MSTP Working Principle and Functions
MSTP (Multi-Service Transfer Platform) (SDH-based Multi-Service Transport Platform) refers to a multi-service node that simultaneously implements the access, processing, and transport of TDM, ATM, Ethernet, and other services based on an SDH platform, providing unified network management. In addition to the functions of standard SDH transport nodes, SDH-based multi-service transport nodes also have the following main functional characteristics.
(1) Access functions for TDM services, ATM services, or Ethernet services;
(2) Transport functions for TDM services, ATM services, or Ethernet services, including point-to-point transparent transport;
(3) Bandwidth statistical multiplexing functions for ATM services or Ethernet services;
(4) Mapping and provisioning functions for ATM services or Ethernet services into SDH virtual containers.
SDH-based multi-service transport nodes can be applied in the access layer and aggregation layer of the transport network according to network requirements; their application in the backbone layer is subject to further study.
The metropolitan area network (MAN) is currently the focus of competition among telecom operators. At present, there are various MAN networking technologies, broadly including SDH-based MANs, Ethernet-based MANs, ATM-based MANs, and DWDM-based MANs. In fact, SDH, ATM, Ethernet, WDM, and other technologies are continuously learning from each other's strengths, complementing one another—aiming to achieve fast transmission while supporting multi-service carriage and providing carrier-class QoS. A trend of convergence is emerging among various MAN technologies.
MSTP Working Principle
MSTP can integrate multiple independent devices, such as traditional SDH multiplexers, digital cross-connect systems (DXC), WDM terminals, Layer 2 switches, and IP edge routers, into a single network device, namely an SDH-based Multi-Service Transport Platform (MSTP), for unified control and management. SDH-based MSTP is best suited as a converged edge node supporting mixed services, particularly TDM-dominated mixed services. It is suitable not only for new operators lacking network infrastructure, for deployment between offices or POPs, but also for large enterprise customer premises. Moreover, even for operating companies that have already deployed extensive SDH networks, SDH-based multi-service platforms can more effectively support packet data services and facilitate the transition from circuit-switched networks to packet networks. Therefore, it will become one of the mainstream technologies for MANs in the near term.
This requires SDH to transform from a transport network into a multi-service platform integrating transport and service networks, i.e., a converged multi-service node. The implementation foundation of MSTP is to fully leverage SDH technology's protection and recovery capabilities for transmitted service data streams and its low latency performance, while upgrading the network service support layer to adapt to multi-service applications and provide intelligent support for Layer 2 and Layer 3 data. That is, transport nodes are integrated with various service nodes to form SDH service nodes with integrated service and transport layers, known as converged network nodes or multi-service nodes, primarily positioned at the network edge.
Characteristics of MSTP
(1) Flexible bandwidth configuration for services. The 10/100/1000Mbit/s interfaces provided by MSTP can meet various user requirements through VC bundling.
(2) Depending on service requirements, it can operate in port group mode or VLAN mode. VLAN mode can be further divided into access mode and trunk mode:
Port Group Mode: All system and user ports on a single board are within one port group. This mode can only be applied to point-to-point services. In other words, when any user port and any system port (since there is only one direction, it is unnecessary to enable all system ports; one is sufficient) are enabled, and a network cable is plugged into any enabled user port, that user port enjoys all the bandwidth and the service can be activated.
VLAN Mode: Divided into access mode and trunk mode.
In access mode, if no VLAN ID is set, the port operates in port group mode, with all system and user ports on the board within one port group.
If a VLAN ID is set, the "port VLAN tag" must be configured. This is because the switching chip adds a VLAN ID to received data packets, which are then transmitted via the system port over fiber to the peer port with the same VLAN ID. For example, if a user port has a VLAN ID of 2, the corresponding user port at the remote site should also be set to VLAN ID 2. This mode can be applied to MSTP services in multiple directions, where ports in each direction must be assigned different VLAN IDs. The user ports and system ports in that direction are then placed into a virtual bridge (the VLAN ID of this virtual bridge must match the "port VLAN tag").
(3) It can operate in full-duplex, half-duplex, and auto-negotiation modes, with MAC address self-learning capability.
(4) QoS configuration:
QoS essentially limits port transmission. The principle is that the transmitting port has multiple transmission queues based on service priority, and transmission of various priority services is completed according to QoS configuration and specific algorithms. Therefore, when a port may transmit services from multiple sources and the total traffic may exceed the transmitting port's bandwidth, the port's QoS capability can be configured, along with corresponding priority settings for various services. When QoS is not configured, bandwidth is evenly distributed and services from multiple sources are transmitted on a best-effort basis.
QoS configuration defines the priority and bandwidth allocation quota for each port when sharing the same bandwidth.
(5) Independent Spanning Tree Protocol operation for each customer.
Advantages of MSTP
(1) At present, the demand from a large number of users remains fixed-bandwidth dedicated lines, mainly 2Mbit/s, 10/100Mbit/s, 34Mbit/s, and 155Mbit/s. These dedicated line services can be broadly divided into fixed-bandwidth services and variable-bandwidth services. For fixed-bandwidth services, MSTP equipment inherits excellent carriage and scheduling capabilities from SDH. For variable-bandwidth services, end-to-end transparent transmission channels can be provided directly on MSTP equipment, fully ensuring service quality. The Layer 2 switching and statistical multiplexing functions of MSTP can be fully utilized to share bandwidth and reduce costs, while VLAN partitioning is used to isolate data and different Classes of Service (CoS) are used to guarantee service quality for key users.
(2) At the MAN aggregation layer, service aggregation from enterprise network edge nodes to central nodes is achieved, characterized by multiple nodes, multiple port types, dispersed user connections, and a large number of ports. By adopting MSTP networking, aggregation or direct access of 10M/100M/1000M POS and 2M/FR services from IP routing equipment can be realized, supporting service aggregation scheduling, integrated carriage, and good survivability. MSTP equipment of different rate levels can be selected according to different network capacity requirements.
Applications of MSTP
MSTP technology has attracted significant attention in existing metropolitan area transport networks, has been deployed on a large scale, and is about to be released as an industry standard. Compared with other technologies, its technical advantages lie in: solving the problem of low efficiency in carrying data services over SDH; solving the problem of low efficiency and high cost in carrying TDM services over ATM/IP; solving the problem of insufficient IP QoS; solving the networking limitations of RPR technology, achieving dual protection and improving service security; enhancing the network concept of data services and improving network monitoring and maintenance capabilities; reducing service selection risks; achieving the networking advantages of reduced investment, unified network construction, and on-demand deployment; and adapting to full-service competition demands with rapid service provisioning.
MSTP enables the transport network to evolve from a supporting network into a bandwidth-operating network with independent operational value, leveraging its mature technical advantages to provide high-quality, low-cost bandwidth resources and meet metropolitan bandwidth demands. Due to its multi-service characteristics, a metropolitan transport network built with B-ADM equipment can provide a rich variety of bandwidth service offerings according to user requirements. Under the MSTP technical framework, B-ADM equipment incorporates operational concepts and intelligent features in network scheduling and equipment aspects, enabling convenient and rapid service establishment, thereby further ensuring the feasibility of bandwidth operations and meeting market demands for metropolitan transport networks.
Requirements for related systems in the "MSTP dedicated line" service:
(1) Network implementation
The networking model for the "MSTP dedicated line" service is that MSTP equipment is deployed at the access end to access services, connected downstream to customer equipment and upstream to local SDH network equipment. The existing transport network (other networks are not considered for now) is used as the bearer network for this service. The MSTP equipment at both ends can adopt (or upgrade) existing MSTP equipment or be newly purchased, depending on the actual conditions of each local network.
(2) Configuration requirements for customer equipment
When point-to-multipoint Ethernet dedicated line services are activated, if customer equipment at branch nodes needs to set VLAN IDs for services, the operator must be informed and VLAN IDs negotiated to ensure that each branch node has a distinct VLAN ID for identification by the aggregation node.
When point-to-point Ethernet dedicated line services are activated, no configuration requirements are imposed on customer equipment.
(3) Interoperability requirements for MSTP equipment
MSTP equipment is selected, procured, and maintained by provincial companies and must have interoperability capabilities.
(4) Requirements for the service bearer network
The "MSTP dedicated line" service only requires MSTP equipment at the access layer of the network; existing SDH transport network resources can be utilized within the network. Since MSTP's support for Ethernet services is implemented through technologies such as GFP, virtual concatenation, and LCAS, all of which require SDH path overhead bytes to carry control information, transparent transmission of SDH path overhead bytes must be ensured. That is, the "MSTP dedicated line" service must not involve the add/drop or cross-connection of 2M circuits, and STM-N interfaces must be used for network connections.
(5) Support for Ethernet services
MSTP (Multi-service Transport Platform) is a metropolitan area transport network technology that organically integrates SDH transport technology, Ethernet, ATM, POS, and other technologies. Based on SDH technology, it aggregates and effectively adapts multiple services, achieving integrated access and transport of multi-services and transforming SDH from a pure transport network into a multi-service platform integrating transport and service networks. From the current state of transport networks, most metropolitan transport networks still rely primarily on SDH equipment. Considering technical maturity, reliability, and cost, SDH-based MSTP technology plays a very important role in MAN applications. With the rapid growth of IP services such as data and broadband in recent years, the development of MSTP technology is mainly reflected in its support for Ethernet services, and the requirements of new Ethernet services are driving the advancement of MSTP technology.
1. How MSTP carries and transports Ethernet services
In the evolution of MSTP technology, based on service application scenarios, the following methods currently exist for carrying and transporting Ethernet services over MSTP:
(1) Transparent transmission of Ethernet services. This is currently a widely used method and was the approach adopted in the early stages of MSTP to achieve transparent transmission of Ethernet services over SDH equipment. This method is solely intended to achieve transparent transmission of Ethernet services, using a protocol (PPP/LAPS/GFP) to directly encapsulate frame signals of non-switched Ethernet services, and then utilizing technologies such as PPP over SDH and inverse multiplexing (dispersing high-speed data streams across multiple low-speed VCs to improve transmission efficiency, e.g., using 5*VC12 concatenation to transmit 10MB/S Ethernet services) to achieve interconnection between two points. Since different manufacturers use different methods for mapping Ethernet services into VCs and adopt different protocols, Ethernet services after transparent transmission must be terminated on equipment from the same manufacturer.
(2) Performing Layer 2 switching on Ethernet services before encapsulation, then mapping them into SDH VCs and transmitting them on the line side, which better adapts to the dynamic characteristics of data services. This method integrates Layer 2 Ethernet frame (MAC frame) switching into the tributary card of SDH equipment. The Layer 2 switch learns the MAC addresses of devices connected to the network and switches frame signals to the correct port based on the destination MAC address. Therefore, MSTP equipment can process Ethernet services as follows:
① MSTP can aggregate multiple low-rate Ethernet services distributed across various locations and transmit them to a single or multiple high-rate Ethernet interfaces at a specific location.
② It enables statistical multiplexing of Ethernet services, effectively utilizing bandwidth on the line side. MSTP can aggregate Ethernet services from multiple Ethernet interfaces into a single high-bandwidth pipe, allowing a single line-side channel to be shared by multiple users. This ensures both peak traffic capacity during Ethernet service bursts and efficient bandwidth utilization (since Ethernet services often have no traffic during many time periods). For example, five Fast Ethernet interfaces can share a 155 MB/s transmission bandwidth on an MSTP, reducing operating costs.
③ Multiple methods can be effectively employed to isolate services of different users, ensuring the security of user data. One method is to provision dedicated channels for users' Ethernet services, i.e., mapping services into separate VCs, thereby achieving effective physical-layer isolation of user services. Alternatively, VLAN tagging can be applied to users' Ethernet services using the 802.1Q standard, isolating user services through VLAN partitioning; when necessary, additional tagging on top of the 802.1Q tag can also be used to isolate user services.
(3) Some MSTP devices combine Layer 3 switching with SDH network elements, representing an extension of the Layer 2 switching approach. In this mode, user service signals are forwarded to the correct port or SDH line-side channel based on IP addresses rather than MAC addresses; it offers the same advantages as Layer 2 switching while effectively isolating broadcast packets arising from MAC addressing. However, Layer 3 switching operates at the service layer, and due to factors such as technology, cost, and network maintenance, this approach is less commonly used in MSTP devices.
(4) Introducing the processing mechanism and functions of RPR (Resilient Packet Ring) into MSTP. RPR is a new MAC layer protocol centered on Ethernet technology, proposed to optimize data packet transmission. It not only effectively supports ring topologies and enables fast recovery upon fiber disconnection or connection failure, but also employs spatial reuse mechanisms to provide efficient bandwidth sharing. It exhibits typical Ethernet characteristics such as high efficiency, simplicity, and low cost in data transmission, and is currently being standardized by the IEEE 802.17 working group. Some time slots can be extracted from the SDH layer of MSTP and mapped from RPR to SDH frame structures using the GFP protocol to construct a logical RPR ring, with services accessed through Fast Ethernet and Gigabit Ethernet interfaces on the RPR line card.
2. Key Technologies for MSTP Carrying and Transporting Ethernet Services
(1) Encapsulation protocols: When carrying and transporting Ethernet services, MSTP must first encapsulate Ethernet signals using a certain protocol. There are many encapsulation protocol options, the most common being PPP, LAPS, GFP, as well as proprietary encapsulation mechanisms from various equipment vendors. PPP (Point-to-Point Protocol) utilizes HDLC (High-Level Data Link Control) for framing; HDLC frames composed of packets are mapped into SDH VCs using byte-synchronous mapping. It is used to carry IP data in POS (Packet over SDH) systems and Ethernet frames in Ethernet over SDH systems. LAPS (Link Access Procedure-SDH) is a link access protocol proposed by Dr. Yu Shaohua of the Wuhan Research Institute of Posts and Telecommunications, adopted by ITU-T as standard X.86. It is specifically used for carrying Ethernet frames over SDH links and is very similar to HDLC. GFP (Generic Framing Procedure) is a link layer standard defined in ITU-T G.7041, capable of carrying all data services. It is an open, universal standard signal adaptation and mapping technology that transparently encapsulates various data signals into existing networks. It can replace numerous different mapping methods, facilitating interoperability between equipment from different vendors. GFP employs different service data encapsulation methods for different service types, including frame-mapped (GFP-F) and transparent-transport (GFP-T) modes. GFP-F maps complete service signal frames into variable-length GFP frames without modifying the encapsulated data, supporting packet-granularity rate adaptation and multiplexing. This mode processes after receiving a complete data frame, requiring buffering and media access control, making it most suitable for variable-length packet data such as Ethernet services. GFP-T employs transparent mapping to process data promptly without waiting for the entire frame to arrive, making it suitable for real-time services and block-coded signal formats with fixed frame lengths.
(2) Virtual concatenation: MSTP devices support configurable bandwidth for Ethernet services in the network, achieved through VC concatenation, i.e., combining multiple VC containers into a larger container. VC concatenation in SDH is divided into contiguous concatenation and virtual concatenation. Contiguous concatenation means the VCs carrying Ethernet services are contiguous in the SDH frame structure and share common overhead. If the VCs carrying Ethernet services are independent in the SDH frame structure with flexibly positioned locations, this is called virtual concatenation. Virtual concatenation enables rate adaptation between Ethernet service bandwidth and SDH virtual containers, allowing combination and utilization of small containers at different rates from VC-12 to VC-4. It enables very fine-grained bandwidth adjustment, effectively providing appropriately sized channels for Ethernet services and achieving dynamic bandwidth adjustment. Compared with contiguous concatenation, it better utilizes SDH link bandwidth, improves transmission efficiency, and avoids bandwidth waste. The most critical aspect of virtual concatenation implementation is the transmission of sequence numbers for the VCs participating in virtual concatenation, ensuring the receiving end can reorder and reassemble the VCs of the service signal.
(3) Link Capacity Adjustment Scheme (LCAS): Defined in ITU-T G.7042, LCAS is a function that dynamically adjusts the number of virtual concatenation members without interrupting services. It can flexibly change the bandwidth of virtually concatenated signals to automatically adapt to traffic variations, particularly suited to the dynamic bandwidth requirements of Ethernet services. Together with virtual concatenation, it is an important indicator for measuring whether MSTP bandwidth is effectively utilized. LCAS utilizes reserved SDH overhead bytes to transmit control information, including six states: fixed, add, normal, EOS (end of VC), idle, and not-used. Through the transmission of control information, the number of VCs is dynamically adjusted to meet Ethernet service bandwidth demands. LCAS automatically maps valid payload onto available VCs, avoiding complex manual circuit cross-connection configuration, improving bandwidth provisioning speed without service impairment. In the event of system failures, it can automatically and dynamically adjust system bandwidth without manual intervention; data transmission can remain normal even when one or several VC paths fail. Therefore, LCAS provides MSTP with an end-to-end dynamic bandwidth adjustment mechanism, significantly improving network utilization while ensuring QoS.
3. Performance Analysis of MSTP Carrying and Transporting Ethernet Services
In the transparent transport function of Ethernet services over MSTP, MSTP utilizes TDM mechanisms to assign SDH VCs to Ethernet ports, exclusively occupying the line bandwidth provided by SDH. This provides excellent bandwidth guarantee and security isolation, making it suitable for Ethernet leased-line services with high QoS requirements and core layer applications. However, this approach is based on fixed time slot structures and lacks dynamic bandwidth allocation characteristics. Service granularity is limited by VC, typically a minimum of 2 MB/s, and cannot support flow control, statistical multiplexing of multiple Ethernet services, or bandwidth sharing. It struggles to adapt to the bursty and rate-variable characteristics of Ethernet services, resulting in low bandwidth utilization and lack of flexibility. In practical applications, for a 10M Ethernet service carried and transported over an actual channel bandwidth of one VC-12, the actual throughput does not exceed E1. When large frames are used without reaching the E1 bandwidth limit, no frame loss occurs on the channel; for small frames, even before reaching the bandwidth limit, the short packet length results in low encapsulation efficiency, and the network element's frame processing software cannot keep up with the large number of small frames, causing frame loss. When the bandwidth limit is exceeded, services experience massive frame loss; when large frames reach the bandwidth limit, service transmission delay suddenly increases significantly.
For the approach using Layer 2 switching for Ethernet service access and aggregation, statistical multiplexing of data transmission, bandwidth sharing, and port aggregation can be achieved. User isolation and rate control are implemented through VLANs, and most current MSTP products support Layer 2 switching. Ethernet services are encapsulated and decapsulated at each service node and undergo Layer 2 switching, allowing service nodes to share common transmission channels and saving central office Ethernet interfaces. The Ethernet line card maps service traffic with different 802.1p values to different queues for processing at the port level, implementing priority policies. Rate limiting (such as minimum and maximum bandwidth) can be set based on ports or VLANs, giving the system a certain bandwidth control mechanism, with surplus bandwidth accessed through contention. However, in terms of Ethernet service protection, reliance on the STP (Spanning Tree Protocol) for fault recovery may take tens of seconds, far exceeding SDH's 50ms self-healing protection time, resulting in relatively slow switching rates. Moreover, Layer 2 switching lacks a comprehensive fairness algorithm for bandwidth contention among services of the same priority, making it difficult to guarantee user bandwidth during network congestion, especially in Ethernet ring network applications.
Embedding RPR modules to implement Ethernet ring networks has been accepted by numerous equipment manufacturers. RPR provides a media-independent interface between the MAC layer and physical layer, implemented on MSTP to achieve fair bandwidth allocation for Ethernet services, service priority processing, and improved bandwidth utilization. RPR implements spatial reuse by restricting Ethernet service data flows to bidirectional flow only between source and destination; the destination node strips data packets addressed to it from the ring, thereby releasing bandwidth on the remaining portion of the ring for other packets, improving bandwidth utilization. Dynamic statistical multiplexing methods ensure fairness of bandwidth access for each node; every node on the ring executes an algorithm ensuring each node receives an equal bandwidth share, preventing excessive traffic from one node from causing congestion for other services on the ring. RPR has automatic topology discovery capability, employing an OSPF-like algorithm to exchange topology identification signaling, automatically detecting any Layer 2 topology changes and enhancing the ring's self-healing capability. Additionally, RPR provides carrier-class fast self-healing of less than 50ms for Ethernet services, protecting against faults caused by node failure or link failure.
4. Development Trends of MSTP Carrying and Transporting Ethernet Services
Looking at current Ethernet services, packet lengths continue to decrease and the proportion of small frames is increasing. The shorter the data packets, the lower the encapsulation efficiency of MSTP in processing small frames and the heavier the system's data processing load. Therefore, the capability of MSTP devices to process small frames must be addressed. Meanwhile, in addition to supporting traditional Ethernet services, MSTP will also support new technical standards of data networks, such as GMPLS signaling.
MSTP technology continues to evolve, and future development will enter the stage of intelligent service development, introducing Automatic Switched Optical Network (ASON) functionality. By utilizing an independent ASON control plane for automatic connection management, rapid response to service demands, automatic service provisioning, automatic network topology discovery, and dynamic bandwidth allocation and other more intelligent strategies can be achieved, greatly enhancing MSTP's own flexibility and capability to effectively support data services.
Summary
In summary, the widespread application of MSTP in metropolitan area transmission networks has invigorated these networks and brought greater profit margins for operators. Major equipment vendors are continuously conducting research and development on MSTP, and the connotation of MSTP is being progressively enriched. It is believed that MSTP still has enormous room for development, and the technological energy inherent in the technology itself holds great potential waiting to be tapped. MSTP will play a decisive role in metropolitan area network construction and become the preferred solution for network deployment.
Alternative Interpretation: Multiple Spanning Tree Protocol
MSTP (Multiple Spanning Tree Protocol)
MST (Multiple Spanning Tree)
Multiple Spanning Tree (MST) is an extension of the IEEE 802.1w Rapid Spanning Tree (RST) algorithm.
By employing Multiple Spanning Tree (MST), multiple spanning trees can be established over trunks, associating VLANs with relevant spanning tree instances, with each spanning tree instance having a topology independent of other instances. MST provides multiple data forwarding paths and load balancing, improving network fault tolerance because a failure in one instance (forwarding path) does not affect other instances (forwarding paths).
A spanning tree instance can only exist in bridges with consistent VLAN instance assignments. A group of bridges must be configured with the same MST configuration information, enabling these bridges to participate in a set of spanning tree instances. Interconnected bridges with identical MST configuration information constitute an MST Region.
Multiple Spanning Tree (MST) uses a modified Rapid Spanning Tree Protocol (RSTP), called the Multiple Spanning Tree Protocol (MSTP).
MSTP (Multiple Spanning Tree Protocol)
Pruning the loop network into a loop-free tree network avoids the proliferation and infinite looping of packets in the loop network, while also providing multiple redundant paths for data forwarding, thereby achieving VLAN data load balancing during the data forwarding process. MSTP is compatible with STP and RSTP, and can compensate for the deficiencies of STP and RSTP. It not only enables rapid convergence but also allows traffic from different VLANs to be distributed along their respective paths, thus providing a better load-sharing mechanism for redundant links.
The characteristics of MSTP are as follows:
MSTP sets up a VLAN mapping table (i.e., the correspondence table between VLANs and spanning trees) to associate VLANs with spanning trees; by introducing the concept of "instances" (integrating multiple VLANs into one set), multiple VLANs are bundled into a single instance, thereby saving communication overhead and resource consumption.
MSTP divides a switching network into multiple regions, with multiple spanning trees formed within each region, and these spanning trees are independent of one another.
MSTP prunes the loop network into a loop-free tree network, avoiding packet proliferation and infinite looping in the loop network, while also providing multiple redundant paths for data forwarding and achieving VLAN data load sharing during the data forwarding process.
MSTP is compatible with STP and RSTP.