New algorithms can be opened up that normally only server farms run due to the high power requirement or massive dataset analysis
- A typical case is when learning what is normal machine behaviour so that it is then able to detect different behavioural patterns and prevent failure or machine stop. This is made possible as it obtains contextual analysis or combines unstructured data analytics . An operation which is known as Machine Learning.
- A dashboard showing a summary from large scale data sets
- Accessible data from everywhere which can be diagnostic information or status.
Application platforms are a type of cloud service (or platform) where customers’ can host their own application as well interconnect with analytics / dashboard etc.,
IoT platforms are a type of cloud service that has adapted to devices as devices connected to the cloud need to be managed:
- Over the air upgrades mean product features can be added over time, problems can be identified and fixed easily
- Identity & security policy management is crucial so all settings and certificates need to be kept update
- Device on boarding / provisioning / commissioning management
- Management device lifecycle
- Connectivity management needs to enable heterogeneous application protocols or media interconnect such as IP, LoRa, SIGFOX, cellular etc.
- Expose interfaces for application platform so access can be obtained to device data for further analytics