CS8791 / Cloud Computing 2022

UNIT-IV
INTER CLOUD RESOURCE MANAGEMENT

 Cloud computing is a term used to describe a paradigm for delivery of computing

services to users on a pay-as-you-go basis through which, users utilize the
Internet and remote data centers to run applications and store data.
 Interconnected Clouds, are also called Inter-Cloud, can be viewed as a natural
evolution of Cloud computing.

Extended Cloud Computing Services

 The overall stack of cloud computing software can be viewed as layers.

 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
 There are six layers of cloud services, ranging from hardware, network, and
collocation to infrastructure, platform, and software applications.

Dr. A.M.Rajeswari, CSE, VCET Page 1 of 13

 CS8791 / Cloud Computing 2022

 The bottom three layers are more related to physical requirements.

o The bottommost layer provides Hardware as a Service (HaaS).
o 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.
o 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.
 SaaS tools apply to distributed collaboration, and financial and human resources
management
 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

Dr. A.M.Rajeswari, CSE, VCET Page 2 of 13

 CS8791 / Cloud Computing 2022

Three Classes of Cloud players:

1. Cloud service providers and IT administrators: Cloud infrastructure
performance is the primary concern.
2. Software developers or vendors: Application performance on a given cloud
platform is most important
3. End users or business users: The quality of services, including security, is
the most important.
 These cloud players vary in their roles under the IaaS, PaaS, and SaaS models.

Inter Connected Clouds:

 Key elements of interconnected clouds is interoperability.

Dr. A.M.Rajeswari, CSE, VCET Page 3 of 13

 CS8791 / Cloud Computing 2022

 Cloud interoperability in practice can be obtained either through brokering or

standard interfaces.
 By using a service broker, which translates messages between different Cloud
interfaces, customers are able to switch between different Clouds and Cloud
providers can interoperate.
 Standardization of interfaces is another common method for realization of
interoperability.
 Both Cloud providers and customers benefit from Cloud interoperability.
 Essential motivations for Cloud interoperability such as avoiding vendor lock-in,
scalability, availability, low access latency, and energy efficiency and etc.
 Cloud interoperability requires cloud providers to adopt and implement standard
interfaces, protocols, formats and architectural components that facilitate
collaboration.
 Even if Cloud interoperability is not supported by Cloud providers, Cloud
customers are still able to benefit from client-centric interoperability facilitated by
user-side libraries or third party brokers

Resource Provisioning and Platform Deployment:

 In Cloud the following resources are provisioned

1. Computer resources (VMs).
2. Storage resources allocation schemes to interconnect distributed computing
infrastructures.
1. Provisioning of Compute Resources (VMs)
 Providers supply cloud services by signing SLAs with end users.
 The SLAs must commit sufficient resources such as
o CPU o Memory o Bandwidth
that the user can use for a preset period.
 Under provisioning of resources will lead to broken SLAs and penalties.
Dr. A.M.Rajeswari, CSE, VCET Page 4 of 13
 CS8791 / Cloud Computing 2022

 Overprovisioning of resources will lead to resource underutilization, and

consequently, a decrease in revenue for the provider.
 Challenging in resource provisioning are comes from the unpredictability of
o Consumer demand
o Software and Hardware failures
o Heterogeneity of services
o Power management
o Conflicts in signed SLAs between consumers and service providers.
 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.
2. Provisioning of Storage Resources:
 The data storage layer is built on top of the physical or virtual servers.
 As the cloud computing applications often provide service to users, it is
unavoidable that the data is stored in the clusters of the cloud provider.
o The service can be accessed anywhere in the world.
o Examples: e-mail systems, web searching application.
 A distributed file system is very important for storing large-scale data.
 In cloud computing, another form of data storage is (Key, Value) pairs.
 Amazon S3 service uses SOAP to access the objects stored in the cloud.

Dr. A.M.Rajeswari, CSE, VCET Page 5 of 13

 CS8791 / Cloud Computing 2022

Resource Provisioning Methods:

 Three cases of static cloud resource provisioning policies are in practice.

a. Over provisioning: The peak load causes heavy resource waste (shaded
area).
b. Under provisioning: Resources (along the capacity line) 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.
c. Constant provisioning: The 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.
 To normalize the resource provisioning and make it more efficient, three methods
are adopted.

Dr. A.M.Rajeswari, CSE, VCET Page 6 of 13

 CS8791 / Cloud Computing 2022

1. The demand-driven method: Provides static resources and has been used in
grid computing for many years.
2. The event driven method: Is based on predicted workload by time.
3. The popularity-driven method: Is based on Internet traffic monitored.

1. 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.
 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.
 Gradually, the utilization rate becomes more stabilized with a maximum of 20 VMs
(100 percent utilization) provided for demand-driven provisioning.

Dr. A.M.Rajeswari, CSE, VCET Page 7 of 13

 CS8791 / Cloud Computing 2022

2. 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 Christmas time
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.
 However, the event-driven method reaches a stable peak of 17 VMs toward the
end of the event and drops quickly in.

3. 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.

Dr. A.M.Rajeswari, CSE, VCET Page 8 of 13

 CS8791 / Cloud Computing 2022

 In Figure, 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).

 The popularity provisioning shown in Figure leads to a similar fluctuation with peak
VM utilization in the middle of the plot.

Dynamic Resource Deployment:

 The cloud uses VMs as building blocks to create an execution environment across
multiple resource sites.
 The InterGrid-managed infrastructure was developed by a Melbourne University
group [19].
 Dynamic resource deployment can be implemented to achieve scalability in
performance.
 The InterGrid is a Java-implemented software system that lets users create
execution cloud environments on top of all participating grid resources.
 Peering arrangements established between gateways enable the allocation of
resources from multiple grids to establish the execution environment.
 In Figure, a scenario is illustrated by which an inter grid gateway (IGG) allocates
resources from a local cluster to deploy applications in three steps:
Dr. A.M.Rajeswari, CSE, VCET Page 9 of 13
 CS8791 / Cloud Computing 2022

o Requesting the VMs,

o Enacting the leases
o Deploying
the VMs as requested.

 Under peak demand, this IGG interacts with another IGG that can allocate
resources from a cloud computing provider.
 A grid has predefined peering arrangements with other grids, which the IGG
manages.
 Through multiple IGGs, the system coordinates the use of Inter Grid resources.
 An IGG is aware of the peering terms with other grids, selects suitable grids that
can provide the required resources, and replies to requests from other IGGs.
 The InterGrid allocates and provides a distributed virtual environment (DVE).
 This is a virtual cluster of VMs that runs isolated from other virtual clusters.
 A component called the DVE manager performs resource allocation and
management on behalf of specific user applications.
 The core component of the IGG is a scheduler for implementing provisioning
policies and peering with other gateways.

Dr. A.M.Rajeswari, CSE, VCET Page 10 of 13

 CS8791 / Cloud Computing 2022

 The communication component provides an asynchronous message-passing

mechanism.
 Received messages are handled in parallel by a thread pool.

Global Exchange of Cloud Resource:

 Enterprises currently employ Cloud services in order to improve the scalability of

their services
 Enterprise service consumers with global operations require faster response time.
 Save time by distributing workload requests to multiple Clouds in various locations
at the same time.
 Cloud computing to be mature, it is required that the services follow standard
interfaces, BECAUSE,
o No standard interfaces/ compatibility between vendor’s systems
o No uniform rate
 Market system, modeled on real-world exchanges.
 University of Melbourne proposed Inter-Cloud architecture (Cloud bus), which
supports brokering and exchange of cloud resources for scaling applications
across multiple clouds.
 Cloud Exchange (CEx) acts as a market maker for bringing together service
producers and consumers.
 It aggregates the infrastructure demands from application brokers and evaluates
them against the available supply
 It supports trading of cloud services based on competitive economic models such
as commodity markets and auctions.
 The Market oriented Cloud will look like

Dr. A.M.Rajeswari, CSE, VCET Page 11 of 13

 CS8791 / Cloud Computing 2022

CEx Components

1. Market directory: Allows participants to locate providers or consumers with the

right offers.
2. Auctioneers: Periodically clear bids and asks received from market participants.
3. Banking Systems: Ensures that financial transactions pertaining to agreements
between participants are carried out.
4. Brokers: Mediate between consumers and providers by buying capacity from the
provider and sub-leasing these to the consumers.
- Has to choose those users whose applications can provide it maximum utility.
- Interacts with resource providers and other brokers to gain or to trade resource
shares.

Dr. A.M.Rajeswari, CSE, VCET Page 12 of 13

 CS8791 / Cloud Computing 2022

- Equipped with a negotiation module that is informed by the current conditions

of the resources and the current demand to make its decisions.
- Consumers, brokers and providers are bound to their requirements and related
compensations through SLAs.
5. SLA: Specifies the details of the service to be provided in terms of metrics agreed
upon by all parties, and penalties for meeting and violating the expectations,
respectively.
6. Providers can use the markets in order to perform effective capacity planning
7. A provider is equipped with a price-setting mechanism which sets the current
price for the resource based on market conditions, user demand, and current level
of utilization of the resource.
8. Negotiation process proceeds until an SLA is formed or the participants decide to
break off.
9. An admission-control mechanism at a provider’s end selects the auctions to
participate in or the brokers to negotiate with, based on an initial estimate of the
utility.
10. Resource management system provides functionalities such as advance
reservations that enable guaranteed provisioning of resource capacity

************************

Dr. A.M.Rajeswari, CSE, VCET Page 13 of 13