Facebook betrieb 2009 rund 30.000 Server für 300 Mio Nutzer, mittlerweile hätte sich die Zahl lt. DatacenterKnowledge.com mind. verdoppelt Wie stellt man sich 30.000 Server vor? http://dennis-wisnia.de/wordpress/wp-content/uploads/2010/06/serververgleich.png einen 50er-Block kann man sich als Rack vorstellen
The Ohio Supercomputer Center provides supercomputing, research and educational resources to a diverse state and national community, including education, academic research, industry and state government. At the Ohio Supercomputer Center, our duty is to empower our clients, partner strategically to develop new research and business opportunities, and lead Ohio's knowledge economy.
"ParallelKnoppix is a modified Knoppix live CD designed for use in creating HPC clusters. You can start up PK on multiple nodes to run a cluster, and customize PK to add or remove applications. "
The Large Synoptic Survey Telescope (LSST) is a project to build an 8.4m telescope at Cerro Pachon, Chile and survey the entire sky every three days starting around 2014. The scientific goals of the project range from characterizing the population of largish asteroids which are in orbits that could hit the Earth to understanding the nature of the dark energy that is causing the Universe's expansion to accelerate. The application codes, which handle the images coming from the telescope and generate catalogs of astronomical sources, are being implemented in C++, exported to python using swig. The pipeline processing framework allows these python modules to be connected together to process data in a parallel environment.
Distributed Sage is a framework that allows one to do distributed computing from within Sage. It includes a server, client and workers as well as a set of classes that one can subclass from to write distributed computation jobs. It is designed to be used mainly for ‘coarsely’ distributed computations, i.e., computations where jobs do not have to communicate much with each other. This is also sometimes referred to as ‘grid’ computing.
Ack. Ppython requires worker threads on each cluster node. I want an ssh private key (no p/w) solution. 1) Start parallel python execution server on all your remote computational nodes:
Disco is an open-source implementation of the Map-Reduce framework for distributed computing. As the original framework, Disco supports parallel computations over large data sets on unreliable cluster of computers.
JPPF is an open source Grid Computing platform written in Java that makes it easy to run applications in parallel, and speed up their execution by orders of magnitude. Write once, deploy once, execute everywhere!