Study of Multi-nodes Failover Used in HA Cluster

WANG Ming-wei,YIN Kang-kai,LI Shan-ping

DOI: https://doi.org/10.3969/j.issn.1001-3695.2005.03.026

2005-01-01

Abstract:Multi-nodes model is introduced. The test results show that the failover technology works well and can quickly and exactly remove the tasks of any invalid working node to the copied one to keep the clustering service continuous.

What problem does this paper attempt to address?

Study of Multi-node Heartbeat Model Used in HA Cluster

YIN Kangkai,Wang Mingwei,LI Shanping

DOI: https://doi.org/10.3969/j.issn.1000-3428.2005.15.037

2005-01-01

Abstract:A multi-node heartbeat model is introduced. It includes several sub models such as trust model, vote model and neighborhood model.The focus is on the multi-node heartbeat model design and implementation. The test result shows that this model works well in small scale clusters.
HAaaS: Towards Highly Available Distributed Systems.

Yaoguang Wang,Weiming Lu,Bin Yu,Baogang Wei

DOI: https://doi.org/10.1109/cluster.2012.59

2012-01-01

Cluster Computing

Abstract:High availability is a valuable property in distributed systems. The master-slave model is used wildly in data management systems for high performance. However, many master-slave systems still have SPOF (Single Point of Failure) for the single master node. We exploit a generalized solution to meet several common use cases for different master-slave systems. The solution makes the high availability as a service (HAaaS), which uses a shared storage infrastructure to make the master stateless and provides an automatic fail over of high-availability service. We deploy the HAaaS in many master-slave subsystems in our unstructured data management system (UDMS) to make the UDMS highly available. The experiments demonstrate the feasibility and efficiency of our solution.
Reliability of Wireless Sensor Networks Based on Redundancy of Cluster-Heads

FENG Dong-qin,LI Guang-hui,QUAN Jian-min,JIN Jian-xiang

DOI: https://doi.org/10.3785/j.issn.1008-973x.2009.05.012

2009-01-01

Abstract:In order to improve the reliability and availability of industrial wireless sensor network(WSN),and to make WSN autonomously work for a long lifetime,a cluster-based routing algorithm for industrial WSN based on the redundancy of cluster-heads was given.When the master cluster-head has not enough energy or has a failure,the redundant cluster-head can take over the data transmission tasks.The marginal nodes use the backup cluster-head and a multi-cluster access scheme to tolerate the failure of cluster-heads.For the isolated nodes that cannot communicate with any cluster-head directly,a multi-hop routing technique based on the probability model was put forward to maintain the normal communication.In addition,the presented algorithm sets the redundant cluster-head into the light-weighted sleeping state,and closes the front end radio frequency components when the redundant testing cycle comes so as to reduce the energy consumption.With the presented sensor nodes,an experimental environment was setup and three experiments were implemented,including cluster-head switching,marginal nodes and isolated nodes communication mechanisms.Experimental results showed that the proposed cluster-based routing algorithm has higher reliability than that of the single cluster-head network model.
NCluster: Using Multiple Active Name Nodes to Achieve High Availability for HDFS.

Zhanye Wang,Dongsheng Wang

DOI: https://doi.org/10.1109/hpcc.and.euc.2013.329

2013-01-01

Abstract:Hadoop HDFS is an open source project from Apache Software Foundation for scalable, distributed computing and data storage. HDFS has become a critical component in today's cloud computing environment and a wide range of applications built on top of it. However, the initial design of HDFS has introduced a single-point-of-failure, HDFS contains only one active namenode, if this namenode experiences software or hardware failures, the whole HDFS cluster is unusable until the recovery of namenode is finished, this is the reason why people are reluctant to deploy HDFS for an application whose requirement is high availability.In this paper, we present a solution to enable the high availability for HDFS's namenode through efficient metadata replication. Our solution has two major advantages than existing ones: we utilize multiple active namenodes, instead of one, to build a cluster to serve request of metadata simultaneously. We implements a pub/sub system to handle the metadata replication process across these active namonodes efficiently. Based on the solution we implement a prototype called NCluster and integrate it with HDFS. We also evaluate NCluster to exhibit its feasibility and effectiveness. The experimental results show that our solution performs well with low replication cost, good throughput and scalability.
Design and implementation of heartbeat in multi-machine environment

Zonghao Hou,Yongxiang Huang,Shouqi Zheng,Xiaoshe Dong,Bingkang Wang

DOI: https://doi.org/10.1109/AINA.2003.1192949

2003-01-01

Abstract:Heartbeat mechanism is widely used in the high availability field of monitor network service and server nodes. In this paper, we design a distributed heartbeat mechanism, which consists of one Master node and multiple Standby nodes, all running the same process of heartbeat but in different states. We experimentally demonstrate that the mechanism improves the scalability over that of the dual computer nodes and supports nice fail-over. Moreover, the process of the heartbeat only needs to maintain few states, which dramatically reduces the overheads of the system and simplifies the management of the nodes. When there are only two nodes in the network, the mechanism will degenerate to dual computer heartbeat. In current version, the mechanism only supports the switch-based architecture; it is easy to make it supporting other architectures such as Ring, etc.
Smart Ring: A Model of Node Failure Detection in High Available Cloud Data Center.

Lei Xu,Wenzhi Chen,Zonghui Wang,Huafei Ni,Jiajie Wu

DOI: https://doi.org/10.1007/978-3-642-35606-3_33

2012-01-01

Abstract:Nowadays most of cloud data centers deploy high available system in order to provide continuous services, so it's very important for a high available cluster to detect the node failure (physical machine failure) accurately and timely in a low bandwidth occupation way. However, compared to the traditional cluster environment, the scale of cloud data center increases rapidly with the use of virtualization, so traditional node failure detection models have already faced several new problems. In this paper, we present a three roles and two layers node failure detection model, named as Smart Ring, which fits cloud data center well and strikes a balance between accuracy, instantaneity and bandwidth occupation. It can simultaneously detect the status of physical machines and virtual machines and deal well with multiple nodes failure and network partition. Our experiment results show that Smart Ring has a better performance than most existing models. © IFIP International Federation for Information Processing 2012.
The Application of Multi-Path Fault Tolerant Algorithm in WSN Nodes

Zhao Jun,Chen Xiang-guang,Xie Ying-xin

DOI: https://doi.org/10.1109/aimsec.2011.6010112

2011-01-01

Abstract:In order to ensure proper work for the end-to-end multiple hops in wireless sensor network (WSN) when some nodes fail without any redundant nodes, based on an airport oil pipeline as the research object, the wireless sensor network that the nodes distributed in a line is studied and the multi-path fault tolerant routing algorithm and the MAC protocol are proposed in this paper. Every node can monitor the operation states of neighboring nodes by the method of mutually detecting. We discuss the differences of network throughput when the states of the sensor node are proper and fail. Field experiments show that the fault tolerant multi-path algorithm proposed can solve the hidden terminal and data collision problems effectively and can be used in the low data rate wireless sensor network.
Research and Implementation of High Availability in Iscsi Multi-level Storage Systems

YU Shengsheng,ZHOU Xin,ZHOU Jingli,ZHANG Jun

DOI: https://doi.org/10.3969/j.issn.1000-3428.2005.05.073

2005-01-01

Abstract:The paper builds a theoretical HA model based on binary component model and Markov process. A quantitative analysis on improving system availability is made. An HA solution is proposed upon the iSCSI multi-level storage system , and a test result is given and analyzed.
A New Method to Improve the Dependability of Key Nodes in Pervasive Computing Environment

Hongwei Wang,Zhibo Wu,Hongwei Liu,Jian Dong

DOI: https://doi.org/10.1109/PCSPA.2010.121

2010-01-01

Abstract:Pervasive computing has been a hot point in researches for decades which integrates digital and physical devices in everyday life. The availability of key nodes which provide service is momentous. HA Cluster computing systems have become significantly practical for being server node. The availability of heartbeats, a main fault detection method, plays a momentous role in existing HA systems. However, the HA system may work on incorrect status when a fault exists in heartbeat links in the HA architecture. This paper presents a HA system with an independent monitor-based mechanism which monitors the leader nodes and manage the system to avoid the fault operation resulting from failures of heartbeat links. Its architecture is discussed, and the dependability of the system is analyzed by a continuous time Markov chain model; the dependability of a HA system and the HA system with an independent monitor module are compared. The results show that the proposed method is reasonable and will significantly improve the availability of the server node to fulfill the requirement in pervasive computing environment.
Multi-Cluster Load Balancing Based On Process Migration

Xiaoying Wang,Ziyu Zhu,Zhihui Du,Sanli Li

DOI: https://doi.org/10.1007/978-3-540-76837-1_14

2007-01-01

Abstract:Load balancing is important for distributed computing systems to achieve maximum resource utilization, and process migration is an efficient way to dynamically balance the load among multiple nodes. Due to limited capacity of a single cluster, it's necessary to share the underutilized resources of other sites. This paper addresses the issues in multi-cluster load balancing based on process migration across separate clusters. Key technology and mechanisms are discussed and then the implementation of a prototype system is described in detail. Experimental results depict that by achieving multi-cluster load balance, surplus resources can be efficiently utilized and the makespan is also greatly reduced as a result.
A Novel Virtual Node Migration Approach to Survive a Substrate Link Failure

Bingli Guo,Chunming Qiao,Yongqi He,Zhangyuan Chen

DOI: https://doi.org/10.1364/nfoec.2012.jth2a.43

2012-01-01

Abstract:A virtual network mapping approach to protect against a link failure by allowing virtual nodes to migrate among their primary locations only is proposed and shown to be superior to previous approaches.
Fault-tolerant Mechanism of the Distributed Cluster Computers

Shang Yizi,Jin Yang,Wu Baosheng

DOI: https://doi.org/10.1016/s1007-0214(07)70107-4

2007-01-01

Abstract:The distributed system with high performance and stability is commonly adopted in large scale scientific and engineering computing. In this paper, we discuss a fault-tolerant mechanism under Linux circumstance to improve the fault-tolerant ability of the system, namely a scheme and frame to form the stable computing platform. In terms of the structure and function of the distributed system, active list and file invocation strategies are employed in the task management. System multilevel fault-tolerance can be achieved by repeated processes in a single node and task migration on multi-nodes. Manager node agent introduced in this paper administrates the nodes using the list, disposes of the tasks according to the nodes' performance, and hence, to be able to make full use of the cluster resources. An evaluation method is proposed to appraise the performance. The analyzed results show the usefulness of the scheme proposed except for some additional overhead of memory consumption.
Design of a Dual-Computer Cluster System and Availability Evaluation

Guo Hui,Zhou Jingli,Li Yue,Yu Shengsheng

DOI: https://doi.org/10.1109/icnsc.2004.1297462

2004-01-01

Abstract:Considering high-availability of the storage system has become a highly desirable property for network attached storage systems, we develop a special software application supporting the dual-computer fault-tolerant scheme. In this paper, we describe the design and implementation of a dual-computer fault-tolerant system. It includes two fundamental techniques: data mirroring and failover. The data mirroring technique is based on logging and tracking scheme. The failover is realized by heartbeat detection and automatically take over functions. According as availability definition, the theoretical analysis for the dual-computer system has been presented. The finds show that the integration of data mirroring and failover techniques can provide higher availability for the network based storage systems. Network managers can turn to this dual-computer scheme as the solution to resolve the increased demand on the file system availability at file server level.
Research on back-up node fault tolerance pattern in WSN

Chuang Ma,Zhibo Wu,Hongwei Liu,Xiaozong Yang

2009-01-01

Abstract:This research introduces back-up node in fault-tolerant wireless sensor networks design. It is expected that a node failure will not affect the normal operation of the entire network. Based on the use of heartbeat detection algorithm to detect node failure, directed diffusion protocol is introduced to extend the working hours of a wireless sensor network.
Research of Scheduling Strategy Based on Fault Tolerance in Hadoop Platform

zhengwu yuan,jinli wang

DOI: https://doi.org/10.1007/978-3-642-41908-9_52

2013-01-01

Abstract:The scheduling problem is a hot issue in the current cloud computing, and node failure conditions should be taken into consideration. Firstly the disadvantages of current Hadoop task scheduling algorithms and fault tolerance in Hadoop platform are discussed in this paper. Then a scheduling strategy based on fault tolerance is presented. According to this strategy, the cluster detects the speed of the current nodes in a cluster, and makes some backups of the intermediate MapReduce data results to a high-performance cache server, and the data is produced by the node that may go wrong soon. Thus the cluster may resume the execution to the previous level quickly when there are several nodes going wrong, the Reduce nodes read the Map output from the cache server or both of the cache and the node, and keeps its high performance. Finally the computer simulation is done. It shows that the strategy presented in this paper is effective under failure tolerance conditions.
Design and implementation of an efficient multipath for a SAN environment

Jun Luo,Ji-wu Shu,Wei Xue

DOI: https://doi.org/10.1007/11576259_12

2005-01-01

Abstract:Multipath provides multiple paths between application Servers and storage devices. Multipath can overcome single point of failure, and improve a system’s reliability and availability. This paper presents a multi-layer Multipath, and describes the design and implementation of a Multipath system in a storage area network (SAN). For an application server, we implemented Multipath in the volume management layer. For a storage server, we implemented Multipath in the SCSI Middle Level layer. This system can make the most use of the storage server’s characteristics to decrease the time of failure discovery and location, and it is independent of lower SCSI cards and storage devices, so it has good compatibility. This paper also proposes methods for choosing paths, automatically recovering paths and balancing the load. We tested the read performance and the average response time, and the results showed that with the load balanced, the read performance improves 17.9% on average, and the average response time decreases 15.2% on average.
Design And Performance Analysis Of Distributed Fault Tolerant Storage Systems

Minghua Jiang,Ming Hu

2007-01-01

Abstract:With the development of distributed systems and network technology, the storage capacity, performance, and scalability of distributed storage systems have increased rapidly. However, distributed storage systems are confronted with great challenges in fault tolerance. As storage systems scale up, they consist of thousands of storage nodes. Multiple failures will occur frequently, which would induce a disaster because of data loss. This paper presents an efficient and reliable fault tolerance mechanism has become a issue that needs to be resolved urgently in distributed storage systems, meanwhile, this paper describes the queue model of storage nodes in the distributed fault tolerant storage system, the mean response time and throughput is given under read/write operation situation.
Enhancing Virtual Infrastructure To Survive Facility Node Failures

Hongfang Yu,Vishal A. Anand,Chunming Qiao,Gang Sun

DOI: https://doi.org/10.1364/nfoec.2011.jwa4

2011-01-01

Abstract:We propose 1-redundant and K-redundant schemes for the design and mapping of survivable virtual infrastructure to recover from facility node failures while minimizing network costs. The efficiency of our solutions is compared using simulation.
Data Center Node Recovery Strategy for Large-Scale Faults

Qian Li,Shan Yin,Yuan Yang,Bingli Guo,Xin Li,Shanguo Huang

DOI: https://doi.org/10.12783/dtcse/cii2017/17271

2017-01-01

DEStech Transactions on Computer Science and Engineering

Abstract:With the increase of data-center scale, the possibility of large-scale fault in data center networks will increase. This paper proposes a nodes recovery strategy to restore the network connections and reduce losses under large-scale disasters.
An Energy-Balanced WSN Algorithm Based on Active Hibernation and Data Recovery.

Changming Liu,Cai Fu,Deliang Xu,Lin Sun,Lansheng Han

DOI: https://doi.org/10.1007/978-3-319-27119-4_51

2015-01-01

Abstract:Nowadays, many algorithms are brought up to elect the optimal nodes as head nodes for WSNs, but the network lifetime may not be optimal because of the lack of maintenance for cluster heads’(CHs) energy. Based on the above consideration, we come up with the CHs’ active hibernation algorithm for dynamic clustering network. First we propose that after the CHs’ energy reaches its threshold, instead of becoming the non-CHs directly, they will enter a FIFO queue in which they can hibernate. Second, in order to avoid the effect that the hibernation has on data collection and network coverage, a data recovery mechanism is designed. Third, a theoretical model is formed to analyze the effect and efficiency improvement of the CHs’ hibernation work. The theoretical analysis and experiments show that without the influence of the redundant nodes, CHs’ hibernation algorithm has better performance in network lifetime and energy balancing.

Study of Multi-nodes Failover Used in HA Cluster

Study of Multi-node Heartbeat Model Used in HA Cluster

HAaaS: Towards Highly Available Distributed Systems.

Reliability of Wireless Sensor Networks Based on Redundancy of Cluster-Heads

NCluster: Using Multiple Active Name Nodes to Achieve High Availability for HDFS.

Design and implementation of heartbeat in multi-machine environment

Smart Ring: A Model of Node Failure Detection in High Available Cloud Data Center.

The Application of Multi-Path Fault Tolerant Algorithm in WSN Nodes

Research and Implementation of High Availability in Iscsi Multi-level Storage Systems

A New Method to Improve the Dependability of Key Nodes in Pervasive Computing Environment

Multi-Cluster Load Balancing Based On Process Migration

A Novel Virtual Node Migration Approach to Survive a Substrate Link Failure

Fault-tolerant Mechanism of the Distributed Cluster Computers

Design of a Dual-Computer Cluster System and Availability Evaluation

Research on back-up node fault tolerance pattern in WSN

Research of Scheduling Strategy Based on Fault Tolerance in Hadoop Platform

Design and implementation of an efficient multipath for a SAN environment

Design And Performance Analysis Of Distributed Fault Tolerant Storage Systems

Enhancing Virtual Infrastructure To Survive Facility Node Failures

Data Center Node Recovery Strategy for Large-Scale Faults

An Energy-Balanced WSN Algorithm Based on Active Hibernation and Data Recovery.