Cloud Computing
Resource Management
In Cloud
Cloud Computing
Different Resources in Computing
Infrastructure, Platform, Application
Resource Types; Physical resource, Logical resource
Resource Management in Cloud
Provider perspective, Consumer perspective
Resource Management for Cloud
Infrastructure, Platform, Application
Challenges in IaaS
Virtualization, Multi-tenancy, Resource management,
Network infrastructure management, Data management,
APIs, interoperability etc.
Resource Management for IaaS
Cloud Computing
Resource Management Objectives:
Resources Management-Challenges for IaaS
(Hardware Resources)
(Logical Resources)
Resource Management Problems in IaaS
Resource provisioning
Resource allocation
Resource requirement mapping
Resource adaptation,
Resource discovery
Resource brokering
Resource estimation
Resource modeling.
Cloud Computing
Resource Management Solutions in IaaS
Resource provisioning Approaches
Open challenges in resource provisioning
Resource allocation Approaches
Open challenges in resource allocation.
Resource mapping Approaches
Open challenges in resource mapping
Resource adaptation Approaches
Open challenges in resource adaptation
Performance Metrics For Resource Management
Reliability, Ease of deployment, QoS:, Delay, Control
Overhead
Different Resources in Computing
5
Resource Types
6
Resources Management
Resources Management play an important role in cloud
computing:
Provider perspective: Provider wants maximum utilization
of his resources with minimal energy cost for maximum profit
without SLA violation and QoS.
Consumer perspective: Consumer wants guarantee for
SLA respect and QoS
7
Resource Management
How Resources (hard and soft ) can be made available to
the users in an efficient way. Efficient in term of:
Profit maximization,
Energy Optimization,
Respecting SLAs
and providing QoS.
8
Resources Management For Cloud
Clouds can be broadly classified as follows:
Infrastructures as a Service (IaaS)
Platforms as a Service (PaaS)
Software as a Service (SaaS)
9
Challenges in IaaS
Challenges in IaaS
Virtualization,
Multi-tenancy
Resource management,
Network infrastructure management,
Data management,
APIs,
Interoperability etc.
One of the major challenges in IaaS is Resources
Management.
10
Resources Management For IaaS
Infrastructure as-a-Service (IaaS) is a most powerful cloud
service, besides others like SaaS, PaaS etc
In IaaS, cloud providers offer resources to users/machines
that include:
computers as virtual machines,
raw(block) storage,
firewalls,
load balancers,
network devices.
One of the major challenges in IaaS is Resource
Management.
11
Resources Management-Objectives
Resource management for IaaS offers following benefits: Or
following should preserved for better Resource management
Scalability:
Quality of service:
Optimal utility:
Reduced overheads:
Improved throughput:
Reduced latency:
Cost effectiveness:
Simplified interface:
12
Resources Management-Challenges
(Hardware)
CPU (central processing unit):
Memory:
Storage:
Workstations:
Network elements:
Sensors/actuators:
Specialized environment:
Cost effectiveness:
Simplified interface:
13
Resources Management-Challenges
(Logical Resources)
Operating system:
Energy:
Network Throughput/bandwidth:
Load balancing mechanisms:
Information security:
Delays:
APIs/(Applications Programming Interfaces):
Protocols:
14
Resources Management-Aspects
Resource management problems include:
Resource provisioning:
Resource allocation:
Resource requirement mapping:
Resource adaptation:
Resource discovery:
Resource brokering:
Resource estimation:
Resource modeling:
15
Resources Management-Aspects
Resource provisioning: It is the allocation of a service
provider's resources to a customer
Resource allocation: It is the distribution of resources
economically among competing groups of people
Resource adaptation: It is the ability or capacity of that
system to adjust the resources dynamically to fulfill the
requirements of the user
Resource requirement mapping: It is a correspondence
between resources required by the users and resources
available with the provider
Resource discovery & selection: It is the identification of
list of authenticated resources that are available for job
submission and to choose the best among them
16
Resources Management-Aspects
Resource brokering: It is the negotiation of the resources
through an agent to ensure that the necessary resources
are available at the right time to complete the objectives
Resource estimation: It is a close guess of the actual
resources required for an application, usually with some
thought or calculation involved
Resource scheduling: A resource schedule is a timetable
of events and resources. Shared resources are available at
certain times and events are planned during these times.
17
Performance Metrics For
Resource Management
Reliability:is defined as the ability of machine, or system to
consistently perform its intended or required function or
mission, on demand without degradation /failure.
Ease of deployment: refers to the easiness in implementing
the system model. The value for ease of deployment has
been assigned as high if the infrastructures are easily
available for deployment.
QoS: refers to a broad collection of networking technologies
and techniques. The goal of QoS is to provide guarantees on
the ability of a network to deliver predictable results. The
higher availability, higher bandwidth, lower latency, and lower
18
error rate offers higher QoS.
Performance Metrics For
Resource Management
Delay: Delay is the time taken from point-to-point in a
network. Higher delay degrades performance of the system
and vice versa.
Control Overhead: Control overhead refers to the extra
consideration required by a system that is not directly related
to data.
19
Open Challenges in
Resource Provisioning
The challenges in resource provisioning are as follows.
How to make the applications hosted on the cloud to be
elastic so that we can achieve economy of scale while
preserving the application-specific Service Level Agreements
(SLAs) such as, response time, throughput?
How do we develop resource prediction models for facilitating
proactive scaling in the cloud so that hosted applications are
able to withstand the variation in workload with least drop in
performance and availability?
20
Data Center Power Consumption
Currently it is estimated that servers consume 0.5%
of the world’s total electricity usage.
Closer to 1.2% when data center systems are factored
into the equation.
Server energy demand doubles every 4-6 years.
This results in large amounts of CO2 produced by
burning fossil fuels.
What if we could reduce the energy used with minimal
performance impact?
21
Motivation for Green Data Centers
Economic Environmental
New data centers run on 70% of the U.S. energy
the Megawatt scale, sources are fossil fuels.
requiring millions of 2.8 billion tons of CO2
dollars to operate. emitted each year from
Recently institutions are U.S. power plants.
looking for new ways to Sustainable energy
reduce costs, no more
sources are not ready.
“blank checks.”
Need to reduce energy
Many facilities are at
dependence until a
their peak operating
more sustainable
envelope, and cannot
energy source is
expand without a new
deployed.
power source. 22
Green Computing
Advanced scheduling schemas to reduce energy
consumption.
Power aware
Thermal aware
Performance/Watt is not following Moore’s law.
Data center designs to reduce Power Usage Effectiveness.
Cooling systems
Rack design
23
Research Directions
There are a number of areas to explore in order to
conserve energy within a Cloud environment.
Schedule VMs to conserve energy.
Management of both VMs and underlying
infrastructure.
Minimize operating inefficiencies for non-essential
tasks.
Optimize data center design.
24
Framework
25
VM scheduling on Multi-core Systems
There is a nonlinear
relationship between
the number of
processes used and
power consumption
We can schedule
VMs to take
advantage of this
relationship in order
to conserve power
Power consumption curve on an Intel
Core i7 920 Server
Schedulin (4 cores, 8 virtual cores with 26
g Hyperthreading)
Power-aware Scheduling
Schedule as many
VMs at once on a
multi-core node.
Greedy scheduling
algorithm
Keep track of cores on
a given node
Match vm
requirements with
node capacity
Schedulin 27
g
485 Watts vs. 552 Watts
V V V V V V V V
M M M M M M M M
Node 1 @ 170W Node 2 @ 105W
Node 3 @ 105W Node 4 @ 105W
VS.
V V V V
M M M M
Node 1 @ 138W Node 2 @ 138W
V V V V
M M M M
Node 3 @ 138W Node 4 @ 138W 28
VM Management
Monitor Cloud usage and load.
When load decreases:
• Live migrate VMs to more utilized nodes.
• Shutdown unused nodes.
When load increases:
• Use WOL to start up waiting nodes.
• Schedule new VMs to new nodes.
Managemen 29
t
VM VM VM VM
1
Node 1 Node 2
VM VM VM VM VM
2
Node 1 Node 2
VM VM VM VM
3
Node 1 Node 2
VM VM VM VM
4
Node 1 Node 2 (offline) 30
Minimizing VM Instances
Virtual machines are desktop-based.
Lots of unwanted packages.
Unneeded services.
Are multi-application oriented, not service oriented.
Clouds are based off of a Service Oriented
Architecture.
Need a custom lightweight Linux VM for service
oriented science.
Need to keep VM image as small as possible to
reduce network latency.
31
Typical Cloud Linux Image
Start with Ubuntu 9.04.
Remove all packages not
required for base image.
No X11
No Window Manager
Minimalistic server install
Can load language support on demand
(via package manager)
Readahead profiling utility.
Reorder boot sequence
Pre-fetch boot files on disk
Minimize CPU idle time due to I/O delay
Optimize Linux kernel.
Built for Xen DomU
No 3d graphics, no sound, minimalistic
kernel
Build modules within kernel directly
VM Image 32
Design
Energy Savings
Reduced boot times from 38 seconds to just 8 seconds.
30 seconds @ 250Watts is 2.08wh or .002kwh.
In a small Cloud where 100 images are created every
hour.
Saves .2kwh of operation @ 15.2c per kwh.
At 15.2c per kwh this saves $262.65 every year.
– In a production Cloud where 1000 images are created
every minute.
Saves 120kwh less every hour.
At 15.2c per kwh this saves over 1 million dollars every year.
Image size from 4GB to 635MB.
Reduces time to perform live-migration.
Can do better.
VM Image 33
Design
