incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rob Vesse <>
Subject Re: [DISCUSS] Rya Incubator Proposal
Date Mon, 07 Sep 2015 09:29:25 GMT
Interesting proposal

As someone already involved with other open source RDF projects both
inside and outside Apache it would be nice to add to the Relationships
with Other Apache Products a bit about what (if anything) you expect the
relationship to other RDF related Apache projects (Jena, Clerezza,
Stanbol, Marmotta, Commons RDF (incubating)) to be?

Apache is about community over code and so never mandates any particular
technical choices (that is always up to the individual communities) but it
would be useful to understand if you see any overlap with existing
projects or any collaboration oppurtunities.  The latter are particular
interesting because one way you can help grow a new community is by
attracting interested users in pre-existing communities who want to work
on the specific problems you are aiming to tackle where their existing
options don't address the problems while your approach does.

It would also be nice to see some discussion in that section about things
like versioning of your major dependencies.  In particular you build on
Accumulo so do you require specific version(s) thereof (since they appear
to maintain 3 release lines currently) or simply require a version with a
specific subset of Accumulo functionality?  How (if at all) does this
translate into risks in terms of adoption, community traction etc e.g.
what happens if you rely on version X and the Accumulo community abandons
that in favour of version Y or if you rely on a specific experimental
feature that never makes it into Accumulo releases?

Also it would be nice if the external dependencies section properly linked
to relevant web pages as right now it has several dead links and in some
cases outdated naming.  For example by Open RDF I assume you mean OpenRDF
Sesame which now lives at (though I'll admit to not
understanding what I'm supposed to call it anymore either!)

Similarly the documentation section mentions papers but doesn't provide
links, while both can be found online easily enough it would be nice to
add the links in



On 03/09/2015 14:03, "Adina Crainiceanu" <> wrote:

>We would like to start a discussion on accepting Rya, a scalable RDF data
>management system built on top of Accumulo. into Apache Incubator.
>The proposal is available online at
> and also at the end of this
>We are looking for additional mentors to help us with the project. Any
>advice and help will be appreciated.
>Thank you very much,
>= Rya Proposal =
>== Abstract ==
>Rya (pronounced "ree-uh" /rēə/) is a cloud-based RDF triple store that
>supports SPARQL queries.
>== Proposal ==
>Rya is a scalable RDF data management system built on top of Accumulo. Rya
>uses novel storage methods, indexing schemes, and query processing
>techniques that scale to billions of triples across multiple nodes. Rya
>provides fast and easy access to the data through SPARQL, a conventional
>query mechanism for RDF data.
>== Background ==
>RDF is a World Wide Web Consortium (W3C) standard used in describing
>resources on the Web. The smallest data unit is a triple consisting of
>subject, predicate, and object. Using this framework, it is very easy to
>describe any resource, not just Web related. For example, if you want to
>say that Alice is a professor, you can represent this as an RDF triple
>(Alice, rdf:type, Professor). In general, RDF is an open world framework
>that allows anyone to make any statement about any resource, which makes
> a popular choice for expressing a large variety of data.
>RDF is used in conjunction with the Web Ontology Language (OWL). OWL is a
>framework for describing models or ontologies for RDF. It defines
>relationships, and/or structure of RDF documents. These models can be used
>to 'reason/infer' information about entities within a given domain. For
>example, you can express that a Professor is a sub class of Faculty,
>(Professor, rdfs:subClassOf, Faculty) and knowing that (Alice, rdf:type,
>Professor), it can be inferred that (Alice, rdf:type, Faculty).
>SPARQL is an RDF query language. Similar with SQL, SPARQL has SELECT and
>WHERE clauses; however, it is based on querying and retrieving RDF
>Work on Rya, a large scale distributed system for  storing and querying
>data, started in 2010.
>== Rationale ==
>With the increase in data size, there is a need for scalable systems for
>storing and retrieving RDF data in a cluster of nodes. We believe that Rya
>can fulfil that role. We expect that communities within government, health
>care, finance, and others who generate large amounts of RDF data will be
>most interested in this project.
>From its inception, the project operated with an Apache-style license, but
>it was open to mostly US government-related projects only. We believe that
>having the project and the development open for all will benefit both the
>project and the interested communities.
>== Current Status ==
>The project source code and documentation are currently hosted in a
>repository on Github. New users are added to the repository upon request.
>=== Meritocracy ===
>Meritocracy is the model that we currently follow, and we want to build a
>larger and more diverse developer community by becoming an Apache project.
>=== Community ===
>Rya has being building a community of users and developers for the past 3
>years. There is currently an active workgroup with monthly meetings and
>number of participants in the meeting is increasing.
>=== Core Developers ===
>The core developers are a diverse group of people who are either
>employees or former / current government contractors from different
>=== Alignment ===
>Rya is built on top of Accumulo, an Apache project.
>== Known Risks ==
>=== Orphaned Products ===
>There is a very small risk of becoming orphaned. The current contributors
>are strongly committed to the project, there is a large enough number of
>developers interested in contributing to the project, and we believe that
>the support for the project will continue to grow from the interested
>=== Inexperience with Open Source ===
>The initial committers have various degrees of experience with open source
>projects - from very new to experienced. This project was open source
>within government from the beginning. We do not expect to have
>in operating under Apache's development process.
>=== Homogenous Developers ===
>The current list of developers form a heterogeneous group, with people for
>academia, government, and industry, collaborating from distributed
>geographic locations. We aim to expand the list of contributors with the
>help of the Apache incubation process.
>=== Reliance on Salaried Developers ===
>Many but not all of the developers working on the project are salaried
>employees, paid to work on this project. They will continue to contribute
>to the open source project. Some of the initial committers continued as
>volunteers even if no longer employed to work on this project and they
>to continue supporting the project.
>=== Relationships with Other Apache Products ===
>Rya uses Apache Accumulo, Hadoop, Zookeeper, Maven.
>=== Apache Brand ===
>Rya has generated interest in the government. It also generated interest
>within academia and industry. We believe that everyone could benefit from
>having Rya as an open source project. Due to its strong ties to Accumulo,
>an Apache project, and due to the values of the Apache Foundation, we
>believe that Apache incubator is the right place for Rya.
>== Documentation ==
>Two peer-reviewed publications [1,2] about Rya were published in 2012 and
>2015. More documentation is available in the code.
>[1] Roshan Punnoose, Adina Crainiceanu, David Rapp. Rya: A Scalable RDF
>Triple Store for the Clouds. Proceedings of the 1st International Workshop
>on Cloud Intelligence, Pages 4:1-4:8, August 2012
>[2] Roshan Punnoose, Adina Crainiceanu, David Rapp. SPARQL in the Clouds
>Using Rya. Information Systems, Volume 48, Pages 181-195, March 2015
>(Available online 23 July 2013)
>== Initial Source ==
>The code is currently available in a private Github repository.
>== Source and Intellectual Property Submission Plan ==
>The source code has been released under the Apache License, Version 2.
>Software grant, and CCLAs have been submitted. ICLAs for initial
>have been submitted or are in progress.
>== External Dependencies ==
> * Open RDF (BSD license)
> * GeoMesa (Apache License, Version 2.0)
> * Accumulo (Apache License, Version 2.0)
> * Hadoop (Apache License, Version 2.0)
> * TinkerPop (Apache License, Version 2.0)
> * IndexingSail (Apache License, Version 2.0)
>== Cryptography ==
>The proposal does not involve any cryptographic code.
>== Required Resources ==
>=== Mailing lists ===
> *
> *
> *
>=== Git Repository ===
>=== Issue Tracking ===
>== Initial Committers ==
> * Roshan Punnoose, roshanp at gmail dot com
> * David Rapp, dnrapp at ncsu dot edu
> * Adina Crainiceanu, adinancr at gmail dot com
> * Aaron Mihalik, aaron.mihalik at gmail dot com
> * Puja Valiyil, pujav65 at gmail dot com
> * Jennifer Brown, jennifer.brown at parsons dot com
> * Steve Wagner, steve.r.wagner at gmail dot com
>== Affiliations ==
> * Roshan Punnoose, Enlighten IT Consulting
> * David Rapp, North Carolina State University
> * Adina Crainiceanu, US Naval Academy
> * Aaron Mihalik, Parsons
> * Puja Valiyil, Parsons
> * Jennifer Brown, Parsons
> * Steve Wagner, Enlighten IT Consulting
>== Sponsors ==
>=== Champion ===
>Adam Fuchs, ASF Member, afuchs at apache dot org
>=== Nominated Mentors ===
>Josh Elser josh dot elser at gmail dot com
>We are seeking additional mentors
>=== Sponsoring Entity ===
>Apache Incubator

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message