Cloud Computing

Lectureslides – III

© 2024 Er. Santosh Panth

UNIT 3

Roles of Grid Computing and Autonomic

Computing in Cloud Computing
Grid Computing
▪ Grid computing is a distributed architecture of multiple computers
connected by networks to accomplish a joint task.

▪ These tasks are compute-intensive and difficult for a single machine to

handle.

▪ Several machines on a network collaborate under a common protocol and

work as a single virtual supercomputer to get complex tasks done.

▪ This offers powerful virtualization by creating a single system image that

grants users and applications seamless access to IT capabilities.
Grid Computing
A typical grid computing network consists
of three machine types:

•Control node/server: A control node is a

server or a group of servers that
administers the entire network and
maintains the record for resources in a
network pool.
•Provider/grid node: A provider or grid
node is a computer that contributes its
resources to the network resource pool.
•User: A user refers to the computer that
uses the resources on the network to
complete the task.
Grid Computing
In grid computing, each computing task is broken into small fragments and distributed
across computing nodes for efficient execution. Each fragment is processed in parallel,
and, as a result, a complex task is accomplished in less time. Let’s consider this equation:
X = (4 x 7) + (3 x 9) + (2 x 5)
Typically, on a desktop computer, the steps needed here to calculate the value of X may look
like this:
•Step 1: X = 28 + (3 x 9) + (2 x 5)
•Step 2: X = 28 + 27 + (2 x 5)
•Step 3: X = 28 + 27 + 10
•Step 4: X = 65
However, in a grid computing setup, the steps are different as three processors or computers
calculate different pieces of the equation separately and combine them later. The steps look
like this:
•Step 1: X = 28 + 27 + 10
•Step 2: X = 65
As seen above, grid computing combines the involved steps due to the multiplicity of
available resources. This implies fewer steps and shorter timeframes.
Types of Grid Computing
1. Computational Grid:
▪ Purpose: To provide substantial computational power by pooling
together processing resources.
▪ Use Cases: Scientific simulations, complex calculations, data analysis.
▪ Examples: SETI@home, Folding@home.
2. Data Grid:
▪ Purpose: To manage and process large datasets distributed across
various locations.
▪ Use Cases: Data-intensive applications, large-scale data analysis,
distributed database management.
▪ Examples: CERN's Large Hadron Collider (LHC) data management,
climate modeling data grids.
3.Collaborative Grid:
▪ Purpose: To enable collaboration among distributed teams and
organizations, allowing shared access to resources and data.
▪ Use Cases: Collaborative research projects, distributed software
development, virtual organizations.
▪ Examples: Virtual observatories, collaborative design projects.
Types of Grid Computing
4. Manuscript Grid Computing:
▪ Purpose: Handle large-scale manuscript digitization, analysis,
and management.
▪ Use Cases: Digital libraries, academic text analysis,
collaborative editing.
▪ Examples: Google Books Project.

5. Modular Grid Computing:

▪ Purpose: Provide flexible, scalable, and maintainable grid
infrastructure.
▪ Use Cases: Scientific research, enterprise IT, cloud services,
healthcare data management.
▪ Examples: Apache Hadoop.
Interaction Models of Grid and Cloud Computing
▪ One of the most scalable interaction models of Grid domains is peer-to-peer,where
most of the Grid participating organizations are both consumers and providers. In
practice, there are usually agreements of resource sharing among the peers.

▪ Furthermore, clients of consumer organizations in Grids use heterogeneous

resources from more than one resource provider belonging to the same Virtual
Organization (VO) to execute their applications.

▪ It is important for participating resource providers and consumers to have common

information models,interaction protocols, application execution states, etc.
▪ The organization of Open Grid Forum (OGF)3has the goal of establishing relevant
and necessary standards for Grid computing. Some proposed standards include Job
Submission Description Language (JSDL), Basic Execution Service (BES) and
others.
▪ Currently, most of the Cloud providers offer their own proprietary service protocols
and information formats.
Distributed Computing in the Grid and Cloud
▪ Operational Focus: Both Grid and Cloud computing manage
distributed computing resources to execute client jobs, but the
Grid is primarily for scientific and HPC applications, while the
Cloud provides scalable, on-demand services.
▪ Services and Protocols: The Grid uses complex, distributed
components to handle job submission, security, machine
provisioning, and data staging. The Cloud offers different
service models (IaaS, PaaS, SaaS) to meet various client
needs.
▪ Interoperability: The Grid excels in interoperability across
different administrative domains, forming virtual organizations.
The Cloud is moving towards a similar model for resource
sharing and overflow management among providers.
▪ Resource Management: The Grid focuses on collaborative
resource sharing among institutions, whereas the Cloud
emphasizes flexibility and scalability in resource provisioning.
▪ Client Services: The Grid provides a specialized environment
for scientific and HPC applications, while the Cloud supports a
broader range of applications through multiple service models.
Layered models and usage patterns in Grid and
Cloud
There are many similarities in Grid and Cloud computing systems. We compare
the approaches by differentiating three layers of abstraction in Grid:
Infrastructure, Platform and Application. Then we map these three layers to the
Cloud services of IaaS, PaaS, and SaaS. An example of the relations among
layers can be seen in Fig.
Layered models and usage patterns in Grid and
Cloud
1. Infrastructure:
▪ This is the layer in which Clouds share most characteristics with the original purpose of
Grid middleware.
▪ In these systems users can provision execution environments in the form of virtual
machines through interfaces such as APIs or command line tools.
▪ The act of defining an execution environment and sending a request to the final
resource has many similarities with scheduling a job in the Grid.

2. Platform:
▪ This layer is built on top of the physical infrastructure and offers a higher level of
abstraction to users.
▪ The interface provided by a PaaS solution allows developers to build additional
services without being exposed to the underlying physical or virtual resources.
▪ The Platform layer allows developers to build new software that takes advantage of
the available resource
Layered models and usage patterns in Grid and
Cloud
3. Applications

▪ There is no clear distinctions between applications developed on Grids and

those that use Clouds to perform execution and storage.

▪ The choice of platform should not influence the final result, since the
computations delegated to the underlying systems can take different shapes
to accommodate to the available APIs and resources
Interoperability in Grids and Clouds
▪ One goal of Grid computing is to
provide uniform and consistent access
to resources distributed in different data
centers and institutions.

▪ This is because the majority of Grids

are formed based on regional as
opposed to local initiatives so
interoperation is a key objective.
interoperability using a uniform access
interface to application users. Software
layers beneath the user interface then
abstract the complexity of the
underlying heterogeneous
supercomputing infrastructure.
▪ One tool that takes this approach for Fig. Meta-brokering architecture
Grid interoperation is meta- brokering .
Concept of Autonomic Computing
▪ "Autonomic Computing" is a new vision of computing initiated by IBM.

▪ This new paradigm shifts the fundamental definition of the technology age from
one of computing, to one defined by data.

▪ Access to data from multiple, distributed sources, in addition to traditional

centralized storage devices will allow users to transparently access information
when and where they need it.
▪ Term “autonomic” is derived from human biology – autonomic nervous system.

➢ Adjusts to many situations automatically without external help.

➢ ANS monitors your heartbeat, checks your blood sugar level and keeps your body
temperature close to 37, without any conscious effort on your part.
Roles of Autonomic Computing in Cloud Computing
There are four areas of Autonomic Computing as defined by IBM. These are as follows:

1. Self-Configuration: The system must be able to configure itself automatically

according to the changes in its environment.
2. Self-Healing: IBM mentions that an autonomic system must have property by
which it must be able to repair itself from errors and also route the functions away
from trouble whenever they are encountered.
3. Self-Optimization: According to IBM an autonomic system must be able to
perform in an optimized manner and ensure that it follows an efficient algorithm for
all computing operations.
4. Self-Protection: the IBM States that an autonomic system must be able to
perform detection, identification, and protection from the security and system
attacks so that systems’ security and integrity remain intact.
Autonomic Computing Architecture

System Models of Autonomic Computing,

Thank you !!