The MOSIX multicomputer operating system for high performance cluster computing

Barak, Amnon; Laadan, Oren

doi:10.1016/s0167-739x(97)00037-x

Cited by 190 publications

(95 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Commodity clusters have received significant attention in bioinformatics, because of their low-budget elements and scalability in respect of the user's requirements [36], [37]. Commodity clusters consist of regular computers (servers), connected through the network, compared to a supercomputer containing many processors.…”

Section: A Commodity Clustersmentioning

confidence: 99%

Biomarker discovery, high performance and cloud computing: A comprehensive review

Blayney

Haberland

Lightbody

et al. 2015

2015 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

View full text Add to dashboard Cite

Abstract-The analysis of biological markers (biomarkers) have the ability to improve clinical outcomes of a disease through prediction and early detection. The constant improvement of next generation sequencing (NGS) technologies coupled with falling equipment price are driving research and application especially in the medical domain. NGS technology is being applied to cancer research promising greater understanding of carcinogenesis. However, as sequence capacity grows, algorithmic speed is becoming an important bottleneck. To understand these challenges we present a review on difficulties currently faced in discovering biomarkers. The review places the spotlight on sequencing technologies and the bottlenecks encountered in these pipelines. Cloud computing and high performance computing technologies such as Grid are summarized in tackling computational challenges presented by technologies and algorithms used in biomarker research.

show abstract

Section: A Commodity Clustersmentioning

confidence: 99%

Biomarker discovery, high performance and cloud computing: A comprehensive review

Blayney

Haberland

Lightbody

et al. 2015

2015 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

View full text Add to dashboard Cite

show abstract

“…The load balancing algorithms were designed from scratch after studying the underlying architecture of this computing cluster and the set of tasks for which it would be used for. Some examples of the tools for cluster computing are MOSIX [41] and [42]. Mosix is a set of adaptive resource sharing algorithms that are geared for performance scalability in a computing cluster of any size, where the only shared component is the network.…”

Section: Linux Cluster Setupmentioning

confidence: 99%

Deploying CPU Load Balancing in the Linux Cluster Using Non-Repetitive CPU Selection

Jameel¹,

Murugananth²,

Chingtham³

2009

IJCEE

View full text Add to dashboard Cite

Abstract-Maintaining load balancing in a computing cluster is an evident problem in distributed systems and research in this field is not new. The challenges in designing the load balancing algorithms are immense. This paper lists some of those challenges in the design of CPU load balancing algorithm and provides solutions to some of them. The algorithm considers one node in the cluster as the Master Server and another as the Load Balancer. The master server maintains the CPU and IP information of each machine. The nodes in the cluster send their CPU status and IP information to the master server after every 30 seconds. The implementation solves "readers-writers" problem exclusively using sockets. If a number of requests are sent before the next central database update, the load balancer selects other less busy nodes in the cluster. This ensures that all nodes are allocated with the new tasks coming from remote systems, thereby maintaining a load balance among the CPUs. This implementation is highly fault tolerant and reliable, guaranteeing a high probability of task completion. Results show that this scheme handles task allocation in much optimized way and with fewer overheads. The implementation can handle CPUs ranging in numbers from 1 to 255.

show abstract

“…After function registration, a trampoline is installed in the beginning of the original function v1. When the function is called (1), execution branches to the execution flow redirection handler (2 tain use counters), and jumps to function v1 copy (3), a clone of the original function. This function image was modified to branch back to the redirection handler (4), where additional post-call bookkeeping is carried out.…”

Section: Execution Flow Redirectionmentioning

confidence: 99%