Image Processing Reference
In-Depth Information
Today, most memory devices and microcontrollers provide an interface for in-system serial pro-
gramming of flash and EEPROM memory. he hardware interface for in-system serial programming
usually consists of a connector with four to six pins that is attached to either an external programming
device or directly to the development PC. hese programming interfaces are often proprietary to par-
ticular processor families; but there also exist some standard interfaces that support a larger variety of
devices. For example, the joint test action group (JTAG) debugging interface (IEEE Standard .)
alsosupportstheuploadofapplicationcode.
While the in-system serial programming approach is much more convenient than the socketed
EPROM method, both approaches are conceptually quite similar, since it is still necessary to establish
a separate hardware connection to the target system. A more advanced approach for uploading appli-
cations is in-programming. In this approach it is possible to program and configure a device without
taking it out of the distributed target system and without using extra cables and hardware interfaces.
In-system configuration is supported by state-of-the-art flash devices, which can reprogram them-
selves in part by using a bootloader program. Typically, whenever a new application has to be set up,
the programming system sends the node a signal causing it to enter a dedicated upload mode. During
the upload phase, usually the node's service is inactive. Failures that lead to a misconfiguration must
be corrected by locally connecting a programming tool.
Alternatively, application code could be downloaded via the fieldbus into the RAM memory at
startup. In this case, only the bootloader resides in the persistent memory of the device and the user-
defined application code has to be downloaded at startup. his approach has the advantage of being
stateless,sothaterrorsinthesystemareremovedatthenextstartup.hus,theengineerscouldhandle
many faults by a simple restart of the system. On the other hand, this approach depends on the con-
figuration instance at startup—the system cannot be started, if the configuration instance is down.
Moreover, the restart time of a system may be considerably longer.
22.8 Management Interfaces
The possibility for performing remote management operations on distributed fieldbus devices is one
of the most important advantages of fieldbus systems. Wollschläger states that “in automation sys-
tems, engineering functions for administration and optimization of devices are gaining importance
in comparison with control functions” [, p. ].
Typical management operations are monitoring, diagnosis, or node calibration. Unlike the pri-
mary fieldbus applications, which often require cyclical, multi-drop communication, these manage-
ment operations usually use a one-to-one (client-server) communication style. For this reason, most
fieldbus systems support both communication styles.
A central question is whether and how this management traffic influences the primary appli-
cation; a problem known as probe effect []. System management operations that influence the
timing behavior of network communication are especially critical for typical fieldbus applications
(e.g., process control loops) that require exact real-time behavior.
The probe effect can be avoided by reserving a fixed amount of the bandwidth for management
operations. For example, in the FF and WorldFIP protocols the application cycle (macrocycle) is
chosen to be longer than strictly required by the application, and the remaining bandwidth is free for
management traffic.
In order to avoid collisions within this management traffic window, adequate mechanisms for
avoiding or resolving such conflicts must be used (e.g., token-passing between nodes that want to
transmit management information, and priority-based arbitration).
In TTP/A, management communication is implemented by interleaving real-time data broad-
casts (implemented by multipartner rounds) with the so-called master-slave rounds that open a
communication channel to individual devices.
