Guidelines to Configure PTP over Ethernet

Keep the following points in mind when you configure PTP over Ethernet for multicast mode of transmission of PTP traffic:

• You can configure a port or a logical interface to be a timeTransmitter clock for PTP over Ethernet to provide packet-based synchronization to base stations that support time and phase alignment; this configuration is compliant with Annexure F of the IEEE 1588-2008 specification.

• Two multicast MAC addresses are used for PTP over Ethernet: 01-1B-19-00-00-00 and 01-80-C2-00-00-0E. The first address is a more standard Ethernet MAC address that is expected to be flooded by all types of Ethernet bridges and switches and also by a large number of base station vendors. A node with this MAC address can be a node that does not process PTP packets. The second address is a reserved address in the IEEE 802.1Q standard for Ethernet encapsulation that is required to be filtered and not forwarded. This MAC address is used to ensure complete end-to-end support of PTP, instead of transmission of packets through any network element that does not support PTP. This address is the default address for G.8275.1 (PTP Profile for time or phase distribution) and a node with this MAC address is a node that supports processing of PTP packets.

• Both of the MAC addresses, 01-1B-19-00-00-00 and 01-80-C2-00-00-0E, are supported on multiple ports simultaneously to enable maximum flexibility and extension of existing networks for future deployments. A single PTP port is configured for one of the MAC addresses at a time.

• PTP packets are sent with the unique MAC address assigned to each port as the MAC source address. In the PTP packet, the Ethernet frame portion of the packet contains the Destination MAC field. This field contains either of the two MAC addresses, 01-1B-19-00-00-00 or 01-80-C2-00-00-0E. Also, the Ethernet frame portion contains the Source MAC field that contains the MAC address of the source port and the Ethertype field that contains the PTP Ethertype value of 0x88F7. Apart from the Ethernet frame, the PTP packet contains the PTP payload.

• When you configure a port for PTP over Ethernet to be a timeReceiver port, a timeTransmitter port, or both by having a stateful port that can be either a timeTransmitter port or a timeReceiver port depending on the states of the other ports in the PTP application, it is possible to build an easily provisioned, redundant PTP service in an Ethernet ring where every node is configured as a boundary clock.

• A boundary clock can function as a timeReceiver clock to a device using IP (such as a TCA Series Timing TimeReceiver or an MX Series router) on one port and can also function as a timeReceiver clock, a timeTransmitter clock, or both on other ports using Ethernet as the encapsulation method. This behavior occurs within a single PTP domain number.

• Best TimeTransmitter Clock Algorithm (BTCA) and the port state machine are supported to determine the states of all the ports in a system and the correct state (timeTransmitter or timeReceiver) for a certain port to process PTP packets.

• PTP over Ethernet supports fully redundant and resilient ring-based configurations of up to 10 nodes for a form of fourth-generation (4G) evolution known as Long-Term Evolution-Time Division Duplex (LTE-TDD). ACX Series routers support a single node or link failure and all nodes maintain a phase accuracy of plus or minus 1.5 microseconds matching a common source.

• You can configure the asymmetry value between the timeTransmitter port and the timeReceiver port, which indicates a value to be added to the path delay value to make the delay symmetric and equal to the path from the timeTransmitter port to the timeReceiver port, on either a dynamic-state port or a timeReceiver-only port.

• You cannot enable PTP over Ethernet on Ethernet interfaces that are configured with 802.1Q VLAN tags or contain a user-configured MAC address.

• While you can configure unique PTP timeReceiver interfaces or timeReceiver ports with different encapsulation mechanisms (such as IPv4 and Ethernet), the boundary clock can use only a single encapsulation method for all of the timeTransmitter ports. Therefore, you must define either IPv4 or Ethernet encapsulation for all the ports or logical interfaces that can possibly function as boundary clock primaries. TimeTransmitter ports select the link-local flag based on each port.

• You can configure a maximum of 128 PTP over Ethernet ports, where up to 4 ports can be configured as timeReceiver and the remaining can be configured as timeTransmitter.

• In PTP over IPv4 deployment, it is necessary to configure certain basic settings on a PTP timeTransmitter port before the PTP timeReceiver ports to connect to the timeTransmitter port. PTP over Ethernet offers a plug-and-play service because any PTP timeReceiver starts receiving packets and can request delay-response packets from the timeTransmitter port after you configure an interface to be a timeTransmitter.

• PTP over Ethernet is compatible with Junos OS releases earlier than Release 12.3X51. When you perform an upgrade to Release 12.3X51 and later from a previous release on an ACX Series router, you can modify the timeReceiver and timeTransmitter ports previously configured for IPv4 to enable PTP over Ethernet based on your network needs.

• You cannot configure a fully redundant PTP ring using IP. A fully redundant PTP ring is supported only when Ethernet encapsulation is used.

• Configuration of dynamic ports in conjunction with Synchronous Ethernet to enable hybrid mode is not supported.

• Multiple PTP timing domains are not supported for PTP over Ethernet, similar to PTP over IPv4. Although a single node can contain interfaces configured for PTP over IPv4 and PTP over Ethernet, both of these interfaces must be part of the same PTP domain.

• SONET/SDH networks define the ability to configure a local priority to a synchronization source in the ITU G.781 standard. Addition of such locally configured priorities to PTP sources to influence BTCA to determine a particular path for PTP packets is not supported.

• Although you can configure a timeReceiver port to use either IP or Ethernet simultaneously, a single timeReceiver port is selected based on the announce messages it receives from the timeTransmitter port and the PTP event packets are exchanged only with a single timeTransmitter port.

• The IPv4 unicast implementation of PTP enables you to limit the number of timeReceiver ports that can be supported simultaneously in the system. With multicast Ethernet-based implementation, in which the timeTransmitter port is not provisioned with the timeReceiver port information, the timeTransmitter port cannot limit the number of timeReceiver ports that it services. This control must be exercised with proper networking planning and design.

• PTP works well with Media Access Control Security (MACsec) encryption enabled on the same port at the same time on supported routers. The following limitations are applicable:

  • The maximum limit for MACsec-enabled logical interfaces (IFL) is 200 per system.

  • The maximum limit for MACsec-enabled ports with physical interfaces (IFDs) and IFLs where MACsec and PTP are enabled together on different ports is 200 per system.

  • The maximum number of IFLs that can be supported on both 1G and 10G ports is 128.

  • PTP in clear text mode is not supported.