Development Trend of Transport Network
IP has been the first choice for operators to constantly roll out new services and enhance their competitiveness. New services such as 3G, WiMAX, triple play, High Definition Television (HDTV), Video on Demand (VoD) and storage services are all delivered over IP-based network architecture. As the fast-growing carrier-class IP services pose higher requirements for existing transport networks, Packet Transport Network (PTN) will become the development trend.
Mainstream PTN Technologies
Two mainstream technologies are available for PTN: Transport Multi-Protocol Label Switching (T-MPLS) and Provider Backbone Transport (PBT).
T-MPLS is a connection-oriented packet transport technology developed on the basis of IP/MPLS. It discards the complex control protocol stack defined by IETF, simplifies the data plane, removes unnecessary forwarding processes and connectionless features that are independent of transport, and adds layered transport network model, protection switching and OAM functions. Moreover, it defines a multitude of adaptation interfaces for client signals, enabling transport resource management and automatic service configuration through the Automatic Switched Optical Networks/Generalized Multi-Protocol Label Switching (ASON/GMPLS) control plane.
PBT is a technology developed on the basis of Ethernet switch that supports L2/L3/L4 processing. It enhances manageability and protection features, adds narrowband service emulation and clock functions, and offers powerful multi-service support capability. Furthermore, it disables traditional Ethernet functions such as MAC address learning, broadcasting and Spanning Tree Protocol (STP), and uses the management plane (or control plane in the future) to have full control over the Ethernet forwarding table.
The PTN network is protected by the OAM mechanism of the transport plane, which offers sub-50ms protection switching capability. The network protection standards involve linear protection and ring protection, as shown in Table 1.
In terms of linear protection, T-MPLS supports trail protection and Sub-Network Connection (SNC) protection, while PBT does not support SNC protection; in respect of ring protection, T-MPLS supports wrapping and steering protection modes, while PBT does not support ring protection.
- Trail protection defined by the ITU-T standard
The trail protection (see Figure 1) defined by the ITU-T standard includes 1+1 and 1:1 trail protection, supporting unidirectional and bidirectional, revertive and non-revertive modes.
In the 1+1 architecture, a permanent bridge is at the source side and a selector at the sink side. The protection connection is dedicated to each working connection. The working traffic is permanently bridged to the working and protection connections at the source side, and transmitted simultaneously to the sink side, where a selection between the working and protection connections is made based on some predetermined criteria, such as signal failure and signal degradation. For 1+1 bidirectional trail protection switching, Automatic Protection Switching (APS) protocol is required to ensure the selectors in both directions can select the same connection. In the 1+1 unidirectional trail protection switching, the APS protocol is not supported.
The 1:1 architecture only operates with bidirectional protection switching. The protection connection is also dedicated to each working connection. However, the traffic is selected by the bridge selector at the source side and transmitted either over the working or protection connection. The selector at the sink side selects the connection that carries the traffic. Since both the source and sink sides need to be coordinated to ensure that the selector bridge at the source side and the selector at the sink side select the same connection, the APS protocol is necessary.
- SNC protection defined by the ITU-T standard
The SNC protection (see Figure 2) defined by the ITU-T standard contains 1+1 and 1:1 SNC protection, supporting unidirectional, bidirectional, revertive and non-revertive modes.
The operating principle of the SNC protection is basically the same as that of the trail protection, and the only difference is that it can protect connections within the subnetwork and support flexible networking.
- Trail protection defined by the IEEE standard
The trail protection defined by the IEEE standard includes 1:1 bidirectional trail protection and 1:1 bidirectional trail protection (load sharing), supporting unidirectional, bidirectional, revertive and non-revertive modes.
The 1:1 bidirectional trail protection uses the Connectivity Fault Management (CFM) OAM mechanism instead of the APS protocol. The protection switching can be triggered as soon as trail failure, such as alarm indication signal (AIS) and remote defect indication (RDI), is detected at either side.
For the 1:1 bidirectional trail protection in load sharing mode, when working and protection connections are in normal operation, the traffic is transmitted simultaneously over working and protection connections; in the event of a certain connection failure, the traffic can only be transmitted over the normal connection.
At present, T-MPLS can support T-MPLS Shared Protection Ring (TM-SPRing) protection, but PBT does not support ring protection.
The T-MPLS ring protection (see Figure 3) requires the completion time for protection against a single failure should be less than 50ms assuming a reference network with a 16-node ring and less than 1200km of transmission distance.
Similar to SDH Multiplex Section Shared Protection Ring (MS-SPRing), T-MPLS ring protection provides two protection modes: wrapping and steering. The wrapping mode is similar to normal SDH MS-SPRing, and the steering mode is similar to SDH cross-ocean protection switching.
- Wrapping mode
When a network node detects a failure, the adjacent nodes at the two sides of the failure will send a switching request through the APS protocol to the opposite nodes in the far-end and near-end directions respectively. When the nodes receive the switching request, the traffic transmitted towards the failed span is switched to the opposite protection connection. After the failure is cleared, the traffic returns to the original working connection.
- Steering mode
When a network node detects a failure, a switching request will be sent through the APS protocol to all nodes in the ring. For the affected traffic, each source service node performs switching from working to the opposite protection connection. After the failure is cleared, the traffic returns to the original working connection.
All protection technologies for PTN mentioned above can meet the requirements for carrier-class protection switching of less than 50ms. T-MPLS offers relatively more complete protection technologies than PBT. As an emerging carrier-class transport network, PTN is being perfected in technologies and standards, which will drive the growth of the relative network protection technologies.