Context-Aware Data Access Models for Mobile Subscribers

Dmitry Namiot - http://servletsuite.com/cv/

Contact information: Email: ; Phone: +7-495-939-23-59; Address: Russia, 119991, Moscow, Leninskie Gory, MSU, Faculty of Computational Mathematics and Cybernetics

2nd educational building, room 360

Position and Organization: senior scientist of Faculty of Computational Mathematics and Cybernetics, Open Information Systems Lab

Context-aware data access models for mobile subscribers

(Declaration of Intent Draft)

Proposal: SkTech.RC/IT/Madnick

Theme: “BIG DATA”: large-scale data gathering & mining

Research issue: data mining services that let define context-aware actions for delivering (discovering) data to mobile subscribers. Nowadays mobile phones are becoming the primary source for possible data collections. The array of data gathered from phones (in “phone as a sensor” concept) is the typical example of schema-less big data. E.g. we can count here environmental sensing and behavioral. Based on various metrics that could be introduced for that vast amount of data we can decide what kind of information snippets could be shown (delivered) for mobile subscribers. The goal is to provide a set of tools that let define (develop) some actions/triggers (e.g. delivering information to mobile phone) depending on the collected context data in the real time. In general it leads to building richer and more personalized mobile experiences.

It should include the following elements (at least):

-  data collection (gathering) modules

-  data persistence mechanisms

-  new metrics for collected data (e.g. proximity as a service, fuzzy logic for data estimation etc.)

-  developers API for using collected data in applications

An example (on the base of existing project SpotEx):

Current model:

Data gathering: Wi-Fi networks info

Metric for data: Wi-Fi proximity

Data delivery (discovery) model, provided for mobile devices: context-aware browser where available content is defined by the proximity rules

Possible research areas that cover big data processing:

Collecting and processing additional sensing data (beyond network interfaces), analyze data for several subscribers simultaneously, historical analysis, new metrics

The relevancy. This project addresses the following hot areas in computing: M2M applications, mobile computing in the real word, context-aware (ubiquitous) computing.

Novel. Context-aware computing for mobile devices is highly fragmented. The amount of practical applications is very low. There are no (almost no) development tools that cover context-aware applications.

Broad in scope. It covers multiple research areas: mobile OS and SDK, big data stores for data persistence, real-time analysis for big data, modern programming development tools and APIs, telecom standards.

Challenging. It requires deep research efforts. There are no existing prototypes at this moment. The existing research materials and projects mostly cover either data gathering only or offline data processing (e.g. visualization) only.

Entrepreneurially promising. The possible commercialization covers the following areas for example: Smart Cities, distributing hyper-local news data to mobile subscribers (e.g. commercial info in malls, news data in campuses and office centers), real world games

Education. The following educational courses could be provided in the connection with the proposed project: mobile OS, mobile SDK, NoSQL databases, data patterns recognition, big data processing. We can develop new multi-disciplinary core courses for sensing data analysis. These courses will serve also as a basic point for PhD students.