Subject Name:- Cloud Computing

Subject Code:- KCS-713

Unit No.:- 4
Lecture No.:- 1
Topic Name: Inter Cloud Resource Management

1
 Content
1:-Extended Cloud Computing Services
– Cloud Service Tasks and Trends
– Software Stack for Cloud Computing
– Runtime Support Services
2:- Resource Provisioning and Platform Deployment
3: Virtual Machine Creation and Management
4:-Global Exchange of Cloud Resources

2
 1:-Extended Cloud Computing Services

Figure 0: A stack of six layers of cloud services and their providers.

3
• The bottom three layers are more related to
physical requirements.
• The bottommost layer provides Hardware as a
Service (HaaS).
• The next layer is for interconnecting all the
hardware components, and is simply called
Network as a Service (NaaS).
• Virtual LANs fall within the scope of NaaS.

4
• The next layer up offers Location as a Service
(LaaS), which provides a collocation service to
house, power, and secure all the physical
hardware and network resources.
• Some authors say this layer provides Security as a
Service (“SaaS”).
• The cloud infrastructure layer can be further
subdivided as Data as a Service (DaaS) and
Communication as a Service (CaaS) in addition to
compute and storage in IaaS.

5
• Cloud players are divided into three classes:
(1) cloud service providers and IT
administrators,
• (2) software developers or vendors, and
• (3) end users or business users.
• These cloud players vary in their roles under
the IaaS, PaaS, and SaaS models.

6
7
• From the software vendors’ perspective,
application performance on a given cloud
platform is most important.
• From the providers’ perspective, cloud
infrastructure performance is the primary
concern.
• From the end users’ perspective, the quality of
services, including security, is the most
important.

8
 Cloud Service Tasks and Trends
• Cloud services are introduced in five layers.
• The top layer is for SaaS applications, as
further subdivided into the five application
areas in Figure 1, mostly for business
applications.
• For example, CRM is heavily practiced in
business promotion, direct sales, and
marketing services.

9
• CRM offered the first SaaS on the cloud
successfully.
• The approach is to widen market coverage by
investigating customer behaviors and revealing
opportunities by statistical analysis.
• SaaS tools also apply to distributed collaboration,
and financial and human resources management.
• These cloud services have been growing rapidly in
recent years.

10
• PaaS is provided by Google, Salesforce.com, and
Facebook, among others.
• IaaS is provided by Amazon, Windows Azure, and
RackRack, among others.
• Collocation services require multiple cloud
providers to work together to support supply
chains in manufacturing.
• Network cloud services provide communications
such as those by AT&T, Qwest, and AboveNet.

11
 Software Stack for Cloud Computing
• Developers have to consider how to design the
system to meet critical requirements such as high
throughput and fault tolerance.
• Even the operating system might be modified to
meet the special requirement of cloud data
processing.
• Based on the observations of some typical cloud
computing instances, such as Google, Microsoft,
and Yahoo!, the overall software stack structure
of cloud computing software can be viewed as
layers.

12
• Each layer has its own purpose and provides the
interface for the upper layers just as the
traditional software stack does. However, the
lower layers are not completely transparent to
the upper layers.
• The platform for running cloud computing
services can be either physical servers or virtual
servers. By using VMs, the platform can be
flexible, that is, the running services are not
bound to specific hardware platforms.

13
• This brings flexibility to cloud computing platforms.
The software layer on top of the platform is the layer
for storing massive amounts of data.
• This layer acts like the file system in a traditional single
machine. Other layers running on top of the file
system are the layers for executing cloud computing
applications.
• They include the database storage system,
programming for large-scale clusters, and data query
language support. The next layers are the components
in the software stack.

14
 Runtime Support Services
• As in a cluster environment, there are also
some runtime supporting services in the cloud
computing environment. Cluster monitoring is
used to collect the runtime status of the entire
cluster.
• The scheduler queues the tasks submitted to
the whole cluster and assigns the tasks to the
processing nodes according to node
availability.

15
• The distributed scheduler for the cloud
application has special characteristics that can
support cloud applications, such as scheduling
the programs written in MapReduce style.
• The runtime support system keeps the cloud
cluster working properly with high efficiency.

16
• Runtime support is software needed in
browser-initiated applications applied by
thousands of cloud customers.
• The SaaS model provides the software
applications as a service, rather than letting
users purchase the software.
• As a result, on the customer side, there is no
upfront investment in servers or software
licensing.

17
• On the provider side, costs are rather low,
compared with conventional hosting of user
applications.
• The customer data is stored in the cloud that
is either vendor proprietary or a publicly
hosted cloud supporting PaaS and IaaS.

18
 2.Resource Provisioning and Platform
Deployment

• The emergence of computing clouds suggests

fundamental changes in software and hardware
architecture.
• Cloud architecture puts more emphasis on the
number of processor cores or VM instances.
• Parallelism is exploited at the cluster node level.
In this section, we will discuss techniques to
provision computer resources or VMs.
19
 Provisioning of Compute Resources (VMs)
• Providers supply cloud services by signing SLAs
with end users.
• The SLAs must commit sufficient resources
such as CPU, memory, and bandwidth that the
user can use for a preset period.
• Under provisioning of resources will lead to
broken SLAs and penalties.

20
• Over provisioning of resources will lead to
resource underutilization, and consequently, a
decrease in revenue for the provider.
• Deploying an autonomous system to efficiently
provision resources to users is a challenging
problem.
• The difficulty comes from the unpredictability of
consumer demand, software and hardware
failures, heterogeneity of services, power
management, and conflicts in signed SLAs
between consumers and service providers.

21
• Efficient VM provisioning depends on the
cloud architecture and management of cloud
infrastructures.
• Resource provisioning schemes also demand
fast discovery of services and data in cloud
computing infrastructures.
• In a virtualized cluster of servers, this
demands efficient installation of VMs, live VM
migration, and fast recovery from failures.

22
• To deploy VMs, users treat them as physical
hosts with customized operating systems for
specific applications.
• For example, Amazon’s EC2 uses Xen as the
virtual machine monitor (VMM). The same
VMM is used in IBM’s Blue Cloud.

23
• In the EC2 platform, some predefined VM
templates are also provided. Users can choose
different kinds of VMs from the templates.
• IBM’s Blue Cloud does not provide any VM
templates.
• In general, any type of VM can run on top of
Xen. Microsoft also applies virtualization in its
Azure cloud platform.

24
 Resource Provisioning Methods
• Figure 1(next slide) shows three cases of static
cloud resource provisioning policies.
• (a)overprovisioning with the peak load causes
heavy resource waste (shaded area).
• (b), underprovisioning (along the capacity line) of
resources results in losses by both user and
provider in that paid demand by the users (the
shaded area above the capacity) is not served and
wasted resources still exist for those demanded
areas below the provisioned capacity.

25
Figure1: Three cases of cloud resource provisioning without elasticity: (a) heavy
waste due to overprovisioning, (b) underprovisioning and (c) under- and then
overprovisioning.

26
• (c), the constant provisioning of resources
with fixed capacity to a declining user demand
could result in even worse resource waste.
• The user may give up the service by canceling
the demand, resulting in reduced revenue for
the provider.
• Both the user and provider may be losers in
resource provisioning without elasticity.

27
• The demand-driven method provides static
resources and has been used in grid
computing for many years.
• The event driven method is based on
predicted workload by time.
• The popularity-driven method is based on
Internet traffic monitored.

28
 Demand-Driven Resource Provisioning
• This method adds or removes computing instances
based on the current utilization level of the allocated
resources.
• The demand-driven method automatically allocates
two Xeon processors for the user application, when
the user was using one Xeon processor more than 60
percent of the time for an extended period.
• In general, when a resource has surpassed a threshold
for a certain amount of time, the scheme increases
that resource based on demand.

29
• When a resource is below a threshold for a
certain amount of time, that resource could
be decreased accordingly.
• Amazon implements such an auto-scale
feature in its EC2 platform.
• This method is easy to implement.
• The scheme does not work out right if the
workload changes abruptly.

30
Figure 2: EC2
performance results
on the AWS EC2
platform, collected
from experiments at
the University of
Southern California
using three resource
provisioning methods.

31
• The x-axis in Figure 2 is the time scale in milliseconds.
• In the beginning, heavy fluctuations of CPU load are
encountered.
• All three methods have demanded a few VM instances
initially.
• Gradually, the utilization rate becomes more stabilized with a
maximum of 20 VMs (100 percent utilization) provided for
demand-driven provisioning in Figure 2(a).
• However, the event-driven method reaches a stable peak of
17 VMs toward the end of the event and drops quickly in
Figure 2(b).
• The popularity provisioning shown in Figure 2(c) leads to a
similar fluctuation with peak VM utilization in the middle of
the plot. 32
 Event-Driven Resource Provisioning
• This scheme adds or removes machine instances based on a
specific time event.
• The scheme works better for seasonal or predicted events
such as Christmastime in the West and the Lunar New Year
in the East.
• During these events, the number of users grows before the
event period and then decreases during the event period.
• This scheme anticipates peak traffic before it happens. The
method results in a minimal loss of QoS, if the event is
predicted correctly. Otherwise, wasted resources are even
greater due to events that do not follow a fixed pattern.

33
 Popularity-Driven Resource Provisioning
• In this method, the Internet searches for popularity of
certain applications and creates the instances by popularity
demand.
• The scheme anticipates increased traffic with popularity.
• Again, the scheme has a minimal loss of QoS, if the
predicted popularity is correct. Resources may be wasted if
traffic does not occur as expected.
• In Figure 2(c), EC2 performance by CPU utilization rate (the
dark curve with the percentage scale shown on the left) is
plotted against the number of VMs provisioned (the light
curves with scale shown on the right, with a maximum of 20
VMs provisioned).

34
 Important Questions
Q1. What is intercloud resource management?
Q2. Elaborate on four major advantages of using virtualized
resources in cloud computing applications. Your discussion
should address resource management issues from the
provider’s perspective and the application flexibility,
cost-effectiveness, and dependability concerns by cloud users.
Q3. Draw a layered diagram to relate the construction of IaaS,
PaaS, and SaaS clouds from bare machine hardware to the
users applications. Briefly list the representative cloud service
offerings at each cloud layer from the major cloud providers
that you know of.

35
 References
Text Books:-
• 1. Kai Hwang, Geoffrey C. Fox, Jack G.
Dongarra, “Distributed and Cloud Computing,
From Parallel Processing to the Internet of
Things”, Morgan Kaufmann Publishers, 2012.
• 2. Rittinghouse, John W., and James F.
Ransome, ―Cloud Computing:
Implementation, Management and Security,
CRC Press, 2017.

36
Thank You

37