ZTE´s Solution and Evolution Strategy for Optical Transport Network

ZTE’s DWDM system can support multiple service interfaces such as SDH, SONET and GE. Its capacity can be smoothly upgraded from 400 Gb/s to 800 Gb/s, and even up to 1.6 Tb/s. Its wavelength range covers C & L wave bands. It can provide transmission beyond 5 000 km through optical fibers G.652 and G.655 without any electrical regeneration. The incorporated technologies include mixed amplification through distributed Raman and EDFA (Erbium-Doped Fiber Amplifier), super out-band FEC (Forward Error Correction), NRZ (Non-Return to Zero) and RZ modulation codes, dynamic power equalization, and distributed dispersion management.

1 Introduction
The turn of the century has witnessed a burst in the planet-wide telecommunication bubble. The market for optical networks is also known to be on the decline. Operators are now seen to be beginning to change their outlook from blind pursuit of greater network capacity and higher bandwidth to financial indexes. Increasing attention is being given to cost-performance index of the system equipment, incorporating of multi-service solutions, and automating the deployment and dispatching function aided by an intelligent network management that helps save on manual labor costs.

  Nowadays, both old operators who have modified their networking strategies and emerging ones choose next generation SDH or metro area WDM technologies and service layer equipment to jointly build their networks. ZTE Corporation thinks that next generation SDH has great vitality, and stating that "SDH is degenerating from a system to an interface" is a hasty deduction. Next generation SDH has two strengths: multi-service and intelligence. Evolution from SDH to MSTP (Multi-Service Transport Platform) suggests multi-service, while intelligence means that GMPLS/ASON (Generalized Multi-Protocol Label Switching/Automatically Switched Optical Network) control planes are used to implement flexible bandwidth allocation, automatic end-to-end service dispatching and dynamic protection and restoration functions. All along WDM was required to provide just large capacity and transparent transport. However, under the changing circumstances, new generation WDM equipment is required to provide quick service access and optical layer protection besides large capacity, high bandwidth and greater service transparency.

2 System Equipment
All globally renowned optical communication system vendors provide DWDM systems with high capacity and ultra long haul transmission capability. ZTE is no exception. ZTE’s DWDM system can support multiple service interfaces such as SDH, SONET and GE. Its system capacity can be smoothly upgraded from 400 Gb/s to 800 Gb/s, and even up to 1.6 Tb/s. Its wavelength range covers C & L wavelength bands. It can provide transmission beyond 5 000 km through optical fibers G.652 and G.655 without any electrical regeneration. The incorporated technologies include mixed amplification through distributed Raman and EDFA (Erbium-Doped Fiber Amplifier), super out-band FEC (Forward Error Correction), NRZ (Non-Return to Zero) and RZ modulation codes, dynamic power equalization, and distributed dispersion management. The characteristics of distributed Raman amplifier include higher gain bandwidth, gain flatness, and automatic gain spectrum adjustment based on signal distribution. Owing to its vast engineering expertise in ultra long haul WDM transmission, ZTE was granted the right to draft G.665, a new standard for the Raman amplifier at ITU-T SG15 (optical and other transmission networks) in October 2003. The standard has now been approved.

  With the highest bandwidth of 400 Gb/s, metro WDM equipment provided by ZTE can offer multi-service OTU (Optical Transponder Unit), T-MUX (Transparent Multiplex), optical multiplex section protection ring and optical channel shared protection ring. It is worth noting that ZTE’s metro WDM equipment switching structure is the most complete while providing perfect protection switching performance along with utmost reliable protection switching ability, thereby making it second to none in the industry. An optical switch with serial/parallel mixed design is used for ZTE’s metro area Optical Add/Drop Multiplexer (OADM) equipment. Making use of  ZTE’s unique "upgrade interface", the metro area OADM equipment can truly ensure "online upgrade" without interruption of services already in use. Moreover, based on the application requirement of metro area access networks and following the new ITU-T standards, G.694.2 and G.695, ZTE has launched its compact Coarse Wave Division Multiplexing (CWDM) equipment with competitive cost-performance ratio. It can provide users with quick capacity expansion and network optimization capabilities.

  ZTE’s SDH-based MSTP equipment can implement multiple rates, including 155 Mb/s, 622 Mb/s, 2.5 Gb/s and
10 Gb/s. On one hand, the MSTP equipment retains the cross-connect capability inherent in SDH and traditional SDH/PDH service interfaces to continue meeting demands of Time Division Multiplexing (TDM) services; whereas, on the other hand, it also supports ATM, transparent transmission on Ethernet, Ethernet L2 switching, Resilient Packet Ring (RPR) and Multiple Protocol Label Switching (MPLS) to meet demands on convergence, grooming and consolidation of data services. To begin with, ZTE’s MSTP equipment uses Generic Framing Procedure (GFP) to guarantee excellent encapsulation. In addition, its virtual concatenation and Link Capacity Adjustment Scheme (LCAS) is adapted to diversify bandwidth granularity and make link capacity adjustment across a certain range.

  Besides Ethernet functions, the RPR functional module has overcome the shortcoming of slow switching speed of Ethernet and has accomplished quick protection switching time not exceeding 50 ms. RPR also provides fairness algorithm to guarantee optimal utilization of link bandwidth and also avoid link congestion. With MPLS features, the networking can be extended from a ring to a grid. Pseudo Wire (PW) is used to implement multi-service access and convergence at the user end, while data is converged into the core data network through tunnels. In this way, all-round or all-network MPLS implementation is finally achieved, which brings the power of MPLS into full play. The private leased line network service for key clients is always an important source of operators’ earnings. The MSTP equipment helps implement point-to-point, point-to-multipoint and multipoint-to-multipoint networking to develop VPN services such as EPL (Ethernet Private Line), EVPL (Ethernet Virtual Private Line), EPLAN (Ethernet Private Local Area Network) and EVPLAN (Ethernet Virtual Private Local Area Network), and provide proper CoS (Class of Service) and QoS according to clients’ requirements. In particular, L2 VPN (Virtual Private Network) (such as Virtual Private LAN Services) that makes use of MPLS functions has a much better cost-performance ratio than the traditional TDM private line.

3 Evolution Strategies
For a long time, traditional optical networks only implemented transmission, multiplexing, cross connection, monitoring and survivability of signals on the user layer. They did not possess switching functionality, and so were less intelligent, i.e., traditional optical networks had nothing "intelligent" about them. If service layer networks are reviewed, "switching" is found to be the basis, no matter if we consider fixed networks or mobile networks.

  Consequently, intelligence without the concept of "dynamic switching" is not real intelligence. Therefore, the introduction of "dynamic switching" into traditional transport networks is a historic breakthrough in the long-term conceptualization of transport networks as well as an important revolution in transport network technology. The transport network, however, need not have its own signaling and routing protocols to support its intelligence. It may make use of related protocols of the fixed data network, such as GMPLS.

  ZTE believes that no matter if we consider fixed switching networks, mobile switching networks, fixed data networks, or optical transport networks, the implementation of intelligence requires introduction of a control plane or control entity into the networks. The core of the control plane/entity must be software, and it is implemented using multiple signaling and routing protocols. As for optical networks, equipment vendors use proprietary protocols to implement the control plane. Proprietary protocols, however, turn out to be barriers to interconnection of products from different vendors, various sub-networks, different operation domains and diversified management domains. Therefore, it is absolutely essential to standardize intelligent optical networks. We have seen that it is just standardized protocols that have helped the large-scale commercial use of intelligent networks of PSTN, GSM, GPRS, CDMA and CDMA 1x. Without globally unified standards, it is impossible for any intelligent network to be put into large-scale commercial use. Fortunately, now we can gladly witness that ITU-T is taking a great effort to make ASON (Automatically Switched Optical Network) standards, while IETF (Internet Engineering Task Force) is developing GMPLS standards, and OIF (Optical Interconnection Forum) is doing its best to make UNI (User Network Interface) standards.

  Intelligent service-layer networks can directly provide end users with services, such as online banking, automatic accounting cards, voice mailboxes and prepayment. In the new scenario, the service that the intelligent optical networks can offer is bandwidth, if at all it can be regarded as a service. They can provide leased bandwidth, wholesale, and so on. They can also implement OVPN (Optical Virtual Private Network) and BoD (Bandwidth on Demand) services depending on users’ demands. Due to the distinctive nature of optical networks, the services they provide cannot be directly used by end users. But incumbent operators can provide such services to new ones; big operators can provide them to small ones; and long haul network operators can provide them to metro area operators. We still clear heads to understand that the early application goal of intelligent optical networks would be to provide quick protection and restoration when there are some troubles and to use standardized signaling and protocols to implement  "end-to-end service deployment", rather than directly offer services.

  Currently, physical connection, also called "hard connection" or "permanent connection" is the only way to connect optical networks with user-layer networks (including traditional PSTN switches, ATM switches, IP routers and image processing equipment). In this way, optical networks just mechanically transport signals from one end to the other. Once such a bearer channel is established, it will remain stable for several months, for half a year, for one year, or even longer. We can take the SDH network as an example. The permanent connection of traditional SDH circuit deployment is actually implemented by manual configurations in the network management system, which not only wastes time (it may take several days) and human resources (that includes a few experienced maintenance and commissioning professionals), but also has low efficiency and is difficult to modify after the deployment is finished.

  However, intelligence means "soft connection" between user-layer networks and optical networks. A
user-layer network should first request an optical network for its bandwidth needs, and then the optical network will quickly respond and provide an optimum transmission channel in a short time. Such connection can not only change the route according to needs, but can also be broken and rebuilt.

  ZTE has developed ASON-based intelligent optical network equipment with the goal of changing traditional permanent connection into soft permanent connection, and even into switching connection, to allow user equipment to initiate a bandwidth request through UNI according to its own requirements. ZTE deploys multiple functional units in the control part of its intelligent optical network, including call controller, connection controller, routing controller, protocol controller, strategy controller, link resource management unit, locating agent and termination adaptor. All control parts have strictly divided tasks but cooperate closely to jointly implement intelligent control functionality.

  Control units distributed to different points communicate each other with I-NNI or E-NNI protocols to quickly establish a connection channel and a bearing channel for the service-layer network in real time. There is no doubt that established channels can be set free and terminated at any time, and can be replaced by new connection channels if they have some faults. As for the network management system of the optical network, both the transport plane and the control plane require management. For example, transmission impairments of the transport plane such as errors, jitter and drift are reported to the network management system, while troubles of the control plane such as those of the signaling network, call failure, connection failure and time-out also need reporting to the network management system. Since the intelligent control layer is added, "Fault management", one of the five management functions of the network management system, may be weakened greatly. ZTE adopts GMPLS as the main protocol of its intelligent optical network equipment. Besides, aiming at the characteristics of the optical network, original MPLS architecture has been expanded in functionality. And the signaling protocol prefers RSVP-TE, while the routing protocol is OSPF-TE and DDRP.

  ZTE also believes that existing user facilities in the network are working maturely and stably, so it is unrealistic to rebuild them for the introduction of intelligent optical networks. The solution is an "intelligent agent". For example, a user device is connected to UNI-C via a simple and easy-to-operate interface, and then the UNI-C is connected to the intelligent optical network via a standard UNI-N interface. In this way, there will be minimum changes to the existing user devices. This is also true for the connection between the traditional optical network and the intelligent optical network. Generally speaking, it is a long evolution from existing optical networks to intelligent optical networks. There is no way to implement an intelligent optical network in a second.

Manuscript received: 2004-12-06