Basic terms

Actual Value

The actual value of a variable. Measured or indirectly calculated (e.g. Measured temperature).

Actual Value - Cumulative

Like Actual Value, but here it is a cumulative value that increases continuously (e.g. total consumed energy)


An Alarm is a detected violation of an alarm configuration, i.e. values from a time series dropped out of a user given threshold boundaries.

Alarm Configuration

An alarm configuration defines in which bounds a time series is expected to act. This information is user given. To any alarm configuration responsible users can be added which in case of an alarm should be notified.


In business administration it is used to refer to something of value that is owned or held by an organization. In our platform it is used in the sense of infrastructure asset management, for which even the ISO 55000 standard exists. In IoT platforms, the term asset is typically used to refer to what physical thing is augmented with IoT technology. For example, a machine would typically be referred to as an asset. (Note: since in our platform 2.0 we want us, our partners and customers to be very flexible in representing reality, we came up with the concept of using containers.)


A channel defines a data stream coming from a device. It has a unit (e.g. Temperature in C) and a data type (e.g. float) as attributes. The data of a channel build a time series.

Channel Type

A channel type contains meta data of the channel. It's attributes are e.g. the unit of data (°C, bar, ...), data type (e.g. float) and a name for this channel type. When creating a device type, it is linked to one or more channel types which are available for this dedicated device type.

Channel Category

Each Channel falls into one category. The available categories are:

  • Actual value
  • Actual value cumulative
  • Setpoint
  • Exogenous factor
  • Status variable

Client (In the OAuth2 sense)

A client represents a system, app, etc. using the platform (see OAuth1 standard). A client has one or more roles which control it's access rights (permissions) to it. Clients are usually used to send data, retrieve data of to build own customized visualizations like own dashboard applications.


Containers are the basic and overarching organizational element within a space in They are used in particular to digitally map organizational structures of real existing objects, e.g. plants, machines or sensors.

A container can also be enriched with descriptive data (also called metadata) according to its container type, e.g. address, hall number or assembly point. Which metadata are possible is defined in the respective container type. The container logic enables a flexible structuring of the platform and replaces the often insufficient mapping performance of a classical tree structure.

Regardless of the container type, it has zero, one or more filters that govern its content. Also independent of the container type it can have a set of tags.

Container Type

The Container Type defines with which descriptive data (metadata) a certain container in the platform can later be enriched. All containers must belong to a container type. First the container type is defined and then the container.


A dashboard is used to gain an overview or insight into a resource. A dashboard is not a resource in itself. It is created based on a dashboard template in the context of a resource, typically a container. The user can navigate between dashboards and thus switch to another container. This can be thought of as a context switch.

Dashboard Template

Dashboards are rendered based on a dashboard template. A dashboard template has zero or more widgets. It also has one or more filters to select resources. In order to create a dashboard template for a pre-determined fixed set of resources, the result set of this filter can optionally be saved, so that the filter is not re-evaluated at runtime (compare to KPI Template).

Data Source

A data source is an entity to retrieve data stored in the platform for some modeling (like KPI computation). You can find them e.g. in the dashboard-manager. We distinguish between different type of data sources (see data source type).

Data Source Type

The available types of data sources, for example time series data source, virtual time series data source, device data source, parking lot.

Datapoint (sometimes aka. Point)

Usually it refers to a channel on a particular device instance and the time series behind that. Note: However, sometimes the term is used to refer to a particular value at a point in time in a time series.


A device is a real or virtual object that can collect data in a structured way and transfer it to the cloud. This structure is defined via a device type. Example: A temperature sensor that records and transmits the temperature in degrees Celsius every 15 minutes. Devices are assigned to containers to structure them.

Device Type

The device type defines the attributes of a device. It has 1 or more channels with indices from 1 to n which are added to a newly created device instance of that type. Each channel corresponds to time series or could be related to an event stream.

Digital Twin

A digital twin is a representation of a physical device or asset.

Domain Knowledge

In Nexocraft terms domain knowledge is the knowledge about the structure of a plant including the location and the overall structure of sensors, modules, machines and plants. This kind of domain knowledge can be injected in a standardized way. In a more abstract sense the term domain knowledge also refers to any kind of unstructured knowledge about the underlying production process that helps to pre-select or finetune the underlying mathematical models and input features.


In an entity is technically represented by a JSON document that conforms to an entity type. Entities are versioned. Version 0 is always the version in the works (aka. under editing). By sealing, version 0 becomes the new version n+1, where n is the number of previously sealed versions.

Entity Type

A type for entities. It is technically represented by a JSON schema.

Exogenous Factor

A value that influences the physical system but cannot be changed intentionally. E.g. the weather


A filter is a search function with which platform resources (e.g. devices, containers, etc.) can be found dynamically according to the search criteria entered. This also includes tags. The criteria of a filter are either all linked with the boolean AND ("corresponds to all following criteria"), or all linked with the boolean OR ("corresponds to one of the following criteria").

Function Element

A pre-configured function in a KPI graph that transforms data coming in at its inlets and yields result data at its outlet(s).

IoT Gateway

A gateway that has the responsibility to connect devices (e.g. sensors) and forward their data to the cloud.


A KPI (Key Performance Indicator) is computed from a time series (e. g. measured values) and other KPIs based on a KPI template.

KPI Graph

A KPI graph is composed of one or more source elements, zero or more function elements, and one or more sink elements.

KPI Template

A KPI Template is a template to calculate KPI's from a set of time series. It can be applied to multiple instances, e.g. you can define a KPI template which is then applied to all devices of a given device type. The KPI template contains a KPI graph which is the description of how the KPI's are computed and a filter which defines the required timeseries for the calculation.

Model Performance

The model performance of an AI project is the measurement of the quality of the models results. There are many different ways to indicate the result quality of AI models. Model KPI's allow comparing the model behaviour to real world behaviour.

Model Sampling Rate

Sensor data is usually available as a raw time series with non-equal time intervals, i.e. the time between two subsequent sensor values fluctuates heavily. However, the mathematical models require an equi-distant distribution of these input values. To account for this the raw input data is sampled to equi-distant intervals where the unified lengths of these intervals is referred to as model sampling rate. The sampling rate does not only define the arrangement of the input data but also the structure of the output data, e.g. predictions are computed in an equi-distant fashion according to the same sampling rate.


A permission consists of a permission verb (read, create, update, delete, grant) and a resource UUID. In this way, permissions can be created with respect to individual resources. To model some general permissions, the platform contains some resources in a special database table, each of which represent some general thing, like for example a platform module, e.g. the KPI Manager, in order to allow a permission to refer to something more general, while technically using a resource uuid to refer to it. In this way, a permission can for example also refer to an app.


A resource is an abstract term. For example, every container, device, device type, data source, etc. is a resource.


The role of one or more users or clients. It is associated with a set of permissions.


The value that a variable is supposed to have. If there is no subordinate control device, this then becomes the manipulating variable that transmits its controlling effect to the physical system (e.g. Desired temperature).

Sink Element

A sink in a KPI graph. Represents a virtual time series computed by the KPI graph. It has one inlet. The sink will not transform the data, it will just be written to a storage.

Simulation Input Dataset

A simulation input dataset is the set of input data required for the model training and the following result computation. This might include measured values given by a devices as well as exogenous factors and user specified values.

Simulation Output Dataset

A simulation output dataset is the set of result data from a model simulation.

Source Element

A source of time series in a KPI graph. Its outlet is typically connected to an inlet of a function element. Each source element has a filter to select a data source.


A Space is the top organizational level within the platform and represents the scope of an IoT project, e.g. a customer project or proof of concept. A platform user needs an invitation and the associated permissions in order to log into a Space and make changes.


A tag is a describing word which could be attached to a resource, like to a container, device, KPI, etc.

Time Series

A time series is a typed series of values over time. For example a series of measurements over time. Can have a fixed time resolution (compare: frequency, sampling rate) such as halfminutely, hourly, daily, weekly, monthly, yearly or a varying time resolution such as raw.


A user representing a natural person using the platform (see OAuth1 standard). A user has one or more roles.

Virtual Time Series

A virtual time series is a time series which was computed and thus not measured by any sensor or device. A well known example of virtual time series are KPIs. In some cases a virtual time series contains just an intermediate result (e.g. from preprocessing) which is used for further computations afterwards.


A widget is not a resource on its own, but it is actually embedded into a dashboard template. It defines a field containing e.g. a graph with the corresponding data shown inside. The kind of widget layout is defined inside the widget type.

Widget Type

Each widget must be of a widget type. Available widget types are for example line chart, bar chart, pie chart, etc.

Did you find it helpful? Yes No

Send feedback
Sorry we couldn't be helpful. Help us improve this article with your feedback.