Energy Efficient and Incremental Analytics in FAME

April 19, 2024

In the heart of the FAME integrated solution lies the Incremental Analytics component. The purpose of this component is twofold: firstly, as its name implies, to provide analytic operations in an incremental fashion. That means that the result of an analytical query (i.e. aggregation, summarization, counting, etc.) is not being calculated each time a request arrives to this component, rather than it is pre-calculated on the-fly, as new data is being ingested at the same time. This allows to provide the result of the request incrementally. The second important purpose of this component is to provide query processing in an energy efficient manner. Towards this direction, LeanXcale SL has developed new innovations under the purpose of the FAME project that have been currently merged into this component and can be exploited for any data user or ML/AI analytic processing. The major impact of these new innovations is the significant reduce of the carbon footprint associated with a particular request for query processing, as a result of the new developments that target to reduce the overall energy consumption.

The Incremental Analytics component relies on the LeanXcale database, which can be considered as the background technology that this component has been built upon. The LeanXcale database can be classified in the NewSQL family of databases that combine the benefits of two worlds: the traditional relational databases that are compliant with the SQL query language and force transactional semantics, ensuring the so called ACID properties, and the NoSQL databases that are meant to be used in cases where data is supposed to be ingested at very high-rates (from batch processing of very high volume to streaming processing of very high velocity) where database transactions are not a first-class citizen and can lead to unnecessary bottlenecks.

As LeanXcale DB can provide complementary capabilities of both SQL and NoSQL worlds, it has been used as the fundamental pillar of the Incremental Analytics component, enhanced with the innovations developed under the FAME project in order to provide both incremental and energy efficient analytics. It is important to highlight that with regards to this component, the term analytics is referred to the database analytic operations that are requested from the high-level ML/AI analytic processing.

More information: