23.03.2017 Views

wilamowski-b-m-irwin-j-d-industrial-communication-systems-2011

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Clock Synchronization in Distributed Systems 18-3<br />

Clock synchronization protocol<br />

code<br />

Timestamp point for<br />

outbound messages<br />

Outbound<br />

message<br />

stack<br />

Inbound<br />

message<br />

stack<br />

Timestamp point for<br />

inbound messages<br />

Access to <strong>communication</strong> media<br />

Synchronization message<br />

Ideal message timestamp<br />

point<br />

FIGURE 18.2<br />

IEEE 1588 node model (PTPNodeModel.pdf).<br />

data flows handle the data transmission for the two directions. The exact timestamp point for both the<br />

DELAY_REQ and SYNC is unimportant for the further processing of the protocol as long as the position is<br />

the same for all <strong>communication</strong> partners.<br />

As soon as a node detects an incoming DELAY_REQ or SYNC message, timestamps for further processing<br />

are generated. If the actual implementation is not able to detect the exact time of incoming messages,<br />

the timestamp has to be corrected in an appropriate way. The time for a message to travel between<br />

the ideal timestamp point and the actual timestamp point for inbound or outbound message point is<br />

referred to as inbound latency and outbound latency for IEEE 1588.<br />

18.4 Service Access Points<br />

Beyond the logical division of subnets, PTP relies on the possibility of distinguishing messages with the<br />

identification of the incoming service access point (SAP) mapped to the ISO model. For the example of<br />

user datagram protocol (UDP), which is defined in the annex of PTP as a possible transport protocol,<br />

this means that synchronization messages (i.e., SYNC and FOLLOW_UP) are sent to another UDP port<br />

than all other messages.<br />

18.5 Ordinary Clocks<br />

Figure 18.3 shows a model for an ordinary PTP clock. Each PTP clock uses a <strong>communication</strong> path for<br />

message exchange with other PTP nodes. This path is accessed in PTP via two different SAPs, namely the<br />

event and the general port. Both hand over messages to the protocol engine of the PTP clock. In order<br />

to model the fact that a clock may not be limited to only one PTP port, the configuration information is<br />

© <strong>2011</strong> by Taylor and Francis Group, LLC

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!