The full timing support architecture is defined across several documents, as shown in the following figure:
In this architecture, every node in the network must support PTP in some form, e.g. as an ordinary clock (Grandmaster or slave), or as a boundary clock or transparent clock (T-BC or T-TC). The architecture itself is described in clause 7.1 of G.8275, and this notes that “physical layer frequency support may be available to stabilize the operation of the T-BCs or telecom transparent clocks (T-TCs)”. The main source of “physical layer frequency support” in today’s networks is SyncE.
Note that this is worded as a permission rather than a requirement. However, the documents that count are much more prescriptive about the need for SyncE.
The Scope for the PTP profile, G.8275.1, states the following: “NOTE: The parameters defined in this version of the profile are chosen based on the case where physical layer frequency support is provided, and the case without physical layer frequency support (i.e. PTP only) is for further study.”
Similarly, the Scope for the performance specification of the T-BC (G.8273.2) states “The current version of this Recommendation focuses on the case of physical layer frequency support. Requirements related to the case without physical layer frequency support (i.e., the PTP only case) are for further study.”
Therefore, while SyncE may technically be an option, for all the cases that have been considered up to now, SyncE support is in effect mandatory.
That is not to say that the network will not work without SyncE, but all the simulations developed to prove that the synchronisation network functions correctly and meets the performance requirements have assumed the presence of SyncE at every node. To date, the case of “pure PTP” has not been considered or simulated, therefore it cannot be assumed that it will work without further evidence.