Emerging as needed
Mobile services have become the most important growing parts of all services provided by telecom carriers. Both the number of mobile users and mobile revenue has increased sharply. In China, China Mobile has made a net profit of $66 billion through over 300 million users in 2006, a 22.1% increase on that of 2005. In Europe, Verizon's mobile service revenue has met an increase of $5 billion in 2005. The company's revenue of mobile services including wireless broadband services has increased by 16.7%, becoming one important drive force for its overall revenue increase. As a result, Verizon took the second place of Total Telecom Global 100.
Aside from the fast increase of users, service modes of mobile communications have also been expanding from voice communications in the past to mobile Internet access, mobile video and so on. It becomes very obvious that mobile Internet is more individualized and multi-media driven, which is represented by the transition from 2G to 3G.
However, such quick developments have posed austere challenges to network carriers. On the one hand, the sharp increase of users and the transition to broadband made the demand for bearer network's bandwidth by wireless core network soar. Only several years in the 21st century, the accessing bandwidth of RAN evolved from K bits to M bits, and that of core bearer network evolved from M bits to G bits. But the evolutions have been flooded by users' ever-growing demands for bandwidth. Operators bear unprecedented CapEx as well as OpEx pressures. On the other hand, telecom carriers traditionally follow a mode of tight coupling. Each service is implemented with one bearer network, which carries only a certain kind services. Considering the continuous emergence of new services, this mode will confine carriers to an infinite loop of continuous reconstructions. Evidently, the traditional TDM/ATM bearer network is obsolete, and then, what is the way out for bearer networks?
Compared with TDM/ATM, IP is more flexible. It can be realized through any existent physical network and does not waste carriers' cables or lines at all. As a typical packet switched network, IP is born with higher efficient bandwidth usage compared with TDM/ATM. This advantage is upgraded by link and package compressing technologies; therefore, IP network is so economic as to meeting the demand for upgrading bandwidth. The open structure of IP enables easy development of various IP-based protocols and services, including data services and video services. Most of the present mobile core network services are already carried by IP, and those services that are still carried by TDM/ATM will get IP interfaces through Package Control Unit (PCU) when R5 begin. When all the services are transmitted in a packet switching network, the upper-layer protocol packages are transparent on the IP bearer network. IP/MPLS technologies can guarantee traffic interconnecting among different equipment. In addition, IP protocols offer peerless support for multicasting. Therefore, IP bearer network is the most economic and reasonable choice for fulfilling the demands for mobile services and realizing multi-service carrying.
Meeting expectations
Traditional IP network only carries Internet services via non-connection and best-effort service modes. It takes much time for self-healing and is relatively vulnerable, and thus cannot meet the requirements of carrier-class mobile voice and video services. In order to use IP network as mobile bearer network, special optimization and improvement must be done in high usability, point-point QoS guarantee and so on, in addition to IP address planning, IGP routing deployment and delicate design of MPLS.
Voice is the foundation and core service of mobile services. For IP network, voice service poses relatively hard requirement of high usability. The following table shows influences of network intermittent blinking on conversation quality.
Table 1 Influences of intermittent blinking on voice servic
Recovery Time |
Influences on Voice Services |
< 50 ms |
No influence |
50 ms to 200 ms |
No influence on signaling if the connection loss probability is less than 5% |
> 2s connection loss threshold) |
Disconnection of voice communications and dedicated line connections |
High availability of IP bearer network can be guaranteed via three types of technologies: equipment reliability technologies, network fault fast detection technologies, and network fault fast protection switching technologies. The three types of technologies should be combined so as to guarantee high availability of IP bearer network.
Carrier-class services require over 99.999% network equipment availability (that is, the unusable service duration each year should be less than 5 minutes). Therefore, mobile bearer network equipments usually adopt control/forward plane splitting and distributed switching architecture. Redundant design is made for key components like main control board, power and fans. These equipments also support hot swap and non-breakage of standby switching services via NSF technologies. Moreover, online upgrading and hot patching are supported through componentization of operating system.
Fast detection and location of fault are necessary for fast protection switching. At present, mainstream fast fault detection technologies are:
- POS port detection: The POS interfaces perceive fault alarms from SDH transmission equipments and detect the faults on directly connected physical ports within 5ms.
- MPLS OAM: equipments can detect connection faults of MPLS LSP within 100ms via MPLS OAM technology.
- BFD: Huawei's routers support the correlations among BFD /MPLS LDP, multiple routing protocols and logic interfaces. Fast fault detection can be made on various connection states, for example links, tunnels, MPLS LSP, and network protocol neighbor relationship. The detection duration can be as short as 30ms. When multiple fault detection technologies are used at the same time, the detection principle is to start from the bottom-layer link fault detection technology. When the bottom-layer fault detection is not applicable, upper-layer network fault detection technologies can be used. Multiple fault detection technologies can be used together to quickly detect faults.
Network fault fast protection switching technologies mainly include IGP FC technology, IP/LDP FRR technology, iteration on demand and by priority technologies of VPN routes, VPN FRR technology and TE FRR technology. By optimizing algorithms of routing protocols, flood method, timer and so on, IGP FC technology greatly reduces the convergence time when network topology changes. Convergence time can be reduced to less than 1s when IGP FC technology is used through the whole network (the specific time depends on the number of network nodes and the complexity of network topology). IP/LDP FRR technology presets and makes backups of next hop or forwards interfaces to switch service traffic to backup links within 50s after faults are detected via fault fast detection technology. Iteration of VPN routes by need and by priority technology quickly completes the convergence of key services or key customer VPN routes according to users' settings, and thus greatly improves the availability of VPN routes. Huawei's VPN FRR technology can switch to backup remote PE within 200ms when working remote PE fails, and the switching speed is not affected by the number of private network routes. As in backbone network, TE FRR and Tunnel Backup technologies can enable fast protection switching among links, nodes and p2p LSP. Huawei's TE Switching solution also greatly reduces configuration complexity of TE Auto FRR.
The comprehensive application of high-availability technologies, fault fast detection technologies and fast protection switching technologies keeps the end -to-end protection switching time of IP bearer network within 200ms, which meets the requirements by services including voice services for high-availability of bearer network.
Table 2 High-availability technologies
Technology |
Technology Mode |
Application Areas |
Effect |
Key parts redundancy |
Equipment reliability |
Main control board, switching network board, power supply, and fans |
Normal working in case of components faults (with alarming) |
Parts hot swap |
Equipment reliability |
Cards, modules, power, and fans |
No power-down and restart in case of changing parts |
Control/forward plane splitting and distributed switching architecture |
Equipment reliability |
Carrier-class routing and switching equipment |
No affects on other cards or modules in case of changing boards or modules, easy to switch main equipments, NSF and so on |
NSF technology |
Equipment and network reliability |
Carrier-class routing and switching equipment |
No service interruption and network topology change in case of active/standby switching |
Operating system componentization |
Equipment reliability |
Carrier-class routing and switching equipment |
Hot patching and online upgrading |
Interface detection |
Fast fault detection |
Physical interface state-inspection |
< 5 ms fault detection |
MPLS OAM |
Fast fault detection |
LDP, LSP state-inspection |
< 100 ms fault detection |
BFD |
Fast fault detection |
State-inspection of physical interfaces, logic interfaces, links, tunnels, MPLS, LSP, and protocol neighborhood |
30 ms to 1000 ms fault detection |
IGP FC |
Fast protection switching |
Whole network configuration |
S-level routes convergence |
IP/LDP FRR |
Fast protection switching |
CE dual homing to PE |
Ms-level protection switching |
VPN FRR |
Fast protection switching |
Remote PE protection switching |
Ms-level protection switching |
BGP/VPN fast convergence |
Fast protection switching |
Iteration on demand and by priority of VPN routes |
Ms-level protection switching |
TE FRR |
Fast protection switching |
Backbone network link, node protection or end-to-end tunnel protection |
Ms-level protection switching |
3GPP classifies mobile network services into four types: "conversational", "streaming", "interactive", and "background". Elements affecting the quality of the four types of services are time delay, delay jitter, and packet loss. The default indexes in engineering implementation in the industry are: the end-to-end delay to be less than 400ms (including terminal and MGW processing delays); the delay jitter to be less than 20ms; the packet loss ratio to be less than 1%. In order to meet the QoS requirements of mobile services, the IP bearer network needs not only optimize design by traffic analysis and ensure light load, but also the help of equipment traffic control technologies, Diff-Serv, MPLS TE and DS-TE technologies.
The most basic method of ensuring QoS is to avoid congestion. In planning an IP bearer network, first design a detailed traffic model, and then design the IP bearer network on the basis of the traffic model. Make sure the network traffic is reasonably distributed and reasonably backed up in case of faults. In general, make sure the link bandwidth peak usage rate of the IP bearer network should not be over 70%; all important service flows should have backup bandwidth in single-point faults.
Light-loading design cannot always prevent congestion. Once congestion happens, control traffic at the network edge, and at the same time, adopt the CAC technology to send the busy tone to users who are temporarily cut off from services. Meanwhile, the Diff-Serv must ensure preferential transmission of high-priority services.
MPLS TE can set up a tunnel across the backbone network according to users' requirements and network resources. It automatically implements maintenance, statistics, property modification (e.g. bandwidth) and backup of the tunnel via signaling negotiation. Carriers can use the MPLS TE technology to precisely control traffic transfer paths, reserve resources and avoid congested nodes. In this way, existing bandwidth resources can be fully used.
Diff-Serv can only reserve resources at a single node but cannot guarantee QoS in the whole path. To reserve resources of the whole path in the differentiated service network, the Diff-Serv technology can be used together with MPLS TE technology to enable DS-TE. Different bandwidth restrictions can be implemented on different types of flows. Each flow can be dynamically adjusted according to bandwidth restriction of the flow.
In addition to the above-mentioned technologies, other technologies like sampling, compression, package and link optimization can be adopted to further improve the efficiency of network bandwidth. In the considerable commercial use of network, quality of voice in IP bearer network is not worse at all than traditional TDM switching network, which suggests that IP bearer network can fully meet the rigid QoS requirements of mobile services.
Technology |
Application Scope |
Effect |
Light-loading design |
Access network, whole backbone network |
Peak usage rate of link bandwidth not over 70%, meanwhile, all key services with backup bandwidth in case of single-point faults |
MPLS TE |
Backbone network end-to-end |
Automatic end-to-end maintenance, statistics, attribute modification, and backup according to users' requirements and network resources |
Diff-Serv |
Ingress PE, independent equipment configuration |
Ensuring preferential forwarding of high-priority services, Traffic control at ingress |
DS-TE |
Backbone Network end-to-end |
Building end-to-end tunnel,
Bandwidth restrictions over different flows,
Dynamic adjustment of traffic according to bandwidth |
Prospect
Multi-service IP bearer network is the first choice for operators to build new networks and their target to rebuild their traditional bearer networks. In recent years, Huawei has participated in or undertaken carriers' network building and rebuilding projects in many countries and regions all over the world. China Mobile IP private bearer network, the largest mobile IP bearer network in the world, has been used for three years. It provides many services including voice services, short message, ring tone, multimedia message, GPRS, WAP and so on. It has stood the tests of harsh natural environment and the rush hours of traditional Chinese reunion festivals like the Spring Festival and the Mid-Autumn Festival. This demonstrates the strong capability as well as high reliability of mobile IP bearer network. It is promising that, with its prominent open nature and strong bandwidth expansion capability, IP bearer network shall play a very important role in the upcoming development of mobile services.
Related literatures: