IoT Time Series Ingest Rate Estimator¶
Idea¶
The IoT Time Series Ingest Rate Estimator helps to interactively estimate the rate at which the data is received at IoT Time Series services. Time Series Inbound Bandwidth (KB/s) in Usage Transparency reflects the same Ingest rate consumed by Tenant.
Ingest Rate¶
Ingest Rate is the timeseries data sent to IoT Time Series service from assets (devices) or applications via MindConnect elements like IoT 2040, MindConnect Nano etc or via direct API calls of IoT Put APIs over a period of time. Ingest Rate is calculated in KB/s.
Basics¶
- Assets: A digital representation of a machine or an automation system with one or multiple automation units (e.g. PLC) connected to Insights Hub. For example, a pump, motor, PLC, an entire tool machine, a production line, a robot, a crane, a car, a windmill and so on.
- Aspect: Aspects are a data modeling mechanism for assets. Aspects group related data points based on their logical association. For example: The pump skid has an aspect "Energy_consumption" that contains the datapoints "power", "current", "voltage" etc. An sspect is specified in Asset Manager and its name should have conjunction to datapoints and a physical asset. An aspect can consist of several variables.
- Variables: Variables are actual data properties of the Assets under aspect. This can be machine values like pressure, temperature, current, voltage etc.
- Record: Collection of values of an aspect variables collected from sensor at a given timestamp.
- Payload: Collection of records sent in one IoT TimeSeries API call which may contain records for one asset or multiple assets.
Estimation of Ingest Rate¶
Time Series Ingest rate is derived from the data ingestion in Time Series PUT API's. It is calculated by identifying the number of variables within an Aspect and their type. Each aspect depending on the number of variables along with sending frequency and sampling frequency will determine the size of individual record for specific aspect. This size then determines what is the overall size and how much data is expected over multiple aspects and asset combination to derive Ingest Rate. For more precise calculation, refer the Ingest Rate Estimator section.
Example Scenario¶
The ingest rate is the size of payload of data received by Time Series at a given second. If, for example, 100 KB payload is received by IoT Time Series API at one second and next for the next 9 seconds there is not data received and then again 100 KB is received and if the same cycle repeats for 1 minute, then average for that minute is 600 KB/ 60 sec which is 10 kbps.
Ingest Rate Estimator (New)¶
The new estimator (Beta Version) can estimate more accurate ingest rate due to granularity and inputs of variable name length, length of variable values for customer scenario. It is available as microsoft excel sheet. The link is available here
Info
Based on customer use-case, only cells marked in Green needs to be updated for estimation of ingest rate
Ingest Rate Estimator (Old)¶
| Aspect Type and Rate | |
|---|---|
| How often do you collect data from sensor into your nanobox | |
| How often you send data from nanobox to Insights Hub |
| Variables | |
|---|---|
| Asset and Aspect | |
|---|---|
Note
- Data from files and bulk import is not considered for ingest rate estimation.
- Above calculation is for estimation purpose only. It is expected to have variation in actual scenario due to various reasons.
- If Aspect and Variables are enabled for qualitycode, additional variables extended by _qc also gets ingested. You can consider qualitycode variable also in calculator for better accuracy.