Ethernet is a very generic term that broadly covers the how data is shifted from one computer using either TCP or UDP protocols. But Ethernet does not define the meaning of the data. ie if TCP delivers a packet of data into my PC, what is it for, which application is it intended for and how is it used in that application?
Common application layers are web browsing, email, ftp, etc. Industrial Ethernet builds on the basic Ethernet data transport layers to define a set of application layers useful for Industrial Automation purposes, eg Control data, Configuration data and Data Collection.
The application to do this originated within Rockwell Automation and is called CIP (Control and Information Protocol). It first appeared in DeviceNet and has been migrated for use in ControlNet (plus the ControlLogix chassis backplane) and Industrial Ethernet.
In essence these networks all share the identical CIP Application Layer; they just use different transport/media layers underneath. Typically the Control Data is an IO connection between say a logic processor and an IO module. In Industrial Ethernet this is handled as a UDP (User Datagram Protocol) that is best suited for fast, low overhead connections on a local segment. By contrast messages for configuration and data collection between devices, eg RSLinx polling for processor tags, is done using a TCP (Transmission Control Protocol) connection that allows for bridging and routing between network segments.
In a nutshell, Ethernet is the pipe; Industrial Ethernet defines the contents of the pipe. Have a look around the Rockwell site and I am sure you will find more info and pretty pictures.