Unit:
Κατεύθυνση Διαχείριση Δεδομένων, Πληροφορίας και ΓνώσηςΠληροφορική
Author:
Georgantopoulos Konstantinos
Supervisors info:
Αλέξανδρος Ντούλας, Επίκουρος Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Αλέξης Δελής, Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Μέμα Ρουσσοπούλου, Αναπληρώτρια Καθηγήτρια, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Original Title:
Data Storage for IoT data
Translated title:
Data Storage for IoT data
Summary:
In several Internet-of-Things (IoT) applications we observe an abundance of data sharing the following characteristics: a) the individual data payloads are typically small in size b) they come in a time-series form c) the measurements we get are very frequent and d) the measurements are typically repeated.
IoT data capture great information in a smart space that can provide us with good insights and knowledge on the operation of the space, as well as the functionality of the devices. Most smart devices upload the whole stream of data to the cloud in order to process it and to run different types of queries. However, given the characteristics of the data (e.g. repeated data) there can be a waste of resources.
The most common representation for storage and analysis of IoT data is the classic relational, record-based approach. Although this approach satisfied users' needs for years, the rapid rise in size of data that are produced by sensors everyday has brought to light the drawbacks of this approach such as the static data model, the slow reading of whole records from the disc, the indexes needed for query optimization and the constant need to reorganize the data as they grow in size.
To address those drawbacks, alternative systems have been developed over the years. One such system is the correlation schema system which is exclusively dependent on data and there is no need for a predesigned schema or schematic constraints. Also the lack of a predesigned schema makes the system more efficient in storage usage.
In this project we plan to investigate and implement the different techniques mentioned earlier for representing and transferring IoT data to a cloud in order to save bandwidth and computational resources.
Main subject category:
Technology - Computer science
Keywords:
storage, data, time series, correlation, database
