Supporting Evacuation Missions...
Transcript of Supporting Evacuation Missions...
![Page 1: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/1.jpg)
U
![Page 2: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/2.jpg)
Supporting Evacuation Missions withOntology-based SPARQL Federation
Audun Stolpe and Jonas HalvorsenSTIDS 12-15 November 2013
![Page 3: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/3.jpg)
Table of contents
1 The Nato Network Enabled Capability (NNEC) concept
2 Example: Evacuation planning
3 System overview
4 Source Selection
5 Reducing the size of the cropping
![Page 4: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/4.jpg)
The NNEC Concept
• Official policy for Command and Control in NATO forces• Goal: Achieve better mission effectiveness• Builds upon extensive information sharing• Shift from “Need to know” to “Responsibiliy to share”
![Page 5: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/5.jpg)
NNEC objectives
• Some technical aims:
• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:
• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 6: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/6.jpg)
NNEC objectives
• Some technical aims:
• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:
• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 7: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/7.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,
• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:
• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 8: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/8.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,• provide a robust scheme for information integration,
• standardise exchange and storage formats
• Some strategic aims:
• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 9: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/9.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:
• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 10: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/10.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:
• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 11: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/11.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:• create a high degree of shared situational awareness,
• based on timely data,• to support decision making during operations
![Page 12: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/12.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:• create a high degree of shared situational awareness,• based on timely data,
• to support decision making during operations
![Page 13: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/13.jpg)
NNEC objectives
• Some technical aims:• support extensive information sharing,• provide a robust scheme for information integration,• standardise exchange and storage formats
• Some strategic aims:• create a high degree of shared situational awareness,• based on timely data,• to support decision making during operations
![Page 14: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/14.jpg)
Presuppositions
• NNEC assumptions:
• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 15: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/15.jpg)
Presuppositions
• NNEC assumptions:
• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 16: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/16.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio
• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 17: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/17.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues
• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 18: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/18.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems
• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 19: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/19.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members
• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 20: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/20.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear
• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 21: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/21.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 22: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/22.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:
• data models are standardised• HTTP can be assumed
![Page 23: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/23.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:• data models are standardised
• HTTP can be assumed
![Page 24: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/24.jpg)
Presuppositions
• NNEC assumptions:• fragile ICT infrastructure, typically IP radio• low bandwidth + latency issues• information is typically distributed across systems• systems are contributed by different coalition members• systems may appear and disappear• information may be mission critical
• Our assumptions:• data models are standardised• HTTP can be assumed
![Page 25: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/25.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 26: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/26.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 27: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/27.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,
2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 28: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/28.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests
3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 29: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/29.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time
4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 30: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/30.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 31: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/31.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 32: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/32.jpg)
Desiderata
We want an approach that:
1 allows a user to access available sources in a unified way,2 minimises the number of required HTTP requests3 allows the relevant sources to be discovered at run-time4 guarantees the soundness/completeness of q. a.
• Points to ontology based data access + rewriting
![Page 33: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/33.jpg)
Table of contents
1 The Nato Network Enabled Capability (NNEC) concept
2 Example: Evacuation planning
3 System overview
4 Source Selection
5 Reducing the size of the cropping
![Page 34: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/34.jpg)
Case description
• JOC — Joint Operations Center
• analyst is monitoring medical evacuation missions• particularly missions that are threatened by enemy activity• information need expressed by the query
“Find all medical evacuation missions and friendly units such that a) themission can be classified as being threatened; and b) that the friendlyunit can handle the specific type of threat that the enemy poses.”
![Page 35: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/35.jpg)
Case description
• JOC — Joint Operations Center• analyst is monitoring medical evacuation missions
• particularly missions that are threatened by enemy activity• information need expressed by the query
“Find all medical evacuation missions and friendly units such that a) themission can be classified as being threatened; and b) that the friendlyunit can handle the specific type of threat that the enemy poses.”
![Page 36: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/36.jpg)
Case description
• JOC — Joint Operations Center• analyst is monitoring medical evacuation missions• particularly missions that are threatened by enemy activity
• information need expressed by the query
“Find all medical evacuation missions and friendly units such that a) themission can be classified as being threatened; and b) that the friendlyunit can handle the specific type of threat that the enemy poses.”
![Page 37: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/37.jpg)
Case description
• JOC — Joint Operations Center• analyst is monitoring medical evacuation missions• particularly missions that are threatened by enemy activity• information need expressed by the query
“Find all medical evacuation missions and friendly units such that a) themission can be classified as being threatened; and b) that the friendlyunit can handle the specific type of threat that the enemy poses.”
![Page 38: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/38.jpg)
Case description
• JOC — Joint Operations Center• analyst is monitoring medical evacuation missions• particularly missions that are threatened by enemy activity• information need expressed by the query
“Find all medical evacuation missions and friendly units such that a) themission can be classified as being threatened; and b) that the friendlyunit can handle the specific type of threat that the enemy poses.”
![Page 39: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/39.jpg)
Information sources
The query involves three operational information systems:
1 JOCWatch: a log of events reported from the field2 MedWatch: medical mission planning and tracking3 Track Source: friendly units: capabilities, location.
![Page 40: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/40.jpg)
Information sources
The query involves three operational information systems:
1 JOCWatch: a log of events reported from the field2 MedWatch: medical mission planning and tracking3 Track Source: friendly units: capabilities, location.
![Page 41: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/41.jpg)
Information sources
The query involves three operational information systems:
1 JOCWatch: a log of events reported from the field
2 MedWatch: medical mission planning and tracking3 Track Source: friendly units: capabilities, location.
![Page 42: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/42.jpg)
Information sources
The query involves three operational information systems:
1 JOCWatch: a log of events reported from the field2 MedWatch: medical mission planning and tracking
3 Track Source: friendly units: capabilities, location.
![Page 43: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/43.jpg)
Information sources
The query involves three operational information systems:
1 JOCWatch: a log of events reported from the field2 MedWatch: medical mission planning and tracking3 Track Source: friendly units: capabilities, location.
![Page 44: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/44.jpg)
Ontology-based data integration
![Page 45: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/45.jpg)
Conceptual relationship between data sources
jw:Hostile <event>
<instigator> <incident> jw:SAFIRE
med:Mission <mission>
<track> <unit> nf:Artil
<posdata> ‘‘49.12352’’ ‘‘51.76352’’
jw:instigator
jw:incident
rdf:type
rdf:type
rdf:type
med:incident
nf:unit rdf:type
nf:posinf
nf:lat
nf:lon
useront:canHandle
JOCWatch
MedWatch
TrackSource
![Page 46: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/46.jpg)
Rewriting: A simple example
Query: ThreatenedMission(x)?
>
Mission
Mission
Incident
Incident
HostileOp
HostileOp
ThreatenedMis
ThreatenedMis
ThreateningInc
ThreateningInc
is-a
is-a
overlaps
overlaps
partOf
partOf
Mission(x) & overlaps(x, y) & Incident(y) & partOf(y,z) &
HostileOp(z)
![Page 47: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/47.jpg)
Rewriting: A simple example
Query: ThreatenedMission(x)?
>
Mission
Mission
Incident
Incident
HostileOp
HostileOp
ThreatenedMis
ThreatenedMis
ThreateningInc
ThreateningInc
is-a
is-a
overlaps
overlaps
partOf
partOf
Mission(x) & overlaps(x, y) & Incident(y) & partOf(y,z) &
HostileOp(z)
![Page 48: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/48.jpg)
Rewriting: A simple example
Query: ThreatenedMission(x)?
>
Mission
Mission
Incident
Incident
HostileOp
HostileOp
ThreatenedMisThreatenedMis ThreateningInc
ThreateningInc
is-a
is-a
overlapsoverlaps
partOf
partOf
Mission(x) &
overlaps(x, y) &
Incident(y) & partOf(y,z) &
HostileOp(z)
![Page 49: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/49.jpg)
Rewriting: A simple example
Query: ThreatenedMission(x)?
>
MissionMission Incident
Incident
HostileOp
HostileOp
ThreatenedMis
ThreatenedMis
ThreateningIncThreateningInc
is-ais-a
overlaps
overlaps
partOf
partOf
Mission(x) & overlaps(x, y) &
Incident(y) & partOf(y,z) &
HostileOp(z)
![Page 50: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/50.jpg)
Rewriting: A simple example
Query: ThreatenedMission(x)?
>
Mission
Mission
IncidentIncident HostileOp
HostileOp
ThreatenedMis
ThreatenedMis
ThreateningInc
ThreateningInc
is-a
is-a
overlaps
overlaps
partOfpartOf
Mission(x) & overlaps(x, y) & Incident(y) & partOf(y,z) &
HostileOp(z)
![Page 51: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/51.jpg)
Rewriting: A simple example
Query: ThreatenedMission(x)?
>
Mission
Mission
Incident
Incident
HostileOpHostileOp
ThreatenedMis
ThreatenedMis
ThreateningInc
ThreateningInc
is-a
is-a
overlaps
overlaps
partOf
partOf
Mission(x) & overlaps(x, y) & Incident(y) & partOf(y,z) &
HostileOp(z)
![Page 52: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/52.jpg)
Table of contents
1 The Nato Network Enabled Capability (NNEC) concept
2 Example: Evacuation planning
3 System overview
4 Source Selection
5 Reducing the size of the cropping
![Page 53: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/53.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q ΣQ Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 54: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/54.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 55: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/55.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
RewriteRewrite
Q Σ
Q Σ
QΣQΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 56: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/56.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources
• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 57: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/57.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then
• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2R2 RnRnR1R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 58: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/58.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and
• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
DistributeDistribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 59: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/59.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources
• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
DistributeDistribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 60: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/60.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that
• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 61: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/61.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 62: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/62.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A
• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
CropCrop
Evaluate
Evaluate
Answer
Answer
. . .
![Page 63: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/63.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣQΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
EvaluateEvaluate
Answer
Answer
. . .
![Page 64: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/64.jpg)
System overview
• Input: SPARQL query Q and an ontology Σ
• Q rewritten to QΣ
• federator selects from available sources• according to signature of QΣ, then• splits Q into subqueries tailored for each source, and• distributes subqueries to the sources• results of subqueries yield snapshots, that• are jointly sufficient for answering Q
• snapshots are joined to form the cropping A• against which, QΣ is finally evaluated
Rewrite
Rewrite
Q Σ
Q Σ
QΣ
QΣ
Distribute
Distribute
R2
R2
Rn
Rn
R1
R1
Crop
Crop
Evaluate
Evaluate
AnswerAnswer
. . .
![Page 65: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/65.jpg)
Sample query
Query pattern:
?mission medics:missionType medics:Rescue.
?mission medics:jocWatchIncident ?incident.
?incident jocw:status ?stat.
Routed to MedWatch:
CONSTRUCT {
?_1 medics:missionType medics:Evac.
?_1 medics:jocwIncident ?_2.
?_3 jocw:status ?_4.
} WHERE {
{ ?_1 medics:missionType medics:Evac.
?_1 medics:jocwIncident ?_2.}
UNION
{ ?_3 jocw:status ?_4.}}
Routed to JOCWatch:
CONSTRUCT {
?_1 jocw:instigator ?_2.
?_3 jocw:status ?_4.
} WHERE {
{ ?_1 jocw:instigator ?_2.}
UNION
{ ?_3 jocw:status ?_4.}}
Each query is built as a union of exclusive/nonexclusive groups
![Page 66: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/66.jpg)
Soundness/completeness
• Note the renaming of variables in the queries• necessary for the soundness of query answering• CONSTRUCT queries are defined to adhere to a logical form• guarantees soundness/completeness in the sense:
Distributing a query Q over a set of sources R yields the exact sameanswer, as if Q were evaluated directly against a single repositorycontaining the union of R.
![Page 67: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/67.jpg)
Table of contents
1 The Nato Network Enabled Capability (NNEC) concept
2 Example: Evacuation planning
3 System overview
4 Source Selection
5 Reducing the size of the cropping
![Page 68: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/68.jpg)
Adapting to a dynamic network topology
• Important design goals:• do not hard-wire a query to a predefined set of sources• sources should be selected once per query• as sources come and go, the federator adapts
• Facilitated by FOL-rewritable ontology languages• FOL-rewritability decouples reasoning from data access• queries are precompiled (potentially cached)• selection and distribution performed afterwards• thus, network topology allowed to change
![Page 69: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/69.jpg)
Detecting available source
• Discovery mechanism
• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:
• approach addresses the NNEC needs• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 70: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/70.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,
• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:
• approach addresses the NNEC needs• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 71: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/71.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location
• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:
• approach addresses the NNEC needs• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 72: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/72.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:
• approach addresses the NNEC needs• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 73: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/73.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:
• approach addresses the NNEC needs• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 74: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/74.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:• approach addresses the NNEC needs
• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 75: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/75.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:• approach addresses the NNEC needs• it is independent of a central registry,
• eliminates the issue of network fragmentation.
![Page 76: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/76.jpg)
Detecting available source
• Discovery mechanism• mDNS for broadcasting and discovering sources,• DNS-SD content description and endpoint location• SPARQL 1.1 SDs and VoID descriptions of source content.
• Advantages:• approach addresses the NNEC needs• it is independent of a central registry,• eliminates the issue of network fragmentation.
![Page 77: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/77.jpg)
Table of contents
1 The Nato Network Enabled Capability (NNEC) concept
2 Example: Evacuation planning
3 System overview
4 Source Selection
5 Reducing the size of the cropping
![Page 78: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/78.jpg)
Shortcomings
• Problem:
• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 79: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/79.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained
• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 80: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/80.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates
• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 81: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/81.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o
• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 82: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/82.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint
• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 83: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/83.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 84: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/84.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 85: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/85.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:
• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 86: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/86.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:• assess the selectivity of triple patterns
• build the cropping incrementally• within the confines of HTTP minimality
![Page 87: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/87.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:• assess the selectivity of triple patterns• build the cropping incrementally
• within the confines of HTTP minimality
![Page 88: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/88.jpg)
Shortcomings
• Problem:• CONSTRUCT queries are too unconstrained• problem especially acute wrt. common predicates• consider e.g. ?a rdf:type ?o• this pattern will most likely be routed to every endpoint• thus, downloading potentially huge amounts of data
• Proposed solution:• assess the selectivity of triple patterns• build the cropping incrementally• within the confines of HTTP minimality
![Page 89: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/89.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns
• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 90: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/90.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 91: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/91.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results
• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 92: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/92.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺
• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 93: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/93.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups
• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 94: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/94.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 95: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/95.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o
• ?s rdf:type ?o
![Page 96: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/96.jpg)
The selectivity of triple patterns
• Heuristics based on structural variations in triple patterns• (s, p, o) ≺ (s, ?, o) ≺ (?, p, o) ≺ . . . ≺ (?, ?, ?)
• ≺-less is less likely to produce large results• compare query patterns with the all-some lifting of ≺• nonexclusive groups more weight than exclusive groups• some patterns are preselected for the lowest level, e.g.
• ?s ?p ?o• ?s rdf:type ?o
![Page 97: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/97.jpg)
Incremental construction of the cropping:
t31, . . . , t3m t31
, . . . , t3m t21, . . . , t2l
t21, . . . , t1l t21
, . . . , t2l t11, . . . , t1k
t11, . . . , t1k t11
, . . . , t1k
R1 . . . Rn
VALUES
VALUESSELECT
SELECTCONSTRUCT
CONSTRUCT
CONSTRUCT
Priority levels
Execution process
![Page 98: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/98.jpg)
Summary of key properties
• federated query answering is provably sound/complete• only one HTTP request is sent to each source• query answering is tractable, i.e. takes only polynomially
many computation steps• source selection is dynamic and once-per-query
![Page 99: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/99.jpg)
Current and future work
• Status:
• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:
• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data• design a general purpose system around the current core
![Page 100: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/100.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization
• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:
• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data• design a general purpose system around the current core
![Page 101: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/101.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure
• a demo that runs against real military databases
• Prioritized future work:
• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data• design a general purpose system around the current core
![Page 102: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/102.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:
• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data• design a general purpose system around the current core
![Page 103: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/103.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:
• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data• design a general purpose system around the current core
![Page 104: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/104.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:• adapt the theory and the implementation to data streams
• address challenges related to lifting of sensor data• design a general purpose system around the current core
![Page 105: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/105.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data
• design a general purpose system around the current core
![Page 106: Supporting Evacuation Missions withstids.c4i.gmu.edu/papers/STIDSPresentations/STIDS2013_talk_T19_StolpeEtAl.pdfDesiderata We want an approach that: 1 allows a user to access available](https://reader034.fdocuments.net/reader034/viewer/2022050415/5f8b34276ab50e65ce496621/html5/thumbnails/106.jpg)
Current and future work
• Status:• a formal theory of federation + join-order optimization• a working implementation in Scala/Clojure• a demo that runs against real military databases
• Prioritized future work:• adapt the theory and the implementation to data streams• address challenges related to lifting of sensor data• design a general purpose system around the current core