DEPARTMENT OF

COMPUTER SCIENCE AND ENGINEERING

A paper on

GRID COMPUTING

Prepared by

R.SRI HARISH GOPI, III B.E

R.SATHEESHBABU, III B.E

[email protected]
ABSTRACT Discovery Service (MDS), Globus Resource

Today we are in the Internet world and Allocation Manager(GRAM).

everyone prefers to enjoy fast access to the

Internet.But due to multiple downloading, there is a INTRODUCTION:

chance that the system hangs up or slows down the

performance that leads to the restarting of the entire What's Grid computing? Grid Computing is a

process from the beginning. This is one of the serious technique in which the idle systems in the Network

problems that need the attention of the researchers. and their “ wasted “ CPU cycles can be efficiently

So we have taken this problem for our used by uniting pools of servers, storage systems and

research and in this paper we are providing a layout networks into a single large virtual system for

for implementing our proposed Grid Model that can resource sharing dynamically at runtime. These

access the Internet very fast. By using our Grid we systems can be distributed across the globe; they're

can easily download any number of files very fast heterogeneous (some PCs, some servers, may be

depending on the number of systems employed in the mainframes and supercomputers); somewhat

Grid. We have used the concept of Grid Computing autonomous (a Grid can potentially access resources

for this purpose. in different organizations).

The Grid formulated by us uses the standard Although Grid computing is firmly ensconced

Globus Architecture, which is the only Grid in the realm of academic and research activities, more

Architecture currently used world wide for and more companies are starting to turn to it for

developing the Grid. And we have proposed an solving hard-nosed, real-world problems.

algorithm for laying our Grid Model that we consider

as a blueprint for further implementation. When IMPORTANCE OF GRID COMPUTING:

practically implemented, our Grid provides the user to

experience the streak of lightening over the Internet Grid computing is emerging as a viable
while downloading multiple files. technology that businesses can use to wring more
profits and productivity out of IT resources -- and it's
going to be up to you developers and administrators
Key words:
to understand Grid computing and put it to work.
Grid Security Interface (GSI), Global Access to
It's really more about bringing a problem to
Secondary Storage (GASS), Monitoring and
the computer (or Grid) and getting a solution to that TYPES OF GRID:
problem. Grid computing is flexible, secure, The three primary types of grids and are
coordinated resource sharing among dynamic summarized below.
collections of individuals, institutions, and resources. ➢ Computational Grid
Grid computing enables the virtualization of A computational grid is focused on
distributed computing resources such as processing, setting aside resources specifically for
network bandwidth, and storage capacity to create a computing power. In this type of grid, most of
single system image, granting users and applications the machines are high-performance servers.
seamless access to vast IT capabilities. Just as an
Ø Scavenging grid
Internet user views a unified instance of content via
A scavenging grid is most commonly
the World Wide Web, a Grid user essentially sees a
used with large numbers of desktop machines.
single, large, virtual computer.
Machines are scavenged for available CPU
Grid computing will give worldwide access to cycles and other resources. Owners of the
a network of distributed resources - CPU cycles, desktop machines are usually given control
storage capacity, devices for input and output, over when their resources are available to
services, whole applications, and more abstract participate in the grid.
elements like licenses and certificates.
Ø Data Grid
A data grid is responsible for housing and
For example, to solve a compute-intensive
providing access to data across multiple
problem, the problem is split into multiple tasks that
organizations. Users are not concerned with
are distributed over local and remote systems, and the
where this data is located as long as they have
individual results are consolidated at the end. Viewed
access to the data.
from another perspective, these systems are connected
to one big computing Grid. The individual nodes can
OUR PROPOSED GRID MODEL:
have different architectures, operating systems, and
software versions. Some of the target systems can be
We are using the Scavenging Grid for our
clusters of nodes themselves or high performance
implementation as large numbers of desktop
servers.
machines are used in our Grid and later planning to
extend it by using both Scavenging and data Grid.
Figure1 gives an idea about the Grid that we have Internet via an Intranet (LAN). By using our Grid
proposed. these large numbers of files are distributed evenly to
all the systems in the Network by using our Grid.
For example we have taken into account of a
PROBLEMS DUE TO MULTIPLE
small LAN that consists of around 20 systems out of
DOWNLOADING:
which 10 systems are idle and 5 systems are using
less amount of CPU(for our consideration) and their
CPU cycles are wasted. And our work begins here, as
While accessing Internet most of us might
we are going to efficiently utilize those “wasted CPU
have faced the burden of multiple downloading and in
cycles” into “working cycles”.
particular with downloading huge files i.e., there can
be a total abrupt system failure while a heavy task is
assigned to the system. The system may hang up and
may be rebooted while some percentage of
downloading might have been completed. This
rebooting of the system leads to download of the file WORKING OF THE PROPOSEDGRID:
once again from the beginning, which is one of the
major problems everyone is facing today. When we are downloading multiple files using

Let us consider N numbers of files of different Internet the Grid formulated by us comes in to action.

sizes (in order of several MBs) are being downloaded A dialog box will appear on the Desktop asking the

on a single system (a PC). This will take user whether to use the Grid or not? If the user selects

approximately some minutes or even some hours to “use the Grid”, then automatically the available

download it by using an Internet connection of system resources in the Network are obtained by the

normal speed with a single CPU. This is one of the Globus Toolkit. The configurations of the idle

tedious tasks for the user to download multiple files at systems are noted and the highest configuration

the same time. Our Grid plays a major role here. system gets the highest priority in the priority Queue.
E.g. If there is a supercomputer with 8 CPUs,
another Supercomputer with 5 CPUs and some other
CONCEPT OF OUR PROPOSED GRID:
PCs with P3-2.0GHz, P4-2.0GHz, P4-2.5GHz, P3-
In order to avoid this problem we have
1.0GHz, P3-1.3GHz, P4-1.5GHz, P3-1.13GHz, P4
formulated our own Grid for such an access to the
-2.4GHz are found in the Network. resources information. Most Grid applications contain
Then the order of priority will be: 1. a tight integration of all these components.
Supercomputer-8 CPUs, 2. Supercomputer-5 CPUs, The Globus Project provides open source
3. P4-2.5GHz, 4. P4-2.4GHz, 5. P4-2.0GHz, 6. P3- software tools that make it easier to build
2.0GHz, 7. P4-1.5GHz, 8. P3-1.3GHz, 9. P3- computational Grids and Grid-based applications.
1.13GHz, 10. P3-1.1GHz. These tools are collectively called the Globus Toolkit.
Now the user can click any number of files to Globus Toolkit is the open source Grid technology for
download. The file size of each file is obtained and is computing and data Grids. On the server side, Globus
stored in the priority Queue based on maximum size Toolkit 2.2 provides interfaces in C. On the client
as highest priority. Now the highest priority file is side, it provides interfaces in C, Java language, and
matched with the highest priority system in the other languages. On the client side, the Java interfaces
Network. The files get evenly distributed to their are provided by the Java Commodity Grid (CoG) Kit.
matched “idle systems”. The downloading gets Globus runs on Linux, AIX, HP-UX, Solaris, and also
completed in those systems and these file gets stored on windows operating systems.
in the common database. The authenticated user can The Globus architecture represents a multiple-
access this database and can retrieve his file that he layer model. The local services layer contains the
has downloaded. operating system services and network services like
The various processes that are taking place in TCP/IP. In addition, there are the services provided
our Grid such as authentication, availability of by cluster scheduling software (like IBM Load
resources, scheduling, data management and finally Leveler) -- job-submission, query of queues, and so
job and resource management are viewed by forth. The cluster scheduling software allows a better
following a standard architecture – The Globus use of the existing cluster resources. The higher layers
Architecture. of the Globus model enable the integration of
multiple or heterogeneous clusters.
EMPLOYING THE GLOBUS ARCHITECTURE
ACCESSING THE INTRANET GRID:
IN OUR GRID:

When any user wants to access our proposed

While planning to implement a Grid project,
Intranet Grid in order to download multiple files over
we must address issues like security, managing and
the Internet, then he should follow certain procedures
brokering the workload, and managing data and
that we consider necessary for the security of our Step [6]. Prepare a RSL
Grid. The main Requirements for document
Processing in Grid Environment are: Step [7]. Submit the program
• Security: single sign-on, authentication, using GRAM services
authorization, and secure data transfer. Step [8]. React to status changes
• Resource Management: remote job from GRAM
submission and management. Step [9]. step [10].Get results
• Data Management: secure and robust data via GASS
movement. Here, we have got the resources available in
• Information Services: directory services of the Network which is automatically done by have the
available resources and their status. Globus Toolkit in the server. When we want to
• Fault Detection: Checking the intranet. download a file this information has to be matched
with the client module and then the downloading has
Portability: C bindings (header files) needed to build
to be carried out in the clients.
and compile programs.
Added module:

Existing Algorithm for Globus Architecture:

Step [10]. the Information about files
to be downloaded.
Step [1]. Create security_proxy
Step [11]. GetMatch the files with
via GSI services
appropriate Machines.
Step [2]. Access a MDS-GIIS
Step [12]. Store files in common
server
database.
Step [3]. Search for required
Step [13]. Retrieval of data from
machine(Rank the machine
database is done after proper
list based on a scheduling
authentication.
policy
You'll also see how Grid services and the very
Step [4]. Prepare the data
framework it all rests on is very much like object-
Step [5]. Transfer the data to the
oriented programming.
target machine by using
GASS services
 the files.
[10]. Assign the 1st file of P2 to the 1st node
PROPOSED ALGORITHM FOR OUR in P1.// first node will be node with
INTRANET GRID: highest configuration and first file will
be the file maximum size.

Steps to perform multiple downloading on the [11]. Start processing // files directed to the

Grid. The host details are got from the server of the appropriate system for accessing

LAN in order to identify the various hosts. The host their wasted CPU cycles.

information is got whenever needed on the priority [12]. Loop

queue basis. [13]. Else:

//module for downloading files [14]. Start timer

[1]. Start lookup // look for file size and [15]. Delay  1 min

resource information [16]. Collect incoming files // the files that

[2]. Declare nres, nfile // no of resources the user clicked to download in this

available and no of files duration.

[3]. Input nres, nfiles [17]. Assign the files  P2

[4]. Input size // the file size [18]. Goto step 8

[5]. Initialize P1  res info // store the [19]. Goto step 1

resource information in priority queue [20]. End // when the user exits from

P1 with highest system configuration as proposed Grid.

priority
[6]. Initialize P2  file size // store the file
information in the priority queue P2
with maximum file size as priority
[7]. If condition (nfiles == nres) // check
whether the no of resources is equal
to no. of files
[8]. Initialize counter
[9]. For (counter =1; counter <= nres;
counter++) // initialize the loop to assign
CONCLUSION:

Grid computing was once said to be fading out

but due to the technological convergence it is
blooming once again and the Intranet Grid we have
proposed adds a milestone for the Globalization of
Grid Architecture, which, leads to the hasty
computing that is going to conquer the world in the
nearest future. By implementing our proposed
Intranet Grid it is very easy to download multiple
files very fast and no need to worry about the security
as we are authenticating each and every step taking
place in our Grid and in particular user to access the
database. Further implementations could be carried
out in the nearest future.

REFERENCES

1. www.wikipedia.com
2. www.google.com
3. www.gridcomputing.com