Spire combines the reliability and scale of HBase with the flexibility of a conventional SQL database. Every aspect of Spire is built to be distributed, so that neither the volume of data nor the number of users is a show-stopping issue. The distributed index and query system results in queries that run in milliseconds, and Hadoop integration makes it easy to process data in bulk. Spire relies on Chef-based cluster management tools to contain the complexity of operational tasks as the size of your cluster grows.
Indexing
Spire leverages the reliability and scalability of HBase. When data is written to an HBase column, it is also written to an indexing queue. This makes Spire indexing extremely fast so that new data is available immediately.
Our real, distributed indexing is integral to the platform and leads to extremely rapid queries.SQL + Search Querying
Spire combines the familiarity and flexibility of SQL with the ease of fulltext search. Unlike other Big Data systems, Spire does not require custom query algorithms or knowing how the data is distributed.
Using Spire, developers craft queries that make sense for the business problem they are solving. For example:
TOP 100 Documents GROUP BY Author, BlogText CONTAINS *laptop
Spire also integrates fulltext, multi-lingual search based on Lucene. And, Spire includes in-database MapReduce, giving developers the maximum flexibility to create any Big Data application.
Data is returned as native JSON or through JDBC, making Spire a web-ready, real-time Big Data platform.
Automated Clusters
In designing Spire, we've paid a lot of attention to the operational and administrative challenges of Big Data. Our goal has been to reduce or eliminate the complexity of managing a distributed environment. Spire uses Chef to automate cluster management. Deploying a Spire cluster is as simple as listing the machines on which to install. Because Chef tracks the state of all servers, Spire applies needed patches and upgrades with minimal downtime — often none at all.
And, Spire services are capable of rolling restarts, avoiding the risks of bringing everything down and then back up again. By automating distributed cluster management, Spire reduces the strain on scarce financial and talent resources.