DRAS-TIC Linked Data: Evenly Distributing the Past
Memory institutions must be able to grow a fully-functional repository incrementally as collections grow, without expensive enterprise storage, massive data migrations, and the performance limits that stem from the vertical storage strategies. The Digital Repository at Scale that Invites Computation...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-07-01
|
| Series: | Publications |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2304-6775/7/3/50 |
| _version_ | 1798038289868914688 |
|---|---|
| author | Gregory Jansen Aaron Coburn Adam Soroka Will Thomas Richard Marciano |
| author_facet | Gregory Jansen Aaron Coburn Adam Soroka Will Thomas Richard Marciano |
| author_sort | Gregory Jansen |
| collection | DOAJ |
| description | Memory institutions must be able to grow a fully-functional repository incrementally as collections grow, without expensive enterprise storage, massive data migrations, and the performance limits that stem from the vertical storage strategies. The Digital Repository at Scale that Invites Computation (DRAS-TIC) Fedora research project, funded by a two-year National Digital Platform grant from the Institute for Museum and Library Services (IMLS), is producing open-source software, tested cluster configurations, documentation, and best-practice guides that enable institutions to manage linked data repositories with petabyte-scale collections reliably. DRAS-TIC is a research initiative at the University of Maryland (UMD). The first DRAS-TIC repository system, named Indigo, was developed in 2015 and 2016 through a collaboration between U.K.-based storage company, Archive Analytics Ltd., and the UMD iSchool Digital Curation Innovation Center (DCIC), through funding from an NSF DIBBs (Data Infrastructure Building Blocks) grant (NCSA “Brown Dog”). DRAS-TIC Indigo leverages industry standard distributed database technology, in the form of Apache Cassandra, to provide open-ended scaling of repository storage without performance degradation. With the DRAS-TIC Fedora initiative, we make use of the Trellis Linked Data Platform (LDP), developed by Aaron Coburn at Amherst College, to add the LDP API over similar Apache Cassandra storage. This paper will explain our partner use cases, explore the system components, and showcase our performance-oriented approach, with the most emphasis given to performance measures available through the analytical dashboard on our testbed website. |
| first_indexed | 2024-04-11T21:38:10Z |
| format | Article |
| id | doaj.art-098e241f89b644fba07c508e9ea8cbc3 |
| institution | Directory Open Access Journal |
| issn | 2304-6775 |
| language | English |
| last_indexed | 2024-04-11T21:38:10Z |
| publishDate | 2019-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Publications |
| spelling | doaj.art-098e241f89b644fba07c508e9ea8cbc32022-12-22T04:01:40ZengMDPI AGPublications2304-67752019-07-01735010.3390/publications7030050publications7030050DRAS-TIC Linked Data: Evenly Distributing the PastGregory Jansen0Aaron Coburn1Adam Soroka2Will Thomas3Richard Marciano4School of Information Studies, University of Maryland, College Park, MD 20742, USAInformation Technology Services, Amherst College, Amherst, MA 01002, USAOffice of the CIO, The Smithsonian Institution, Washington, DC 20002, USASchool of Information Studies, University of Maryland, College Park, MD 20742, USASchool of Information Studies, University of Maryland, College Park, MD 20742, USAMemory institutions must be able to grow a fully-functional repository incrementally as collections grow, without expensive enterprise storage, massive data migrations, and the performance limits that stem from the vertical storage strategies. The Digital Repository at Scale that Invites Computation (DRAS-TIC) Fedora research project, funded by a two-year National Digital Platform grant from the Institute for Museum and Library Services (IMLS), is producing open-source software, tested cluster configurations, documentation, and best-practice guides that enable institutions to manage linked data repositories with petabyte-scale collections reliably. DRAS-TIC is a research initiative at the University of Maryland (UMD). The first DRAS-TIC repository system, named Indigo, was developed in 2015 and 2016 through a collaboration between U.K.-based storage company, Archive Analytics Ltd., and the UMD iSchool Digital Curation Innovation Center (DCIC), through funding from an NSF DIBBs (Data Infrastructure Building Blocks) grant (NCSA “Brown Dog”). DRAS-TIC Indigo leverages industry standard distributed database technology, in the form of Apache Cassandra, to provide open-ended scaling of repository storage without performance degradation. With the DRAS-TIC Fedora initiative, we make use of the Trellis Linked Data Platform (LDP), developed by Aaron Coburn at Amherst College, to add the LDP API over similar Apache Cassandra storage. This paper will explain our partner use cases, explore the system components, and showcase our performance-oriented approach, with the most emphasis given to performance measures available through the analytical dashboard on our testbed website.https://www.mdpi.com/2304-6775/7/3/50distributed databaselinked data platformFedora Commons repositoryhorizontal scaling |
| spellingShingle | Gregory Jansen Aaron Coburn Adam Soroka Will Thomas Richard Marciano DRAS-TIC Linked Data: Evenly Distributing the Past Publications distributed database linked data platform Fedora Commons repository horizontal scaling |
| title | DRAS-TIC Linked Data: Evenly Distributing the Past |
| title_full | DRAS-TIC Linked Data: Evenly Distributing the Past |
| title_fullStr | DRAS-TIC Linked Data: Evenly Distributing the Past |
| title_full_unstemmed | DRAS-TIC Linked Data: Evenly Distributing the Past |
| title_short | DRAS-TIC Linked Data: Evenly Distributing the Past |
| title_sort | dras tic linked data evenly distributing the past |
| topic | distributed database linked data platform Fedora Commons repository horizontal scaling |
| url | https://www.mdpi.com/2304-6775/7/3/50 |
| work_keys_str_mv | AT gregoryjansen drasticlinkeddataevenlydistributingthepast AT aaroncoburn drasticlinkeddataevenlydistributingthepast AT adamsoroka drasticlinkeddataevenlydistributingthepast AT willthomas drasticlinkeddataevenlydistributingthepast AT richardmarciano drasticlinkeddataevenlydistributingthepast |