High Avalibility Design And Implement Case Of IP Packets Fowarding Equipment

Because of the high flexibility of IP(Internet Protocol)technology,communication and telecom systems are transitioning to all-IP networks,including voice bearer networks and data transmission networks.With the increase in the amount of IP network transmission data,the IP data forwarding devices,known as routers,are significant increase in data forwarding capabilities.For example,a typical forwarding capability of a single device has exceeded 10 T bits per second.Similar to the traditional telecom equipment,the failure of the IP data forwarding hardware device will cause the IP network service breakdown.In order to reduce the failure rate and impact of IP data forwarding equipment,equipment manufacturers have proposed a series of solutions to improve the reliability of service,such as a large number of OAM(Operation and Maintenance)protocols.The hardware level standards such as ATCA(Advanced Telecom Computing Architecture)are defined for hardware reliability.These solutions make the IP network system has a certain degree of redundancy,auto-recovery ability,lead to improve the IP network robustness.However,equipment manufacturers and network operators are still beset by some of the key hardware module failure problems.Based on the hardware architecture design and reliability function model of IP data forwarding equipment,this paper identifies several key hardware failures that affect the reliability of IP data forwarding equipment,carries out some high availability solutions,including:1.Isolated power supply block failure of the line processor unit.We carry out a reliability solution of the board-level power supply design,which is based on the parallel redundancy idea using the uc3902 current sharing solution of TI(Texas Instruments).2.Crystal oscillator failure of the line processor unit.A parallel redundancy design is carried out using the AD9574 synthesizer of ADI(Analog Device),and a programmable logic device is designed to help build a high reliability clock system.3.Memory subsystem failure of the line processor unit,including soft error and hardware error.Based on the memory reliability feature of the general-purpose processor,the fault isolation and auto-recovery scheme is designed for the memory subsystem of the forwarding accelerator.The key solutions of IP data forwarding devices above are carried out to improve the robustness of the equipment.The actual effect of the high availability solutions has been verified by fault inject test.The high availability design including the failure mode disclosure and fault-tolerant solutions of the power supply,the crystal oscillator,and the memory subsystem could be used for reference for similar hardware systems such as data-center servers.The hardware infrastructures of the future cloud systems could use the solutions above to improve their reliability and availability.