Why DEX

• DEX core technology is based on a combination of different advanced techniques to provide a high performance both in ideal situations, as well as in stress conditions where the database size is larger than the memory available.


• DEX partitions the graph in multiple light and independent data structures with a small memory footprint that are really easy to manage.


• Thanks to DEX data structures, the query engine only brings from disk to memory the part of the database that is required for the query, reducing the I/O considerably.


• Nodes and edges in DEX are represented internally by bitmaps, which are more compact and require less memory space by using compression techniques.


• Moreover, our bitmap implementation allows very fast and efficient manipulations and combination operations such as union, intersection, etc.


• Usually graph databases solutions store each node and edge in a complex object. The partitioned approach allows for the decomposition of each object in simpler elements in such a way that repeated values can be simplified to a single occurrence and its collection of references. This compression reduces again the memory requirements allowing a more efficient memory management. It improves also the memory locality of the data structures, and provides very fast query and selection operations thanks to the direct access to deduplicated values and their references.


• DEX C++ core engine avoids the overhead in execution time and memory resources, which appear in solutions implemented using interpreted languages based on virtual machines such as Java.


• DEX core engine provides a faster execution of DEX programs thanks to the code optimization provided by the compiler.


• Provides direct access to the OS buffers; no more decoding and encoding of data buffers to the high-level programming languages typical of other solutions.


• Its high performance and speed allows for the construction of applications with a high concurrency of user queries with low latencies and very fast response times, in the range of microseconds for direct calls and basic traversals.


• Low-level API written in C++ with a direct access to the core engine functionalities and data.


• DEX API is natively available for .Net, C++ and Java programming languages.