Summarization of the key technologies of the third generation MSTP

Absrtact: Firstly, the three stages of MSTP technology development are outlined, then the paper introduces the key technologies of virtual cascade, general frame Regulation, link capacity adjustment mechanism and intelligent adaptation layer in the third generation MSTP technology, and introduces and compares the MPLS and RPR technology of the intelligent adaptation layer.

Key words: Metropolitan area Network, MSTP,MPLS,RPR

I. Overview of MSTP

In recent years, the growing number of IP data, voice, image and other services to transport demand makes the user access and network broadband technology quickly popularized, but also promote the transmission backbone of large-scale construction. Because the traffic environment of the service has changed greatly, the metropolitan area network with the main purpose of carrying voice has been unable to meet the requirement of traffic and convergence in capacity and interface capability. Therefore, the multi-service transfer platform (MSTP) technology came into being.

The development of MSTP technology is mainly embodied in the support of Ethernet service, the QoS requirement of the new Ethernet service promotes the development of MSTP. It is generally believed that the development of MSTP technology can be divided into three stages.

The first generation of MSTP is characterized by providing Ethernet point-to-point transmission. It is to map the Ethernet signal directly to the virtual container (VC) in SDH for point-to-point transmission. In the provision of Ethernet transmission leased line business, due to business granularity limited to VC, the general minimum of 2mbit/s Therefore, the first generation of MSTP can not provide different Ethernet services QoS differentiation, traffic control, multiple Ethernet traffic statistics multiplexing and bandwidth sharing and Ethernet business layer protection functions.

The second generation of MSTP is a feature that supports Ethernet two-tier switching. It is a data frame exchange based on the Ethernet link layer between one or more user Ethernet interfaces and one or more independent point-to-point links based on the virtual container of the SDH. Compared with the first generation of MSTP, the second-generation MSTP has made a number of improvements that provide support for multiple Ethernet based on 802.3X traffic control, multiuser isolation and VLAN partitioning, STP based Ethernet business layer protection, and priority forwarding based on 802.1p. Most of the MSTP products currently in use belong to the second-generation MSTP technology. However, compared with the Ethernet business requirements, the second generation of MSTP still exist many deficiencies, such as not providing good QoS support, business bandwidth granularity is still limited by VC, the business layer based on STP protection time is too slow, VLAN function is not suitable for large metropolitan area public network applications, It is not possible to realize the fair access of different position nodes on the ring, and the traffic control based on 802.3x is only for point-to-point link, etc.

The recent advent of the third generation of MSTP is characterized by support for Ethernet QoS. In the third generation MSTP, a number of new technologies are introduced, such as the intermediate Intelligent adaptation layer, the universal frame Regulation (Gfp:generic Framing Procedure) High Speed Package protocol, the virtual cascade and the link capacity adjustment mechanism (LCAS). Therefore, the third generation MSTP can support QoS, multicast, user isolation and bandwidth sharing capabilities to achieve business level agreement (SLA) enhancements, congestion control, and fair access. In addition, the third generation MSTP also has a fairly strong scalability. It can be said that the third generation of MSTP provides comprehensive support for the development of Ethernet business.

Second, third generation MSTP key technology

1. Virtual cascading

The Cascade concept of VC is defined in the ITU-T g.7070, which is divided into two kinds: neighboring cascade and Virtual cascaded. The various VC used in SDH to carry Ethernet business are continuous in the frame structure of SDH, share the same channel cost (POH), this kind of situation is called neighboring Cascade, sometimes it is directly referred to as Cascade. Each VC used in SDH to carry Ethernet business is independent in the frame structure of SDH, and its position can be flexibly handled, which is called virtual cascade.

In principle, cascading and virtual cascade can be regarded as the technology of combining several small containers into a larger container to transmit data services. The rate adaptation between Ethernet bandwidth and SDH virtual channel can be realized by cascading and virtual cascade technology. In particular, the virtual cascade technology, can be from VC-4 to VC-12 and other small containers for the combination of use, can do very small size of the bandwidth regulation, the corresponding cascade maximum bandwidth can also be in a very small range of adjustment. The characteristic of virtual cascade technology is to realize the use of SDH economy to provide the appropriate size channel to the data service effectively, avoid the waste of bandwidth, this is also the most advantage of virtual cascade technology.

2. General Frame procedure

GFP is a link layer standard defined in the ITU-T g.7041  It can transmit variable length packets in byte-synchronized links and transmit fixed-length data blocks, which is a simple and flexible data adaptation method.

GFP adopts a frame-bound method similar to ATM technology, which can transparently encapsulate various data signals, and facilitate the interconnection of multi-vendor devices; GFP introduced the concept of multiple service levels, and realized the statistic Multiplexing and QoS function of user data.

GFP uses different business data encapsulation methods to encapsulate different business data, including Gfp-f and gfp-t two ways. The Gfp-f package method is suitable for packet data, encapsulates the whole packet data (PPP, IP, RPR, Ethernet, etc.) into the GFP load information area, does not make any changes to the encapsulated data, and decides whether to add the load area detection domain as needed. The GFP-T package is suitable for block data with 8B/10B encoding, extracts a single character from a received block of data, and maps it to a fixed-length GFP frame.

3. Link capacity adjustment mechanism

Lcas is a function defined in the ITU-T g.7042 to dynamically adjust the number of virtual cascade without interrupting the flow of data, which provides a way to smoothly change the bandwidth of the virtual cascade signal in the transport network to automatically adapt to the business bandwidth requirements.

Lcas is a two-way protocol that dynamically adjusts the bandwidth of the service by exchanging the control packets between the receiving and sending nodes in real time. The control pack can be represented by a fixed, increased, normal, EOS (which means the VC is the last VC of the virtual Cascade Channel), Idle, and not using six kinds.

Lcas can automatically map the effective net load to the available VC, in order to achieve the continuous adjustment of bandwidth, not only improve the bandwidth allocation speed, no damage to the business, and when the system fails, you can dynamically adjust the system bandwidth, without human intervention, in order to ensure the quality of service under the premise of significantly improve network utilization. Under normal circumstances, the system can be implemented through network management to increase or delete virtual Cascade Group members, to ensure that "do not lose the package", even if the "broken fiber" or "alarm" and other reasons for the virtual Cascade Group member deletion, can guarantee only a small number of packet loss.

4. Intelligent Adaptation Layer

Although Ethernet services are supported in second-generation MSTP, they do not provide good QoS support, one of the main reasons is that existing Ethernet technologies are connectionless. In order to introduce QoS in Ethernet service, the third generation MSTP introduces an intelligent adaptation layer between Ethernet and sdh/sonet, and uses this intelligent adaptation layer to deal with the QoS requirements of Ethernet service. The realization technology of intelligent adaptation layer mainly includes Multiprotocol label switching (MPLS) and resilient Packet Ring (RPR).

(1) Multi-Protocol label Exchange

MPLS, a multi-protocol label Exchange Standard protocol, developed by Cisco in 1997 and made by the IETF, uses layer 2.5 switching technology to organically combine third-tier technologies (such as IP routing) with second-tier technologies (such as ATM, Frame Relay, etc.). This makes it possible to provide point-to-point transmission on the same network, to provide multicast, to provide the best possible services from Ethernet, and to provide real-time Exchange services with high QoS requirements. The MPLS technology uses the tag to unify the upper layer data, so that the different types of packets can be hosted by SDH. The essence of this process is through the introduction of the intermediate Intelligent adaptation layer, the router is marginalized, while the switch is placed in the network Center, the Ethernet's business requirement is adapted to the SDH channel through one route and multiple exchange, and the whole performance of the network is greatly improved by using GFP High Speed Encapsulation protocol, virtual Cascade and Lcas.

The third generation MSTP device based on MPLS not only can realize the End-to-end traffic control, but also has the fair access mechanism and the reasonable bandwidth dynamic allocation mechanism, and can provide the unique End-to-end service QoS function. In addition, by embedding two-layer MPLS technology, allowing different users to use the same VLAN ID, fundamentally solve the VLAN address space constraints. Again, because of the tag mechanism in MPLS, routing calculation can be based on Ethernet topology, which greatly reduces the number and complexity of routing devices, optimizes the transmission efficiency of Ethernet data in MSTP, and achieves optimal configuration and optimal use of network resources.

(2) Resilient packet ring

RPR is the IEEE definition of how to optimize data exchange in the ring topology of the MAC layer protocol, RPR can host Ethernet services, IP/MPLS services, video and dedicated line services, the purpose is to better deal with the problem of circular topology data flow. RPR ring consists of two optical fibers, in the loop on the packet processing, for each node, if the destination of the data flow is not the node, the data stream simply forward, which greatly improve the processing performance of the system. By performing a fair algorithm, each node on the ring can enjoy a fair amount of bandwidth, greatly increasing the bandwidth utilization of the loop, and the switching of business protection on one optical fiber has no effect on the business on the other.

RPR is a new type of Mac protocol, which is specially constructed for ring topological structure, and has the characteristics of flexibility and reliability. It can adapt to any standard (e.g. SDH, Ethernet, DWDM, etc.) of the physical layer frame structure, can effectively transfer voice, data, images and other types of business, support SLA and two-tier and three-tier functions, provide multi-level, reliable QoS service  support dynamic network topology update. The algorithm of OSPF can be used to exchange topological recognition signaling between nodes and has the mechanism of preventing packet death cycle, which increases the self-healing ability of loop. In addition, RPR also has a strong compatibility and good scalability, with TDM, SDH, Ethernet, POS and many other kinds of multirate ports, can host IP, SDH, TDM, ATM, Ethernet and other protocols of the business  can also easily increase transmission lines, transmission bandwidth or insert new network nodes, Good adaptability to new business, protocol, or physical layer specifications that may arise in the future. Again, because the RPR loop each node grasps the loop topology structure and the resource situation, and adjusts the loop bandwidth allocation situation according to the actual situation, therefore the network management personnel does not need to the node resource allocation to carry on the too many intervention, reduces the artificial disposition to bring the human error. RPR enables operators to provide telecommunication service at lower cost in the metropolitan area, which is a very suitable technology for the backbone layer and the convergence layer of the metropolitan Area network.

(3) Comparison between MPLS technology and RPR technology

MPLS Technology and RPR technology have their advantages and disadvantages. MPLS technology breaks through the 4096 address space limit of VLAN in the core node through LSP tag stack, and it can provide good support for Ethernet service QoS, SLA enhancement and network resource optimizing utilization, while RPR technology provides fast packet ring protection for all distributed access, supports dynamic bandwidth allocation, Space reuse and additional business. From the view of the optimization function of the whole metropolitan Area network resources, MPLS technology can optimize the resources from the whole metropolitan area network structure, and achieve the best statistic multiplexing, and RPR technology can only optimize the network resources from the local (within a loop) rather than from the whole network structure. From the complexity of the equipment of the whole metropolitan Area Network, the use of MPLS technology can avoid the introduction of the third layer routing device in the whole metropolitan area, while the RPR device will inevitably introduce the third-tier routing device when it is connected with the loop. From the point of view of protection recovery, although MPLS technology can provide network recovery function, RPR can provide higher network recovery speed.

At present, RPR technology has been adopted by most manufacturers and has a comparative advantage in the market.

Third, the conclusion

As the demand for business types and bandwidth grows further, metropolitan area network must be flexible and reliable, with large capacity and good scalability, support multi-protocol and multiple services, flexible circuit scheduling and business management capabilities, so that operators in the protection of former investment, but also flexible and rapid network expansion and new business, Then reduce operating costs, increase business income, enhance their competitive advantage. In the construction of the metropolitan Area Network, virtual Cascade can provide a more flexible channel capacity organization to better meet the transmission characteristics of the data service; Lcas can provide a series of protocols and steps for dynamically changing the capacity of a transmitting channel; GFP can provide a simple and efficient way to encapsulate data services organized in frames , MPLS and RPR can provide a large amount of data service transfer capabilities. It can be said that the third generation of MSTP provides a very ideal solution for metropolitan Area Network construction.