cloud schedualing

cloud schedualing

• 1.
  In the NameOf GOD Guide : Student : Fall 93 1
• 2.
  A Ranking ChaosAlgorithm for dual scheduling of cloud service and computing resource in private cloud A R T I C L E I N F O Article history: Received 20 September 2012 Received in revised form 22 January 2013 Accepted 15 February 2013 Available online 16 March 2013 2
• 3.
  Abstract  ServiceComposition Optimal Selection (SCOS)  Optimal Allocation of Computing Resources (OACR)  Dual Scheduling of Cloud Services and Computing Resources (DS-CSCR)  Ranking Chaos Optimization (RCO)  Private cloud 3
• 4.
  Introduction  AddressingSCOS and OACR step by step with two different algorithms independently becomes very cumbersome and inefficient.  We propose the idea of combining two stages decision-making into one and put forward the concept, Dual Scheduling of Cloud Service and Computing Resource (DS-CSCR), in private cloud.  To achieve high efficient one-time decision in DS-CSCR, a new Ranking Chaos Optimization (RCO) is designed. 4
• 5.
  Structure of privatecloud and actors in conglomerate 5
• 6.
  The modeling ofDS-CSCR in private cloud  The characteristics and QoS indexes of cloud services  The characteristics and QoS indexes of software cloud services  The characteristics and QoS indexes of hardware cloud services  The characteristics and QoS indexes of VMs  Problem formulation of DS-CSCR in private cloud 6
• 7.
  Flowchart RCO begin initialization Ranking section Individual Chaos Dynamic Heuristic evaluation Global updating Stoppin g criteria Return best strategy yes no 7
• 8.
  The complexity ofthe operators in GA and RCO. GA Roulette wheel selection O(n2) Crossover O(nm) Mutation O(nm) O(n2) O(m) O(1) O(1) RCO Ranking selection O(n log n) Individual chaos O(n) Dynamic heuristic O(n*(max(p, s))) O(n log n) O(1) O(s) O(p) 8 Algorithm | time complexities of operators | n→∞ | m→∞ | s→∞ | p→∞ n:population m : size P :computing resource S: availableCloudService
• 9.
  Searching capability ofRCO for solving DS-CSCR  from bad to good is: GA < CGA < CO < RCO2 < RCO1 < RCO Simple chaos optimization can get much better solution than the traditional GA and improved CGA. 9
• 10.
  Time consumption andstability of RCO for solving DS-CSCR 10
• 11.
  sort of stabilityof the six algorithms Sort of stability of the six algorithms from bad to good is: CGA < CO < RCO2 < RCO1 < RCO < GA 11
• 12.
  Disadvantage of DS-CSCR  quite cumbersome  mutual relations between cloud services and underlying computing resources are always ignored. 12
• 13.
  Advantages of DS-CSCR  With the new DS-CSCR and RCO, cloud services and computing infrastructures can then be quickly combined and shared with high efficient decision. 13
• 14.
  Conclusions  NewDS-CSCR model was presented in private cloud for high efficient one time decision. Properties of software/hardware cloud services, VMs and computing resources are deeply analyzed.  For addressing the complex dual scheduling problem (DS-CSCR),a new intelligent algorithm – RCO was presented. Individual chaos operator was designed as the backbone operator of the algorithm.  RCO with these three operators showed remarkable performances in terms of searching ability, time complexity and stability in solving the DS-CSCR problem in such private cloud compared with other algorithms. 14
• 15.
  Future work The QoS properties for manufacturing capabilities and human resources are needed to be studied.  RCO presented in this paper still has some disadvantages.  As a new improved intelligent algorithm, its effectiveness in various other complex combinatorial optimization problems remains to be further explored and validated. 15
• 16.
  question 16
• 17.
  Thanks for yourattention 17