A Novel Scheduling Mechanism For Hybrid Cloud Systems

Large scale and complexity in structure are two major trends in current cloud environments. Bringing cloud computing to its full potential requires a paradigm shift toward cloud resiliency. A major challenge in designing resilient clouds is to avoid single point of failures in cloud management systems. The state-of-the-art cloud management systems are mostly designed by relying on centralized managing systems with complicated master nodes, which are prone to single point of failures. In this paper, we propose a novel resilient hybrid cloud management system to ensure offering consistent cloud services to users, which consists of two tiers: a peer-to-peer (P2P) tier and a centralized tier. In the centralized tier, each local cloud has a centralized management system, while these local clouds form a P2P cloud in the P2P tier. Our hybrid cloud system benefits from auto scaling across the entire system while it easily manages each local cloud. We propose an innovative user request scheduling mechanism, dedicated to our system, considering multi-objectives: minimizing user requests average makespan, while ensuring load balancing. Finally, we conduct extensive experiments to evaluate the performance of our proposed hybrid cloud management system.

[1]  Aameek Singh,et al.  Server-storage virtualization: Integration and load balancing in data centers , 2008, 2008 SC - International Conference for High Performance Computing, Networking, Storage and Analysis.

[2]  Francesco Tiezzi,et al.  The Autonomic Cloud: A Vision of Voluntary, Peer-2-Peer Cloud Computing , 2013, 2013 IEEE 7th International Conference on Self-Adaptation and Self-Organizing Systems Workshops.

[3]  Jelena V. Misic,et al.  Performance Analysis of Cloud Computing Centers Using M/G/m/m+r Queuing Systems , 2012, IEEE Transactions on Parallel and Distributed Systems.

[4]  Chao Li,et al.  Fuxi: a Fault-Tolerant Resource Management and Job Scheduling System at Internet Scale , 2014, Proc. VLDB Endow..

[5]  Edward G. Coffman,et al.  Computer and job-shop scheduling theory , 1976 .

[6]  Matei Zaharia,et al.  Job Scheduling for Multi-User MapReduce Clusters , 2009 .

[7]  Randy H. Katz,et al.  Improving MapReduce Performance in Heterogeneous Environments , 2008, OSDI.

[8]  Arnold O. Allen.  Probability, Statistics, and Queueing Theory , 1978 .

[9]  Richard M. Karp,et al.  Load Balancing in Structured P2P Systems , 2003, IPTPS.

[10]  Minghua Chen,et al.  Queuing models for peer-to-peer systems , 2009, IPTPS.

[11]  Gregory R. Ganger,et al.  alsched: algebraic scheduling of mixed workloads in heterogeneous clouds , 2012, SoCC '12.

[12]  Chita R. Das,et al.  Modeling and synthesizing task placement constraints in Google compute clusters , 2011, SoCC.

[13]  Abhishek Verma,et al.  Large-scale cluster management at Google with Borg , 2015, EuroSys.

[14]  Alysson Neves Bessani,et al.  The TClouds architecture: Open and resilient cloud-of-clouds computing , 2012, IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN 2012).

[15]  Michael Abd-El-Malek,et al.  Omega: flexible, scalable schedulers for large compute clusters , 2013, EuroSys '13.

[16]  H. Constantin,et al.  MARKOV CHAINS AND QUEUEING THEORY , 2011 .

[17]  Özalp Babaoglu,et al.  Design and implementation of a P2P Cloud system , 2012, SAC '12.

[18]  François Baccelli,et al.  Elements Of Queueing Theory , 1994 .

[19]  Thomas L. Saaty,et al.  Elements of queueing theory , 2003 .