The cloud computing model changes many aspects of platform expectations due to resource sharing. Some of the traditional methods of application development on distributed systems can get distorted because of this. Also, the cloud model gains traction due the feasibility of workload consolidation. However, the consolidation methods can co-host divergent applications that can adversely get affected by generic resource provisioning models. This work explores how to use application specific resource provisioning on cloud platforms. The basic intent is to understand if common provisioning stubs can be evolved that can support variety of applications without effecting their performance expectations.
One such initiative was exploring Cloud-IoT frameworks. Real-time events associated with human activities is catching rapid pace due to feature and function rich mobiles like the smartphones. These devices not just facilitate users to seek access to information derived from large data sources, but also enable them to add data as sources to equally large amount of data repository. One such interesting and widely used application is a location based navigation service. This application derives its utility based on large amount of geo-spatial data repository overlaid with current events happening in the neighborhood of user interest. However, because of amount of data that needs to be handled for processing, most of the times these applications process either static data or use trends based on historical information. What really increases the utility of these locations based services is modulation to real-time changing events. This work identifies three key characteristics that need to be addressed to fulfill this gap, namely timeliness in responding to user queries based on changing events, data repository to support such functionality while preserving the desired user privacy. Here we present the three facets to a navigator application based on the above three characteristics and demonstrate how its utility can be improved using compute clouds.
- J.Lakshmi, S. K. Nandy, Aakriti Gupta, Pavan Kumar Akulakrishna, Vaibhav Ankush Kachore, Cloud Framework for IoT associated GPS location based applications, Technical Report, Cloud Systems Lab, SERC, IISc, Bangalore-12, India, Oct. 2015.(PDF)
Complete details of the efforts in this direction are listed here:
- Application scheduling frameworks for supporting real-time response
A.Gupta, L. J. and S. K. Nandy, “Real Time Routing in Road Networks,” 2014 IEEE Fourth International Conference on Big Data and Cloud Computing, Sydney, NSW, 2014, pp. 9-16. doi: 10.1109/BDCloud.2014.85 (PDF)(Talk)
Real Time Routing in Road Networks, Aakriti Gupta, MTech (Computational Science) Project Report, June 2014 (PDF)
- Storage Organization for delivering application performance
Pavan Kumar Akulakrishna, Lakshmi J, and Nandy S. K. 2014. Efficient Storage of Big-Data for Real-Time GPS Applications. In Proceedings of the 2014 IEEE Fourth International Conference on Big Data and Cloud Computing (BDCLOUD ’14). IEEE Computer Society, Washington, DC, USA, 1-8. DOI=http://dx.doi.org/10.1109/BDCloud.2014.49 (PDF)(Talk)
Efficient Storage of Big-Data for GPS Applications, Pavan Kumar Akulakrishna, MTech (Computational Science) Project Report, June 2014 (PDF)
- Application specific security constructs on Cloud setups
V. A. Kachore, J. Lakshmi and S. K. Nandy, “Location Obfuscation for Location Data Privacy,” 2015 IEEE World Congress on Services, New York City, NY, 2015, pp. 213-220. doi: 10.1109/SERVICES.2015.39 (PDF) (Talk)
Location Obfuscation for Privacy in Navigation Applications, Vaibhav Ankush Kachore, MTech(Computational Science) Project Report, June 2015 (PDF)