Resource-Constrained Reasoning Using a Reasoner Composition Approach

Tracking #: 545-1748

Wei Tai
John Keeney
Declan O'Sullivan1

Responsible editor: 
Guest Editors Semantic Web For All

Submission type: 
Full Paper
To increase the interoperability and accessibility of data in sensor-rich systems, there is a proliferation recently of using Semantic Web technologies within sensor-rich systems. Quite a range of such applications have emerged, such as hazard monitoring and rescue, context-aware computing, environmental monitoring, field studies, internet of things, and so on. These systems often assume a centralized paradigm for data processing, which however do not always hold in reality especially when the systems are deployed in a hostile environment. At runtime, the infrastructure of systems deployed in such an environment is also prone to interference or damage, causing part of the infrastructure to have limited network connection or even to be detached from the rest. A solution to such problem would be to push the intelligences, such as semantic reasoning, down to the device layer. A key enabling part for this solution is to run semantic reasoning on resource-constrained devices. This paper shows how reasoner composition (i.e. to automatically adjust a reasoning approach to preserve only a “well-suited” amount of reasoning for a particular ontology) can achieve resource-efficient semantic reasoning. Two novel reasoner composition algorithms are introduced and implemented. Evaluation indicates that the reasoner composition algorithms greatly reduce the resources required for OWL reasoning, facilitating greater semantic reasoning on sensor devices.
Full PDF Version: 


Solicited Reviews:
Click to Expand/Collapse
Review #1
By Aidan Hogan submitted on 10/Oct/2013
Review Comment:

The authors have done a reasonable job addressing the comments of the reviewers. I had asked for five main revisions:

1. Extend related work in cardinalities and rule-based reasoning ...

This now looks fine.

2. Add evaluation of larger ontologies ..

I still would have like to have seen LUBM used just for the intra-reasoner experiments, but understand that the authors wanted to use the same ontologies for both configurations.

3. Clarify the issue of completeness ...

The discussion added is quite informal (rather than characterise the cases, an example is given), but the issue is now at least raised, which I'm okay with.

4. Clarify issues of what is stored in memory ...

This is fine.

5. Address minor comments ...

The minor comments have been addressed, but still I found several typos and poorly worded sentences. The authors should more thoroughly proof-read the paper and should make efforts to improve readability (esp. since two are native speakers). Also, the legends in Figure 10 do not correspond with the text (S/SP) and I would recommend increasing the size of Figures 13 (a) and (b) to span two columns as they are currently too cramped to comfortable read the bars.

In any case, these remaining comments are minor.

Review #2
By Christophe Dupriez submitted on 14/Oct/2013
Review Comment:

This is a very good article explaining how rules based system may be better implemented in embedded systems. It is only bad that Oracle did not continue to improve the Sun SPOT platform: nowadays, this article would be instant worldwide success if it was based on the latest Java 8 running on a Raspberry Pi! Google Nexus Android smartphone could also be a good choice. For the next research update?
Congratulations for the work done!

Review #3
By Christophe Guéret submitted on 05/Nov/2013
Review Comment:

This submission is a revised version of the paper #446-1620. My comments on this first version were related to the readability of the paper and the potential negative impact of the proposed optimisations. I am pleased to read that both concerns have been addressed is this improved, more readable, submission.

Just one minor remark: it seems to me that in Figure 10 the legend for the cross should read "SL+TP" instead of "S+TP".