Escolar Documentos
Profissional Documentos
Cultura Documentos
K.Sreekala(2007-2011)
IIIrd IT
From XENIX group
CONTENTS:
Ø Abstract
Ø Introduction- grid computing
Ø Importance of grid computing
Ø Types of grid
Ø Our proposed grid model
Ø Employing the globus architecture in our proposed grid
Ø Accessing the proposed grid
Ø Algorithm for proposed grid
Ø Implementation our proposed algorithm
Ø Conclusion
Ø Bibliography
is the only Grid Architecture currently
used world wide for developing the
Grid. And we have proposed an
algorithm for laying our Grid Model that
GRID COMPUTING-THE HASTY we consider as a blueprint for further
implementation. When practically
COMPUTING TO ACCESS
implemented, our Grid provides the user
INTERNET
to experience the streak of lightening
ABSTRACT: over the Internet while downloading
Today we are in the Internet multiple files.
world and everyone prefers to enjoy fast
access to the Internet. But due to Key words:
multiple downloading, there is a chance Grid Security Interface (GSI), Global
that the system hangs up or slows down Access to Secondary Storage (GASS),
the performance that leads to the Monitoring and Discovery Service
restarting of the entire process from the (MDS), Globus Resource Allocation
beginning. This is one of the serious Manager (GRAM).
problems that need the attention of the INTRODUCTION :
researchers. What's Grid computing? Grid
So we have taken this problem Computing is a technique in which the
for our research and in this paper we are idle systems in the Network and their “
providing a layout for implementing our wasted “ CPU cycles can be efficiently
proposed Grid Model that can access the used by uniting pools of servers, storage
Internet very fast. By using our Grid we systems and networks into a single large
can easily download any number of files virtual system for resource sharing
very fast depending on the number of dynamically at runtime. These systems
systems employed in the Grid. We have can be distributed across the globe;
used the concept of Grid Computing for they're heterogeneous (some PCs, some
this purpose. servers, maybe mainframes and
The Grid formulated by us uses supercomputers); somewhat autonomous
the standard Globus Architecture, which
(a Grid can potentially access resources Just as an Internet user views a unified
in different organizations). instance of content via the World Wide
Although Grid computing is Web, a Grid user essentially sees a
firmly ensconced in the realm of single, large, virtual computer.
academic and research activities, more Grid computing will give
and more companies are starting to turn worldwide access to a network of
to it for solving hard-nosed, real-world distributed resources - CPU cycles,
problems. storage capacity, devices for input and
output, services, whole applications, and
IMPORTANCE OF GRID more abstract elements like licenses and
COMPUTING: certificates.
Grid computing is emerging as a For example, to solve a compute-
viable technology that businesses can intensive problem, the problem is split
use to wring more profits and into multiple tasks that are distributed
productivity out of IT resources -- and over local and remote systems, and the
it's going to be up to you developers and individual results are consolidated at the
administrators to understand Grid end. Viewed from another perspective,
computing and put it to work. these systems are connected to one big
It's really more about bringing a computing Grid. The individual nodes
problem to the computer (or Grid) and can have different architectures,
getting a solution to that problem. Grid operating systems, and software
computing is flexible, secure, versions. Some of the target systems can
coordinated resource sharing among be clusters of nodes themselves or high
dynamic collections of individuals, performance servers.
institutions, and resources. Grid
computing enables the virtualization of TYPES OF GRID:
distributed computing resources such as The three primary types of grids
processing, network bandwidth, and and are summarized below:
storage capacity to create a single system
image, granting users and applications Ø Computational Grid
seamless access to vast IT capabilities.
A computational grid is focused and data Grid. Figure1 gives an idea
on setting aside resources about the Grid that we have
specifically for computing power. In proposed.
this type of grid, most of the
machines are high-performance PROBLEMS DUE TO MULTIPLE
servers. DOWNLOADING:
While accessing Internet most of
us might have faced the burden of
Ø Scavenging grid multiple downloading and in particular
A scavenging grid is most with downloading huge files i.e., there
commonly used with large numbers can be a total abrupt system failure while
of desktop machines. Machines are a heavy task is assigned to the system.
scavenged for available CPU cycles The system may hang up and may be
and other resources. Owners of the rebooted while some percentage of
desktop machines are usually given downloading might have been
control over when their resources are completed. This rebooting of the system
available to participate in the grid. leads to download of the file once again
Ø Data Grid from the beginning, which is one of the
A data grid is responsible for major problems everyone is facing
housing and providing access to data today.
across multiple organizations. Users Let us consider N numbers of
are not concerned with where this files of different sizes (in order of
data is located as long as they have several MBs) are being downloaded on a
access to the data. single system (a PC). This will take
approximately some minutes or even
OUR PROPOSED GRID MODEL: some hours to download it by using an
We are using the Scavenging Internet connection of normal speed with
Grid for our implementation as large a single CPU. This is one of the tedious
numbers of desktop machines are tasks for the user to download multiple
used in our Grid and later planning to files at the same time. Our Grid plays a
extend it by using both Scavenging major role here.
CONCEPT OF OUR PROPOSED
GRID:
In order to avoid this
problem we have formulated our own
Grid for such an access to the Internet
WORKING OF THE
via an Intranet (LAN). By using our Grid
PROPOSEDGRID:
these large numbers of files are
When we are downloading
distributed evenly to all the systems in
multiple files using Internet the Grid
the Network by using our Grid.
formulated by us comes in to action. A
For example we have taken into
dialog box will appear on the Desktop
account of a small LAN that consists of
asking the user whether to use the Grid
around 20 systems out of which 10
or not? If the user selects “use the Grid”,
systems are idle and 5 systems are using
then automatically the available system
less amount of CPU(for our
resources in the Network are obtained by
consideration) and their CPU cycles are
the Globus Toolkit. The configurations
wasted. And our work begins here, as
of the idle systems are noted and the
we are going to efficiently utilize those
highest configuration system gets the
“wasted CPU cycles” into “working
highest priority in the priority Queue.
cycles”.
E.g. If there is a supercomputer
FIGURE 1: LAYOUT OF OUR
with 8 CPUs, another Supercomputer
INTRANET GRID: with 5 CPUs and some other PCs with
P3-2.0GHz, P4-2.0GHz, P4-2.5GHz, P3-
1.0GHz, P3-1.3GHz, P4-1.5GHz, P3-
1.13GHz, P4 -2.4GHz are found in the
Network.
Then the order of priority will security, managing and brokering the
be: 1. Supercomputer-8 CPUs, 2. workload, and managing data and
Supercomputer-5 CPUs, 3. P4-2.5GHz, resources information. Most Grid
4. P4-2.4GHz, 5. P4-2.0GHz, 6. P3- applications contain a tight integration of
2.0GHz, 7. P4-1.5GHz, 8. P3-1.3GHz, 9. all these components.
P3-1.13GHz, 10. P3-1.1GHz. The Globus Project provides
Now the user can click any open source software tools that make it
number of files to download. The file easier to build computational Grids and
size of each file is obtained and is stored Grid-based applications. These tools are
in the priority Queue based on maximum collectively called the Globus Toolkit.
size as highest priority. Now the highest Globus Toolkit is the open source Grid
priority file is matched with the highest technology for computing and data
priority system in the Network. The files Grids. On the server side, Globus
get evenly distributed to their matched Toolkit 2.2 provides interfaces in C. On
“idle systems”. The downloading gets the client side, it provides interfaces in
completed in those systems and these C, Java language, and other languages.
file gets stored in the common database. On the client side, the Java interfaces are
The authenticated user can access this provided by the Java Commodity Grid
database and can retrieve his file that he (CoG) Kit. Globus runs on Linux, AIX,
has downloaded. HP-UX, Solaris, and also on windows
The various processes that are operating systems.
taking place in our Grid such as The Globus architecture
authentication, availability of resources, represents a multiple-layer model. The
scheduling, data management and finally local services layer contains the
job and resource management are operating system services and network
viewed by following a standard services like TCP/IP. In addition, there
architecture – The Globus Architecture. are the services provided by cluster
EMPLOYING THE GLOBUS scheduling software (like IBM Load
ARCHITECTURE IN OUR GRID: Leveler) -- job-submission, query of
While planning to implement a queues, and so forth. The cluster
Grid project, we must address issues like scheduling software allows a better use
of the existing cluster resources. The · Portability: C bindings
higher layers of the Globus model enable (header files) needed to build
the integration of multiple or and compile programs.
heterogeneous clusters.