History[05][03]

Sparksee was formerly known as DEX. Development around DEX was first started by the DAMA-UPC technology transfer team from Universitat Politècnica de Catalunya (Barcelona, Spain) in 2006. Intended for large graph analysis, DEX was already used in various industrial projects and gained popularity from different areas such as medical and bibliographical network analysis. In 2008, its first version was released, and DEX officially changed its name to Sparksee on its 5th release in 2014.

Checkpoints[06]

Sparksee supports consistent checkpointing, meaning all recovery log files are flushed to disk when the checkpoint is invoked. Users can configure the checkpoint frequency.

Concurrency Control[07]

Sparksee supports a multiple read and single write model and locking is performed on the graph level. All transactions, unless explicitly initialized as write, start as read ones and are transformed into write transactions when all the other read transactions finish.

Data Model[05][07]

Sparksee represents data and schema as a graph and map data operations to graph operations as well. In the actual implementation, nodes and relations are stored as objects.

Indexes[08]

In Sparksee, the index itself stores the data instead of pointers, to ensure faster data retrieval.

Isolation Levels[07]

Sparksee supports full ACID transactions and the default isolation level is SERIALIZABLE. And the granularity is at graph level.

Query Interface[04]

Sparksee uses custom APIs for data retrieval and returns object instances as outputs.

Storage Architecture

Storage Model[07]

Sparksee has native graph storage, meaning the storage system is specifically built for graph data.

System Architecture[05]

Sparksee offers a shared-nothing architecture with master-slave replication and synchronous updates.