Apache Rya – A next generation database for a new generation of drones
UNITED STATES NAVAL ACADEMY, Annapolis, MD. – Apache Rya (incubating) is an open source database management system. Rya is one of the most advanced database projects underway in the Department of the Navy today. The latest naval application of Rya was recently highlighted in a CBS 60 Minutes episode on the new generation of drones. With its ability to store billions of linked information sets and return answers to most computer-based questions in under a second, Rya demonstrated an ability to support advanced tactical communications between Marine Captain Jim Pineiro and two autonomous robots and Navy Lt. Cdr. Rollie Wicks and a missile boat to accomplish a mission through manned-unmanned teaming. The joint manned-unmanned team accomplished portions of the mock naval Reconnaissance, Surveillance, and Target Acquisition (RSTA) mission in multiple orders of magnitude faster with artificial intelligence and Rya than they would have without it. Efforts are underway to transition Rya as an embedded technology that can support autonomous swarms of smaller robots.
The Rya technology will help humans to communicate with robots and artificial intelligence in a more natural way. Rya is currently being applied in a project to allow Marines to communicate verbal tactical orders to a robot operating under battlefield conditions. Rya was designed to store and search for information represented in a format close to human language. Historically, databases have been a collection of tables containing data with no semantic information and limited relationships. The explosion in size and complexity of information has led to a shift in what and how information needs to be stored. This shift represents a transition from communications of unrelated information components to semantically related expressions. This closely mimics the human language, which is structured as “triples” or the combination of three components consisting of a subject, predicate, and object. For example, to issue the command “Robot 1, map area Delta”, one could write the triple “<Robot 1, map, Area Delta>”, where “Robot 1” is the subject, “map” is the predicate, and “area Delta” is the object of the triple. The Resource Description Framework (RDF) is a standard for such a data representation. In its most basic form, RDF contains triple information of the form <subject, predicate, object>. Therefore, an RDF triple store is a database designed for storing and accessing complex contextual information where the basic unit of information is a “phrase” containing the three components of subject, predicate, and object.
In addition to advancing human to machine communications, Rya is being used for artificial intelligence projects involving semi-autonomous content production operations. Combined with artificial intelligence, Rya has demonstrated the ability to support the acceleration of content production operations by one to three orders of magnitude. Rya will begin supporting national agency content production operations in early 2017.
Rya started as a research project at the Laboratory for Telecommunication Sciences and continued at the United States Naval Academy (USNA). After the publication of the first peer-reviewed article about Rya , many military, government, and commercial entities became interested in using Rya for their applications. Rya has been an open source project at the Apache Software Foundation since 2015. Dr. Adina Crainiceanu, Associate Professor of Computer Science at USNA, and Jennifer Brown, program manager for semantic technologies at Parsons, are leading the research and development of Rya. As the project is now open source, contributions are received from around the world, and the pace of development is increasing.
Funding for some of the development of Rya is provided by U.S. Department of Defense agencies. This support has led to multi-agency collaboration designed to answer important research questions in the areas of databases with applications in support of robotics, artificial intelligence, and content production operations. As the Rya technology continues to mature, the size and diversity of the Rya community continues to increase.
Do you want to collaborate or use Rya on your project? Contact the Apache Rya development team at email@example.com.
1. Apache Rya (incubating). http://rya.apache.org
2. Caleb Meier, Puja Valiyil, Aaron Mihalik, Adina Crainiceanu. Rya: Optimizations to Support Real Time Graph Queries on Accumulo. Accumulo Summit, 2015
3. Caleb Meier, Puja Valiyil, David Lotts, Aaron Mihalik, Adina Crainiceanu. Rya: Accumulo Indexing Strategies for Searching Semantic Networks. Accumulo Summit, 2016
4. Roshan Punnoose, Adina Crainiceanu, David Rapp. Rya: A Scalable RDF Triple Store for the Clouds. Proceedings of the 1st International Workshop on Cloud Intelligence, August 2012
5. Roshan Punnoose, Adina Crainiceanu, David Rapp. .SPARQL in the Cloud Using Rya Information Systems, Volume 48, March 2015
6. RDF. http:/www.w3.org/RDF/
** = The opinions expressed here are solely those of the author, and do not necessarily reflect those of the Department of the Navy, Department of Defense or the United States government.