As the widely used combination of TCP/IP (e.g. for web videos or general http) does not provide the required data throughput and latency, the GigE Vision standard was designed under the umbrella of AIA (Automated Imaging Association). This standard is UDP/IP based: a lightweight, high-speed protocol that does not require prior communication. The GigE Vision protocol introduces reliable communication with cameras/devices while supporting high-speed streaming of image data over standard Gigabit Ethernet networks. Unlike the IIDC 1394 (FireWire/DCAM) standard, GigE Vision is mainly a transmission protocol standard. It merely describes the nature of the communication with the camera control and data streaming. Feature descriptions and access are handled by the referenced GenICam standard (refer to chapter 6.8). UDP/IP, unlike TCP/IP, also supports point-to-multipoint connections. This means that a camera can transmit image data to several computers at the same time (multicast). There is always a master application that maintains control of the camera, but other computers/applications can read the camera and image data. In this way, complex processing situations can easily be split among several computers, or monitoring stations can be set up that do not obstruct processing.
The GigE Vision standard uses many networking options and does not restrict GigE Vision devices to image sources/cameras: for example, devices such as strobe controllers or transparent preprocessing boxes, which are easy to handle by the combination of GigE Vision / GenICam. Other potential software solutions include the CVB GigE Vision Server that make applications behave like a GigE Vision device/camera and thus enable other hardware to be GigE Vision compatible, or enable any kind of processing or configuration.
In general the GigE Vision standard consists of three parts:
- GigE Vision - Device Discovery
- GigE Vision - GVCP (GigE Vision Control Protocol)
- GigE Vision - GVSP (GigE Vision Streaming Protocol)
GigE Vision performance
The performance level of a GigE Vision system depends strongly on the hardware and software components and their configuration. Below we describe a few performance considerations: