1. International Indexed & Refereed Research Journal, January, 2013 ISSN 0975-3486, RNI- RAJBIL- 2009-30097, VOL- IV * ISSUE- 40
Research Paper—Computer Science
TCP/IP Security Protocols for High Performance
Grid Computing Architecture
January,2013 * Abdur Rakib
* University: Cmj University, Meghalaya
A B S T R A C T
Grid computing is a term referring to the combination of computer resources from multiple administrative domains to attain
a common goal. The grid can be thought of as a distributed system with non-interactive workloads that involve a large
number of files. In this paper, we propose a solution for various security issues found in High Performance Grid Computing
Architecture. We analyze different network layers available in Grid Protocol Architecture and identify various security
disciplines at its different layers. We also analyze various Security Suites available for TCP/IP Internet Protocol Architecture.
The paper intends to achieve two major tasks.
Keywords: Reliability, Performance, Grid Computing, Resources, Security, Modeling.
Introduction: detection is an important component of any modern
Grid computing began in the academic security system because it is considered as a second
research community and the national defense industry, line of defense against bugs and security holes as well
where researchers needed to process large amounts of as providing protection against insiders. This thesis
data as quickly as possible for data intensive projects. studies the problem of intrusion detection in Grid
It is an innovative approach that leverages existing IT environments since it is considered as an important
infrastructure to optimize compute resources and security issue. It introduces flexible cooperative
manage data and computing workloads. Using the distributed intrusion detection architecture for
original concept of grid computing, arrays of compu- computational Grids. This work is based on the study
tational power are constructed from a network of many of latest Grid projects and intrusion detection systems
small and widespread computers and used to perform to deliver architecture that suits and benefits form the
large calculations and operations that can be decom- underlying computational Grid evironment.A prototype
posed into independent units of work. This approach implementation of the proposed architecture for the
allows massive computational projects to achieve re- purposes of validation and verification is also
sults that otherwise could not be completed even on introduces.
today's largest computers. As the concept has evolved, According to Foster and Kesselman a grid is
grid computing gained rapid acceptance in the com- a system that conforms to three specific categories: it
mercial marketplace in a manner similar to the emer- coordinates resources that are not subject to centralized
gence of the Internet. Organizations with both large control, it uses standard, open, general-purpose
and small networks have been adopting grid techniques protocols and interfaces, and it delivers nontrivial
in order to reduce execution time as well as to enable quality of service. Next generation scientific exploration
resource sharing. With the rapid advance in science, requires computing power and storage that no single
engineering, and business; people seek more compu- institution alone is able to afford. Additionally, easy
tational power and resources to solve their problems access to distributed data is required to improve the
more efficiently in terms of accuracy, time, secure and sharing of results by scientific communities spread
money. around the world. The proposed solution to these
The field of Grid computing was born to fill the challenges is to enable different institutions, working
gap between available technology and increasing in the same scientific field, to put their computing,
demand for computational power. The Grid provides a storage and data resources together in order to achieve
powerful computational environment by coupling the required performance and scale. Grid is a type of
distributed resources to enable seamless aggregation parallel and distributed system that enables the sharing,
and sharing to create more powerful resource. The term selection, and aggregation of services of heterogeneous
distributed here does not refer only to geographical resources distributed across "multiple" administrative
locations but also to administration that may span domains based on their availability, capability, perfor-
multiple organizations. Security issues were addressed mance, cost, and users' quality-of-service requirements.
from the beginning of the Grid computing because of As Network performance has outpaced computational
their importance to the success of such field. Intrusion power and storage capacity, this new paradigm has
40 RESEARCH ANALYSIS AND EVALUATION
2. International Indexed & Refereed Research Journal, January, 2013 ISSN 0975-3486, RNI- RAJBIL- 2009-30097, VOL- IV * ISSUE- 40
evolved to enable the sharing and coordinated use of unused CPU capacity in all participating machines to
geographically distributed resources. This chapter be allocated to one application that is extremely com-
presents an introduction to Grid giving focus to the putation intensive.
main requirements and challenges that must be ad- Our purpose in this article is to argue that the
dressed in setting up this new paradigm of distributed Grid concept is indeed motivated by a real and specific
computing. problem and that there is an emerging, well-defined
The ancestor of the Grid is Metacomputing. Grid technology base that addresses significant aspects
This term was coined in the early eighties. The idea of of this problem. In the process, we develop a detailed
Metacomputing was to interconnect a collection of architecture and roadmap for current and future Grid
computers held together by state-of-the-art technol- technologies. Furthermore, we assert that while Grid
ogy and "balanced" so that, to the individual user, it technologies are currently distinct from other major
looks and acts like a single computer. The constituent technology trends, such as Internet, enterprise,
parts of the resulting "metacomputer" could be housed distributed, and peer-to-peer computing, these other
locally, or distributed between buildings, even conti- trends can benefit significantly from growing into the
nents. There are new economic as well as business problem space addressed by Grid technologies. The
factors that are contributing to the heightened interest real and specific problem that underlies the Grid concept
in the development and implementation of grid com- is coordinated resource sharing and problem solving
puting. Because of the Internet and the way business in dynamic, multi-institutional virtual
is conducted today, we are inundated with data. As the organizations. The sharing that we are concerned with
data flood gates open wider, the window of opportu- is not primarily file exchange but rather direct access to
nity for capturing and turning this data into informa- computers, software, data, and other resources, as is
tion grows shorter and shorter. required by a range of collaborative problem-solving
Computing applications in many industries and resource brokering strategies emerging in industry,
involve processing large volumes of data and/or per- science, and engineering. This sharing is, necessarily,
forming repetitive computations that exceed existing highly controlled, with resource providers and
server platform capabilities. In order to use data analy- consumers defining clearly and carefully just what is
sis to achieve business intelligence and improve deci- shared, who is allowed to share, and the conditions
sion making, data must be analyzed in a timely manner. under which sharing occurs.
Today's business requirements often demand a much Conclusion:
larger sample size for analysis or perhaps use of the A computer cluster is a group of linked
entire data source for maximum accuracy. Thechallenges computers, working together closely so that in many
that IT shops face today, including budget cuts, server respects they form a single computer. The components
consolidation, hardware provisioning and overall of a cluster are connected to each other through fast
administration, are all factors driving interest in and local area networks. Cluster computing is the technique
implementation of grid computing. of linking two or more computers into a local area
The convergence of recent hardware and network in order to take the advantage of parallel pro-
software advances has made resource virtualization cessing. In this paper, we propose a solution for various
possible and made it easier to construct a grid. On the security issues found in Cluster Computing Architec-
hardware side, these advances include networked ture. We analyze various Security Suites available in
storage devices and low-cost, modular hardware com- TCP/IP Security Protocol Suite. The paper intends to
ponents (such as blades); on the software side, they achieve two major tasks. First, it defines variousSecurity
include improvements in networking, Web services, Disciplines on different levels of Cluster Computing
databases, application servers and management Architecture. Second, it proposes different Security
frameworks. A parallel processing architecture in which Suites available in different layers of TCP/IP Security
CPU resources are shared across a network, and all Protocol Suite applicable for different Security
machines function as one large supercomputer. It allows Disciplines defined in different levels of Cluster
Computing Architecture.
R E F E R E N C E
[1] http://www.cloudbus.org/~raj/papers/gridtech.pdf pg 1-30
[2] http://www.uhd.edu/academic/colleges/sciences/ccsds/pdf/7_Things_About_Grid_Computing.pdf pg 1-2
[3] http://www.cs.kent.edu/~farrell/grid06/lectures/grid01.pdf pg 1-11
[4] http://www.redbooks.ibm.com/redbooks/pdfs/sg246778.pdf 1-268
[5]http://www.nature.com/nature/webmatters/grid/
[6] Steven M. Bellovin et. al. "Probable Plaintext Cryptanalysis of the IP Security Protocols", In Advances in
Cryptology:Proceedings of CRYPTO 96 , pages 1-15.
RESEARCH ANALYSIS AND EVALUATION 41