely XML Parser, XML Encoder, XML Indexer, Data Manager and Query Processor. All these components are coupled tightly and play an important role in answering queries specified by the user. Extensive experiments quantify the practical benefits of TwigINLAB2, the main algorithm in Query Processor, which amounts to a significant performance gain compared to other approaches.