What Does the Data Say About the ClusterDuck Protocol v3?
The ClusterDuck Protocol (CDP) has evolved to version 3 thanks to the support of our open source community. There are many new things in the ClusterDuck Protocol version 3. Bloom filter is implemented [read more about it here], the foundation for simple commands from the cloud is built in, API is created to extract data from the network and event types are integrated in the packets or messages that the Ducks send.
In this post, we will visualize the data that was collected from CDPv3. The data is from a small deployment in the heart of Greenpoint, Brooklyn (headquarters of Owl Integrations) and contains about three months worth of data (from March 1, 2022 — May 31, 2022). It was extracted using the new APIs and the data is currently open sourced [get the data]. The data cleaning/ processing was live streamed [watch the live stream].
In the OWL office, we deployed a 4-Duck ClusterDuck Network, 3 Sensor Ducks (BMPDUCK1, IRDUCK01, and HEALTHDK) and 1 PAPADUCK. Each Duck has a unique purpose. It uses different sensors to provide situational awareness of the environment. The BMPDUCK1 is providing temperature and pressure readings, the IRDUCK01 is providing motion detection readings, the HEALTHDK is providing Duck health data, and the PAPADUCK is in charge of pushing all that data to the cloud. The sensors sends readings at a determined time except for the IRDUCK01. The IRDUCK01 sends alerts every time it detects motion.
Adding an event-type simplifies extracting key data points. For example, using the label ‘bmp’ for temperature and pressure data, we are able to extract only that data. Previously, all sensors had the ‘sensor’ event type so it was difficult and arduous to filter out data when desiring a specific sensor.
In the figure above, the number of messages received from an event type is plotted. An event type of ‘HEALTH’ provides some insights on the status of Ducks (more on that later). The IRDUCK01 and HEALTHDK sends health data which is why it has the most number of messages. We are getting health packets from two ducks. ‘BMP180’ is the event type that collects temperature and pressure. ‘GPS’ collects the GPS coordinates of the Duck. ‘PIR’ is motion detection. When it detects motion, it sends a ping to the cloud. ‘STATUS’ is when a general message is sent. Originally all the messages were labeled as ‘STATUS’. There should not be many ‘STATUS’ messages anymore since it was broken down by event types.
Extracting the data from the ‘BMP180’ event type, we get temperature and pressure sensor readings. Below we plotted the daily average readings.
Yes the office can get that hot sometimes! We operate from an old building without any central heating or cooling. Depending on how the grumpy the building is that day, it can get really hot or get really cold.
The motion detected can be considered as foot traffic. Logically, the foot traffic in the office makes sense. The weekends have the least amount of foot traffic (there are times where people had to come in on a Sunday). Friday has the least amount of foot traffic from the weekdays since Friday is optional to come to office.
Plotting the motion detection by the hour has a different perspective of the foot traffic. There are some detections in the early AM which might be concerning. But other than that, the numbers make sense. On the 12th-15th hour that is when the majority of us go to lunch or come back so there would be a lot of foot traffic. Then there is some traffic during the morning when we come in and during the evening when we leave.
Health data is important to monitor our network. The heart of the ClusterDuck Network are the Ducks. Monitoring its status is crucial to maintain a good and resilient network. These are some data points collected from the health packets from all the Ducks from the office. The HEALTHDK and IRDUCK01 are the Ducks providing health data. There seems to be a consistency between the Ducks. Both Ducks show a steady board temperature with an increase at the end and a slight decrease in volts. Based on previous deployments, the data seems normal. It will be continually monitored to see if any issues arise. There is more information we can extract from the health packets that will be discussed in Part II of the Data Analytics.
At Owl Integrations, we aim to be the world’s number one integrator, incorporating distributed network infrastructure technology to support any data need. In this small office deployment we integrated sensors to collect temperature and pressure readings and motion detection to provide situational awareness of the environment. Imagine what data you can get by integrating any sensor with our communication technology.