Resense: Transparent Record and Replay of Sensor Data in the Internet of Things (Poster)
0 likes207 views
This work was supported by the German Ministry for Education and Research and the European Union to develop Resense, a framework that allows researchers to transparently emulate sensors and replay recorded sensor data from Internet of Things (IoT) experiments. Resense provides an easy-to-use software and graphical user interface to set up and execute repeatable IoT experiments involving heterogeneous sensors on a testbed of Raspberry Pis. The paper demonstrates Resense by replaying real-world sensor data from a sports analytics application.
![This work was supported by the German Ministry for
Education and Research as Berlin Big Data Center
(01IS14013A) and the European Union’s Horizon 2020, under
the Marie Skłodowska-Curie grant agreement No 765452.
22nd International Conference on Extending Database Technology, March 26-29, 2019, Lisbon, PT
Resense: Transparent Record and Replay of Sensor Data
in the Internet of Things
Dimitrios Giouroukis Julius Hülsmann Janis von Bleichert Morgan Geldenhuys Tim Stullich
Felipe O. Gutierrez Jonas Traub Kaustubh Beedkar Volker Markl
Open Source Repository
Resense open source repository:
https://github.com/TU-Berlin-DIMA/resense
Abstract
Conducting repeatable and scalable experiments on Internet of Things
(IoT) test beds requires (i) manual fine tuning with extensive A/B
testing and (ii) edge case / rare event testing that leads to extremely
long test durations.
Being able to replay sensor data on real test beds enables researchers
to run repeatable experiments without these problems.
In this paper we make the following contributions:
Our proposed Resense framework allows for replaying sensor data
using emulated sensors and provides an easy-to-use software for
setting up and executing IoT experiments.
1. We show how to transparently emulate sensors in order to
record and replay sensor data.
2. We provide an easy to use software for setting up and executing
IoT experiments involving heterogeneous sensors.
3. We demonstrate recording and replaying real world sensor data
in the context of sports analytics on a set of Raspberry Pis.
Resense Architecture
• Master/Slave approach to administrating nodes
• Users can control Resense through a Graphical User Interface
• Deployment of sensor data and configuration to edge nodes
• Replay and record data on demand
For external applications, it is transparent whether they read
from physical or emulated sensors.
Edge Node Internals
①. Resense Recorder reads values from a physical sensor.
②. The value is stored for later use in replays by the system.
③. Resense Replay reads the values stored by the Recorder.
④. Resense Replay writes the values to emulated sensors that
are allocated in kernel memory.
⑤. External applications can still read from both, physical and
emulated sensors, since they access them the same way.
Experiment Dashboard
Resense’s UI is split into three main panels:
①. Control panel, shows list of experiments and sensors
involved in the current configuration.
②. Monitoring panel, displays statuses of current recordings
and experiments that are running.
③. Live view panel, shows time series data for sensor
readings for currently running experiments.
Experiment Configuration
"experiment": "edbt-2019",
"nodes": [
{
"edgeId": "edge-2",
"host": "192.168.1.2",
"user": "pi",
"password": "raspberry",
"sensors": [
{"sensorId":"ball","type":["Integer"],"address":3},
{"sensorId":"referee_left","type":["REAL"],"address":4}
]
},{...}]
An example configuration that portrays part of a sensor
testbed setup that replays data from a football match.](https://image.slidesharecdn.com/edbt-resense-poster-190402024513/85/Resense-Transparent-Record-and-Replay-of-Sensor-Data-in-the-Internet-of-Things-Poster-1-320.jpg)
Resense: Transparent Record and Replay of Sensor Data in the Internet of Things (Poster)
- 1. This work was supported by the German Ministry for Education and Research as Berlin Big Data Center (01IS14013A) and the European Union’s Horizon 2020, under the Marie Skłodowska-Curie grant agreement No 765452. 22nd International Conference on Extending Database Technology, March 26-29, 2019, Lisbon, PT Resense: Transparent Record and Replay of Sensor Data in the Internet of Things Dimitrios Giouroukis Julius Hülsmann Janis von Bleichert Morgan Geldenhuys Tim Stullich Felipe O. Gutierrez Jonas Traub Kaustubh Beedkar Volker Markl Open Source Repository Resense open source repository: https://github.com/TU-Berlin-DIMA/resense Abstract Conducting repeatable and scalable experiments on Internet of Things (IoT) test beds requires (i) manual fine tuning with extensive A/B testing and (ii) edge case / rare event testing that leads to extremely long test durations. Being able to replay sensor data on real test beds enables researchers to run repeatable experiments without these problems. In this paper we make the following contributions: Our proposed Resense framework allows for replaying sensor data using emulated sensors and provides an easy-to-use software for setting up and executing IoT experiments. 1. We show how to transparently emulate sensors in order to record and replay sensor data. 2. We provide an easy to use software for setting up and executing IoT experiments involving heterogeneous sensors. 3. We demonstrate recording and replaying real world sensor data in the context of sports analytics on a set of Raspberry Pis. Resense Architecture • Master/Slave approach to administrating nodes • Users can control Resense through a Graphical User Interface • Deployment of sensor data and configuration to edge nodes • Replay and record data on demand For external applications, it is transparent whether they read from physical or emulated sensors. Edge Node Internals ①. Resense Recorder reads values from a physical sensor. ②. The value is stored for later use in replays by the system. ③. Resense Replay reads the values stored by the Recorder. ④. Resense Replay writes the values to emulated sensors that are allocated in kernel memory. ⑤. External applications can still read from both, physical and emulated sensors, since they access them the same way. Experiment Dashboard Resense’s UI is split into three main panels: ①. Control panel, shows list of experiments and sensors involved in the current configuration. ②. Monitoring panel, displays statuses of current recordings and experiments that are running. ③. Live view panel, shows time series data for sensor readings for currently running experiments. Experiment Configuration "experiment": "edbt-2019", "nodes": [ { "edgeId": "edge-2", "host": "192.168.1.2", "user": "pi", "password": "raspberry", "sensors": [ {"sensorId":"ball","type":["Integer"],"address":3}, {"sensorId":"referee_left","type":["REAL"],"address":4} ] },{...}] An example configuration that portrays part of a sensor testbed setup that replays data from a football match.