CLUSTER COMPUTING INTRODUCTION Cluster is a widely used

CLUSTER COMPUTING

INTRODUCTION Cluster is a widely used term meaning independent computers combined into a unified system through software and networking Clusters are typically used for High Availability (HA) for greater reliability or High Performance Computing (HPC) to provide greater computational power than a single computer can provide. Clusters are composed of many commodity computers, linked together by a high-speed dedicated network

Cluster categorization • High-availability (HA) clusters • Load balancing clusters • High-performance (HPC) clusters • Grid Cluster

Cluster Architecture A cluster is a type of parallel or distributed processing system that consists of a collection of interconnected stand-alone computers working together as a single, integrated computing resource

Cluster Computing Features • • • Network technologies Network Types Communication Protocols Operating system Single System Image (SSI) Quorum

Production of High Throughput Cluster Computing Applications • Divide and Conquer • Data Management • Shared Storage Architectures

HOMER CLUSTER Used for e-mail and information resources Architecture A cluster is made up of six components • Compute servers • File servers • Password servers • Application servers • Mail servers • Reference systems

Hardware Topology :

Dell high performance computing combines multiple Symmetric Multi-Processor (SMP) computer systems together with high-speed interconnects to achieve the rawcomputing power of supercomputers HPCC Architecture

HPCC Building Block

BENEFITS • scale to very large systems • replacing a "bad machine" within a cluster is trivial yields much higher Availability LIMITATIONS • • • Typically latency is very high and bandwidth relatively low. Currently there is very little software support for treating a cluster as a single system. Problems exist in the interactions between mixed application workloads on a single timeshared computer APPLICATIONS • execution platform for a range of application classes • to execute many Internet applications • execution environments for applications such as weather modeling, automobile crash simulations, life sciences, computational fluid dynamics etc.

Comparing with other distributed computing Characteristic Cluster Grid P 2 P Resource Management (i. e. memory, objects, storage, network access, etc) Centralized Distributed Resource Ownership Singular (Often locked to a single node to prevent data corruption) Singular or multiple, varies from platform to platform Singular, multiple, or distributed, depending on circumstance and architecture Method of Resource Allocation / Scheduling Centralized, allocated according configuration Decentralized N/A, there is no single permanent host for centralized data or resource management. Everything is transient. External Representation Single Image Single or multiple image(s) Unknown, it is circumstantial Inter-Operability Guaranteed within a cluster Enforced within a framework Multiple competing standards Suggested Equipments Mostly high-end, high capability systems High-end or commodity systems Any type, including wireless device and embedded systems. Scaling 2 - 16 way (Although, theoretically 128+ is possible) Two to thousands units connection Theoretically, infinite (In actuality, it depends on network backbone transmission speed, number of clients, and type of transmission protocol…. . ) Discovery Mechanism Defined membership (Static or Dynamic) Centralized index, as well as, multiple decentralized mechanisms. Always decentralized discovery mechanism.

CONCLUSION Cluster computing has become a major part of many research programs because the price to performance ratio of commodity clusters is very good. Also, because the nodes in a cluster are clones, there is no single point of failure, which enhances the reliability to the cluster.

THANK Q