PTP transparent clock VLAN change capabilities

PTP runs on a single VLAN, and previously, you could not change VLANs if you used a transparent clock (TC). You had to use a PTP boundary clock (BC) if you wanted to change VLANs.

However, when you use a boundary clock, if the topology changes, the address of the BC also changes, which causes devices using PTP to have time faults. Using a PTP transparent clock when end devices exist on a different VLAN from the grandmaster clock, instead of a boundary clock, enables you to build redundant networks that can tolerate faults and does not cause time faults on those end devices.

Feature availability and requirements

Beginning with the Cisco IOS XE 17.13.1 release, when you use Precision Time Protocol (PTP), you can use the transparent clock to change VLANs on a Device Level Ring (DLR). Previously, if you used the transparent clock while using PTP, you could not change VLANs. To change VLANs, you needed to switch to a boundary clock.

This feature is supported only on Cisco Catalyst IE3400 Rugged Series Switches and Cisco Catalyst IE3400 Heavy Duty Series Switches. See Restrictions for multiple VLAN support and PTP over DLR.

This guide provides information about PTP over DLR, including guidelines and configuration steps. For detailed information about PTP and DLR, see the guides on cisco.com:

• The chapter "Precision Time Protocol" in this guide.

• The chapter "Device Level Ring" in Redundancy Protocol Configuration Guide, Cisco Catalyst IE3x00 and IE3100 Rugged, IE3400 Heavy Duty, and ESS3300 Series Switches

PTP transparent clock characteristics

The Cisco switch supports two PTP VLAN behaviors. The default configuration uses a single VLAN. The multi-VLAN feature must be enabled explicitly.

• When PTP packets are received from the grandmaster clock (GMC) and ordinary clocks (OCs), they are terminated, and new PTP packets are generated based on the PTP port VLAN configuration. That means that PTP packets with different VLAN IDs can be supported on a transparent clock.

• The PTP header and body data from the received packets are copied to the newly created packets at the PTP layer. As a result, DLR ring switches and PTP endpoints connected to DLR ring switches do not see any PTP parent change after a DLR ring fault or recovery.

• The PTP port VLAN configuration and behavior are the same as those in boundary clocks.

  • On access ports, the access VLAN is used to transmit and receive PTP packets.

  • On trunk ports, by default, the trunk native VLAN is used to transmit and receive PTP packets.

  • On trunk ports, the interface level CLI command ptp vlan vlan_id is supported to configure the trunk nonnative VLAN used to transmit and receive PTP packets. PTP packets are transmitted with a VLAN tag when the nonnative VLAN is configured as the PTP VLAN.

The ptp mode e2etransparent multiple-vlan is introduced in the Cisco IOS XE 17.13.1 release, applicable only for Cisco Catalyst IE3400 Rugged Series Switches and Cisco Catalyst IE3400 Heavy Duty Series Switches that support DLR.

Note	For details about boundary and other clocks, see the section 'PTP Clocks' in the chapter 'Configuring Precision Time Protocol guide.

Common PTP over DLR use cases

You can use PTP over DLR in a variety of different topologies for different purposes. This section provides information for two common use cases and topologies.

Single ring topology with grandmaster clock off ring:

The preceding illustration depicts a single-ring topology with the GMC off the ring and connected to the supervisor. The supervisor operates in transparent clock mode. The ring has no breaks in it, and, as a result, the supervisor blocks all traffic on Port 2 (P2) except for control traffic. The supervisor and DLR-enabled devices A, B, and C, are all in transparent clock mode. The configuration prevents devices on the ring from seeing changes within their PTP system when the Layer 1 infrastructure has a fault. It also causes DLR to open up P2.

In the illustration, a fault occurs between devices A and B. Before the fault, device C would receive PTP messages on P1 because P2 is blocking. After the fault, P2 on the supervisor goes to forwarding state. As a result, the device will receive PTP messages from the supervisor on P2.

If the supervisor is a device with multiple VLANs using normal transparent clock operation, it is not possible to support the forwarding of PTP messages on different VLANs. Configure the DLR supervisor node in transparent clock mode with multiple VLAN support. This configuration allows the use of multiple VLANs while maintaining a single time source on the ring.

DLR ring connected to spanning tree protocol topology:

The preceding diagram shows a single DLR ring used with the Spanning Tree Protocol (STP), with the GMC off the ring. The supervisor operates in BC mode, and the ring nodes and STP switches operate in TC mode.

When operating as a boundary clock (BC), all TCs and ordinary clocks view the supervisor as the PTP parent—specifically, the parent and the port number that transmits the PTP messages. In the diagram, the PTP messages are sent from Port 1. TCs and ordinary clocks use the PTP ID to recognize the PTP parent based on the PTP clock ID of the supervisor, the port number (for example, P1), and other attributes. All devices are synchronized to the PTP parent based on messages transmitted from P1.

When a fault occurs in the ring, the supervisor recognizes the fault and opens P2. This action allows all traffic to use that port, enabling the topology to reconverge and restoring connectivity to devices. The supervisor sends PTP messages with the port number set to P2 instead of P1. The change causes the ordinary clocks (the end devices) to perceive a change of the PTP parent. End devices may resynchronize with the new PTP parent, and this transition can be disruptive.

Configure the DLR supervisor node in TC mode with support for multiple VLANs to overcome this limitation. Doing so would allow the use of multiple VLANs while maintaining a single time source on the ring. Because all nodes in the DLR ring are in TC mode, the parent ID and parent port number do not change when a failure occurs in the ring.