Green Cloud Computing - IUST Personal Webpages

advertisement
Green Cloud Computing
Hadi Salimi
Distributed Systems Lab,
School of Computer Engineering,
Iran University of Science and Technology,
hsalimi@iust.ac.ir
Fall 2010
Green Cloud Computing
Vision and Potential of Cloud Computing
- Computer networks grow up and become sophisticated !
- Computing utilities based on a service provisioning model expected
- Users can access applications as services from anywhere in the
world on demand
Page  2
Green Cloud Computing
Challenges of Cloud Infrastructure
- On-demand resource provisioning in response to time-varying workloads
- Data centers are expensive to maintain, High energy costs and huge
carbon footprints
- Lowering the energy usage of data centers is a complex issue
Page  3
Green Cloud Computing
Challenges of Cloud Infrastructure(cont.)
- Cloud resources need to be allocated not only to satisfy QoS, but also to
reduce energy usage.
- Rising energy cost is a potential threat as it increases the TCO & ROI of
Cloud infrastructure setup by providers.
Page  4
Green Cloud Computing
Cloud Infrastructure
Page  5
Green Cloud Computing
Green Cloud Architectural Elements
- a) Consumers/Brokers
- b) Green Resource Allocator
 Green Negotiator
 Service Analyzer
 Consumer Profiler
 Pricing
 Energy Monitor
 Service Scheduler
 VM Manager
 Accounting
- c) Virtual Machines
- d) Physical Machines
Page  6
Green Cloud Computing
Energy-Aware Data Centre Resource Allocation
- The problem of VM allocation can be divided in two parts
 admission of new requests for VM
 optimization of current allocation of VM
- To solve it we apply modification of the Best Fit Decreasing algorithm
- The complexity where is the number of VMs that have to be allocated and
is the number of hosts
 select VMs that need to be migrated
 chosen VMs are placed on hosts
- The first heuristic is Single Threshold
Page  7
Green Cloud Computing
Open Challenges
- Ability to transfer VMs between physical nodes using live migration
- Minimal number of nodes according to current resource requirements
- Performance degradation and, thus result in SLA violation
Page  8
Green Cloud Computing
Energy-aware Dynamic Resource Allocation
Challenge
- Resized and consolidated
- Provide high performance while meeting SLA
 Excessive power cycling of a server could
 Turning resources off in a dynamic environment
 Ensuring SLA brings challenges
Page  9
Green Cloud Computing
QoS-based Resource Selection and Provisioning
Challenge
- QoS-aware resource selection plays an important role in Cloud computing
- explore the trade-offs between QoS and energy saving
Page  10
Green Cloud Computing
Optimization of Virtual Network Topologies Challenge
- Communicating VMs may costly data transfer between physical nodes
- To observe the communication between VMs and place them on the
same or closely located nodes
Page  11
Green Cloud Computing
Autonomic Optimization of Thermal states and
Cooling System Operation Challenge
- High temperature leads to a reduced system reliability and availability,
decreased lifetime of devices
o Leverages the temperature variations between different workloads
o Hardware level thermal management techniques
Page  12
Green Cloud Computing
Efficient Consolidation of VMs for Managing
Heterogeneous Workloads Challenge
- Different types of applications can be allocated on one physical computer
node
- What kind of applications can be allocated to a single host that efficient
usage of the resources.
Page  13
Green Cloud Computing
Concluding
Cloud computing plays a significant role in the reduction of data center
energy consumption costs
Consumers are increasingly becoming conscious about the environment
Page  14
Green Cloud Computing
References
 Rajkumar Buyya, et al. "Energy-Efficient Management of Data Center
Resources for Cloud Computing: A Vision, Architectural Elements, and
Open Challenges."
Page  15
Green Cloud Computing
Download