(SEM VII) THEORY EXAMINATION 2023-24 HIGH PERFORMANCE COMPUTING

KCS073 – HIGH PERFORMANCE COMPUTING

B.Tech (SEM VII) – Exam-Oriented Answers

SECTION A

(2 Marks × 10 = 20)

a) Cluster computing vs Grid computing

Cluster computing uses tightly coupled computers in a single location, while grid computing connects geographically distributed, loosely coupled systems to share resources.

b) High performance grid

A high performance grid is a distributed computing system that integrates multiple high-speed resources to solve complex computational problems efficiently.

c) Capabilities of Open Grid Service Architecture (OGSA)

OGSA supports service discovery, resource management, security, interoperability, and scalable grid services.

d) Security issues in grid computing

Security issues include authentication, authorization, data integrity, confidentiality, and secure resource sharing across multiple organizations.

e) Cluster in computing environment

A cluster is a group of interconnected computers that work together as a single system to improve performance, availability, or scalability.

f) Cluster middleware

Cluster middleware is software that enables communication, coordination, and resource management among nodes in a cluster.

g) Advantages of parallel programming

Parallel programming improves performance, reduces execution time, enhances scalability, and efficiently utilizes hardware resources.

h) Advantages of Beowulf model

Beowulf clusters are low-cost, scalable, use commodity hardware, and provide high performance for parallel applications.

i) Types of cloud models

Cloud models include Public Cloud, Private Cloud, Hybrid Cloud, and Community Cloud.

j) Advantages of cloud computing

Cloud computing offers scalability, cost efficiency, flexibility, high availability, and reduced infrastructure maintenance.

SECTION B

(10 Marks × Any 3 = 30)

a) Architecture of high performance grid

High performance grid architecture consists of resource layer, middleware layer, service layer, and application layer.
Advantages: Resource sharing, scalability, cost efficiency
Applications: Scientific simulations, weather forecasting, bioinformatics

b) OGSA capabilities and security concerns

OGSA provides standardized service interfaces, resource discovery, and lifecycle management.
Security concerns include identity management, secure communication, access control, and trust management.

c) Role of resource management in HPC

Resource management allocates CPU, memory, storage, and network efficiently. It ensures load balancing, scheduling, fault tolerance, and optimal system utilization.

d) Role of MPL in parallel programming

Message Passing Libraries (MPI/MPL) allow processes to communicate via messages. They support synchronization, data exchange, and scalability in distributed systems.

e) Cloud deployment models

Deployment models include public, private, hybrid, and community clouds. Each model varies in cost, control, security, and scalability.

SECTION C

(10 Marks × Any 1 = 10)

3(a) Models and protocols in grid computing

Grid computing uses computational grids, data grids, and service grids.
Protocols include GridFTP, GRAM, GSI, and MDS for data transfer, job execution, security, and resource discovery.

3(b) Peer-to-Peer (P2P) computing

P2P computing allows nodes to act as both clients and servers.
Advantages: Scalability, fault tolerance, resource sharing, decentralization.

SECTION D

Job schedulers in GT4 architecture

GT4 uses schedulers like PBS, LSF, Condor, and SGE.
Advantages: Efficient job execution, load balancing, resource optimization.

GSI security layer

Grid Security Infrastructure (GSI) provides authentication, authorization, encryption, and secure communication using certificates.

SECTION E

Cluster analysis                                              Cluster analysis groups similar data points.
 

Steps to cut hierarchical clustering:

Build dendrogram                                          Choose threshold

Cut tree                                                           Form flat clusters

Cluster computing and types                         Cluster computing combines multiple systems to act as one.
 

Types:                                                            High Performance Clusters

Load Balancing Clusters                                 High Availability Clusters