Database Reference
In-Depth Information
One of the more important aspects of the enterprise offering is OpsCenter.
OpsCenter allows users not only to see what is going on in their system, but also
to browse the data in Cassandra and perform normal administrative operations.
Graphs can easily be created to monitor the performance of the cluster and even
individual ColumnFamilys. OpsCenter also provides an alerting system for noti-
fications of changes in your cluster. On the administration side, OpsCenter also
greatly increases the ease with which backups and restores can be completed.
Acunu
Acunu is a big-data company that offers low-latency analytics as a service. Its of-
fering of Acunu Analytics is based on Cassandra. If you remove a lot of the oper-
ational management from a big-data system, you can focus more on the core com-
petency of your business. Acunu also offers support for Cassandra and traditional
consulting.
The analytics system that Acunu has built was created to ingest high volumes
of data at a high velocity. Your data is preprocessed as it enters the system. All
filters, transformations, and cursory analytics are done here. This includes things
like rolling up cubes of sums, averages, totals, top-K, and other common column
store cubes. Pre-aggregating and grouping of queries is commonly done to ensure
that when common queries are run, the responses come back quickly. The speed
is achieved by storing the results of queries on the way in to ensure that queries
don't need to be recalculated at query time.
Acunu also offers the option of approximate aggregates. These are things like
count distinct and top-K that normally work well off of indexes in column stores
but are typically very difficult or too resource intensive to do at the big-data level.
If you are willing to accept an estimate of things like top-K and distinct counts in
your application, this is an incredibly useful feature. An example of this feature
would be something like unique visitors to a Web site or most valuable customers
in a retail application. Additionally, there is the capability of supporting
GROUP
BY
-style queries that can be used for hierarchical fields such as URLs or even
geospatial queries.
Acunu also has a few plug-and-play capabilities. Event streams such as those
from Apache Flume can easily be ingested through a RESTful API. Acunu also
plugs in very well with Storm, MQ, and many common evented frameworks. Once
the data has been brought into the system, you can take advantage of Acunu's AQL
(Acunu Query Language), an SQL-like query language, or build JavaScript pre-
