Database Reference
In-Depth Information
processors to further filter events. This allows you to create custom groupings of
data or events that more closely follow your application's needs.
Since all of the data passed into Acunu is ultimately stored in its raw form after
it has been ingested and processed, additional analysis can be done on the full
data set. Acunu is built on Cassandra and can therefore easily plug into Hadoop
or many other common data warehouses. This gives you the ability to run ad hoc
analytics or even general-purpose batch analytics.
Out of the box, Acunu also gives users the ability to create dashboards made
from visualizations of the data. These visualizations come in the form of pie
charts, line graphs, histograms, and other common data visualization techniques.
In addition to these visualizations, there is also a visual query builder to aid in
constructing complex queries. Once you have created these queries, you can turn
them into reusable widgets. These can be used either on another part of the Acunu
dashboard or as JavaScript, which can then be embedded into any HTML page.
Titan by Aurelius
Titan is a distributed graph database that leverages the engineering sophistication
of Cassandra to encode and query graph structures efficiently at scale. Graphs are
composed of vertices (dots, nodes) and edges (lines, arcs) and are leveraged when
modeling domains where querying and analyzing relationships between entities or
interaction between agents is important. The structure and evolution of the inter-
actions can often provide insights that can be used to predict the future state of
the system. Moreover, graphs, and the databases that persist them, provide utility
as data management solutions where many-to-many relationships exist—a person
having many friends, a product having many features, or a hierarchy having many
branches—with cyclic, recursive data paths allowed.
Example application domains for graph databases include social networks, re-
commendation engines, biological systems, and financial transaction networks.
Titan stores graphs as a distributed adjacency list. Each row in Cassandra rep-
resents a vertex and its adjacency list; that is, a vertex's incident edges and prop-
erties. An edge is represented by an edge label (e.g., friendship), edge properties
(e.g., rating), and a reference to the row ID of the adjacent vertex pointed to by
that edge. Each edge is serialized and compressed into a column stored in the ver-
tex's row.
Traversing, the fundamental operation on a graph, is the process of moving
from vertex to vertex (i.e., row to row). To make traversing easy for developers
