Image Processing Reference
In-Depth Information
Top of TCP/IP
Top of Ethernet
Modified Ethernet
NRT
protocol
RT
protocol
NRT
protocol
NRT
protocol
RT
protocol
RT
protocol
TCP/UDP
IP
TCP/UDP
IP
TCP/UDP
IP
Ethernet
Ethernet
Modified Ethernet
FIGURE .
Possible structures for RTE.
Redundancy recovery time indicates the maximum time from a single permanent failure to the
network becoming fully operational again. In this case of a permanent failure, the delivery time of a
message is replaced by the redundancy recovery time.
21.4 Practical Realizations
Standard Ethernet is not able to reach the requirements of the RTE. here exist different propositions
tomodifytheEthernettechnologybytheresearchcommunity[].hemarkethasadoptedalso
additional technical solutions. All the solutions included in the standardization are presented here in
a short description.
Communication interfaces are structured in different layers. In Figure ., a simplified structure
of a communication protocol is described. Common to all Ethernet network is the universal cabling
infrastructure. Non-real-time (NRT) applications make use of the Ethernet protocols as defined in
ISO - [], and the TCP/UDP/IP protocol suite. They use typical Internet protocols like, e.g.,
HTTP for FTP for the NRT applications. To build an RTE solution, there are in principle three diferent
approaches as shown in Figure .. The first is to keep the TCP/UDP/IP protocols unchanged and
concentrate all RT modiication in the top layer; here this solution is called “on top of TCP/IP.” In the
second approach, the TCP/UDP/IP protocols are bypassed and the Ethernet functionality is accessed
directly (on top of Ethernet); in the third approach, the Ethernet mechanism and infrastructure itself
are modified to ensure RT performance (modified Ethernet).
21.4.1 Realization of the on Top of TCP/IP Protocols
SeveralRTEsolutionsusetheTCP/UDP/IPprotocolstackwithoutanymodiication.Withthisproto-
colstack,itispossibletocommunicateovernetworkboundariestransparently,alsothroughrouters.
Therefore, it is possible to build automation networks reaching almost every point of the world in the
same way as the Internet technology. However, the handling of this communication protocol stack
requires reasonable resources in processing power and memory and introduces nondeterministic
delays in the communication.
In the international standard IEC - [], all CPs have to list also at least one typical set of
performance indicators as defined in the standard. his allows the end user an easier selection of an
appropriate network for his application.
Search WWH ::




Custom Search