An Access Control System for Time Series Data in NoSQL Databases

Noury, Amir | 2015

459 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 47994 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Amini, Morteza
  7. Abstract:
  8. An important class of applications which have been rapidly growing recently is the one that create and use time series data. These types of data sets are ordered based on the timestamps associated to their data items. In practice, traditional relational databases are unable to satisfy the requirements of these data sets; however, NoSQL databases with column-wide data structure are appropriate infrastructure for them. These databases are very efficient in read and write operations (especially for time series data, which are ordered) and are able to store unstructured data. Time series data may contain valuable and sensitive information; hence, they should be protected from the information security point of view. In this thesis, we identify security requirements of time series data and survey how wide-column databases handle these requirements. One of the most important security requirements is a proper access control system enforcing access policies which are composed based on the data content with various temporal granularities (e.g. hour, day, and month). To this aim, an access control model for time series data based on wide-column data structure is proposed in this thesis. This model not only satisfies two aforementioned requirements but also provides inference control and hierarchical access control. In order to prove the applicability of the proposed model, a prototype of an access control system (based on the proposed model) was implemented as a secure access layer on top of OpenTSDB time series database and its efficiency was evaluated. The experimental results justify the soundness of the proposed model and show 4% to 20% overhead on response time for read queries. In this evaluation, the effect of various parameters on the overhead of access control have been considered; parameters like the number of value constraints, the number of data points, the fraction of data that are accessed, and the type of the operator employed in the query to access data
  9. Keywords:
  10. Access Control ; NoSQL Database ; Column-wide Databases ; Time Series Data ; Time Series Database

 Digital Object List