| Distributed Computing Development Platforms |
| Platform Logo | Platform Information |
|---|---|
 |
Cosm "is
a set of open
protocols and applications designed to allow computers all over the world to
work together on projects." It can be used to create any kind of distributed
computing based research project. It was formally launched on September 21,
2000. Development of Phase 2 of the project began in 2003, and is in alpha
testing as of November, 2006. Phase 2 contains "CosmFS, the full security
system, more utility layer, and a far more advanced client."
See a list of third-party distributed projects using Cosm. |
 |
Sun Microsystems' N1 Grid Engine finds idle resources on a LAN of Sun Solaris systems and uses them for your distributed application. See a September 27, 2000 press release about it. N1 Grid Engine is available for Solaris, Linux, Mac OSX, AIX, HP/UX, Irix, and Windows Linux. |
 |
Parabon Computation has a great developer resource page with links for a demo version of their Frontier Software Development Kit (SDK), an API and other documentation about the Frontier SDK, sample applications built on the SDK, a developer discussion list and a link for technical support, and a link to resources for Java development. |
| The Programmer's Playground provides free software tools for application developers and end users to create non-commercial, Internet-based, distributed computing applications. Commercial users can use it through a special license agreement. | |
|
Dcom "is
a skeleton set of C programs and files which allows you to convert your
favorite long running mathematical problem programs into distributed
computing applications. It supports signing for safe automatic program
executible updates, runs on X86 Linux and Windows, and its clients are
self-installing executibles."
This is a bare-bones development framework with no readme or instructions on how to use it. You will also need the following in order to use it:
Version 1.13 is available as of November 13, 2001. | |
| The Dispense Package is a free, "out-of-the-box" software package which enables you to quickly and easily set up a simple distributed computing project to run on the Internet. It was designed on FreeBSD and should work on Linux. It isn't yet ported to other flavors of Unix or Windows. | |
| Fida is "a simple framework of developing and deploying independently distributed applications that can harness otherwise idle computing processors across the Internet. It follows the standard client-server model based on TCP/IP protocols. Its component-based architecture making it efficient and flexible to extend Fida to a wide range of distributed scientific and engineering applications." It is based on the C language, but you can use any language that can be wrapped in or called by C code. | |
 |
DOGMA is a research tool being developed at Brigham Young University. Based on Java, it provides globally-accessible web-based user interfaces for distributed computing applications, firewall and proxy server support, integrated security features, the ability to manage dedicated clusters and supercomputers for closed-access projects, the ability to customize the software for the underlying hardware, and many other features. |
 |
"The Grid Computing and Distributed Systems (GRIDS)
Laboratory at the University of Melbourne is actively engaged in the
design and development of next-generation parallel and distributed
computing systems and applications. The Lab's flagship "Open Source" project,
called the Gridbus Project, is developing
technology that enables GRID computing and
BUSiness. The Gridbus project team is developing cluster and grid
technologies (middleware, tools, and applications) that deliver
end-to-end quality of services depending on user requirements. They
include Economic Grid Scheduler, Cluster Scheduler (Libra), Grid
modeling and simulation (GridSim), Data Grid broker, GridBank, and GUI
tools for workflow management and composition of distributed
applications from (legacy) software components. The Gridbus scheduling
system aggregates or leases of services of distributed resources
depending on their availability, capability, performance, cost, and
users quality-of-service requirements. The Gridbus technology
development is driven by requirements of various applications including
Drug Design, High Energy Physics, and Brain Activity Analysis. The
World-Wide Grid (WWG) testbed used in this research contains resources
from organizations around the globe.
"The Gridbus project sponsorers include: Sun Microsystems Inc., Victorian Partnership for Advanced Computing, University of Melbourne, and Australian Research Council. The project welcomes collaboration from indviduals and organisations around the world, who are interested in getting involved in this open source R&D project." |
| The ECMNet client/server software was originally designed for the ECMNET project. The software has been used successfully for other projects such as Minimal Equal Sums of Like Powers. Version 2.0k of the software (including source code) is available as of June 21, 1999. | |
 |
cyBrain is
"a cross-platform, C++ framework to run complex calculations on a cluster of
computers.... Algorithms are implemented as plugins [which] can be placed in
a GUI (Graphical User Interface), linked with the mouse" and executed
immediately. "If other computers are connected to the cyBrain network, the
plugin can be distributed on those machines to gain more computing power. The
networking stack of cyBrain supports IPv4 and IPv6 both as client and server,
to reduce network load it can use IP multicasting to distribute high bandwith
data like video streams."
"As a first concrete implementation cyBrain provides plugins for image processing based on genetic algorithms: A video stream is captured on a firewire bus and sent to the clients. On the client a genetic algorithm is locating objects in the video frame, the object itself can be described as a simplified vector-image. The calculation is distributed on a cluster of computers to increase frame rate." Version 0.7-9 of the software is available as of December 13, 2002. |
 |
Q2ADPZ
(Quite Advanced Distributed Parallel Zystem) "is an
open source implementation of a system for distributed computing. The system
allows the management/use of the computational power of idle computers in a
network. The users of the system can send computing tasks to these computers
to be executed, which can be in the form of a dynamic library, an executable
program or any program which can be interpreted (Java, Perl, etc.). Platforms
supported are Linux, Unix, Win32 and MacOS X.
"The system is a client-master-slave architecture, using message based communication. Messages between the components of the system are in XML format, and can optionaly be crypted for security reasons." The platform is in active development, with a new version available almost every day. |
 |
XtremWeb
is a distributed computing infrastructure/platform which
allows people to create their own GRID or public projects to process many
different kinds of problems. All project software for the servers, the
clients, and the website, is open-source, is free, and is available for
download through the website.
On January 12, 2004, the platform successfully passed the 1 million jobs per week test. Version 1.2.rc7 of the Linux client was last updated on December 19, 2003. |
| GnutellaHTC will "use the processing power and resources of computers that are already connected to existing P2P file sharing networks" for High-Throughput Computing, "by spreading tasks for remote execution." High-Throughput Computing is a computing process which uses a large amount of computing power for a long period of time. The platform is in development. | |
 |
Alchemi [.NET Grid Computing Framework]
is "a grid computing framework for Windows with the
primary goal of being easy to use. It provides a) a programming environment
to develop grid applications and b) the runtime machinery to construct grids
and run grid applications.
"Alchemi allows flexibility in application composition. You can employ the powerful grid thread model to write grid applications in any .NET-supported language. Alternatively, existing applications can be grid-enabled either programmatically or declaratively. "Constructing uni-level grids (clusters) is as simple as installing the Alchemi Manager on one machine and installing the Alchemi Executor on each machine that is to be part of the grid. The flexible execution model allows Executors to be configured as 'dedicated' (execution is managed by the Manager) or 'non-dedicated' (execution is managed by the user or triggered by a screen saver). "A web services interface allows interoperability with custom grid middleware. "A console application allows you to monitor applications and submit applications grid-enabled using the grid job model for execution. Sample applications including a parallel Mandelbrot set generator are included. "Alchemi is open source, released under the GPL license. "Alchemi is being developed as part of the Gridbus Project at the Grid Computing and Distributed Systems Laboratory at the University of Melbourne, Australia." Version 0.6.2 of the software is available as of April 16, 2004. Join a discussion mailing list about the platform. |
 |
Eluder
is "Extreme Lightweight Utilization of DisparatE
Resources, ... a pluggable client/server application designed to allow
for quick and efficient implementation of collective computing projects,
enabling teams of various sizes and resources to make use of Idle CPU and
memory in their organization or of Internet users." It is designed to be
an entry-level alternative to BOINC.
The platform consists of a Win32 client and an Apache/ModPerl server with MySql backend. It is an open-source project freely available under GPL. Version 0.1 of the software is available as of July 25, 2004. |
 The Karoo Project is a development framework which allows a single software application to be scaled across multiple systems. It will enable software developers to easily create distributed computing applications and other applications which can be scaled to any size. The framework provides queues as a convenient form of multi-threading and provides a messaging and management system which allow many applications to cooperate as one virtual application. The project will eventually implement a "cloud" style front end (it will use a javascript library for application nodes to communicate with a single proxy node via http: the proxy node will convert HTTP messaging to karoo-style messaging). As of September 2, 2009, the project is in active development, in pre-alpha stage. The project owner welcomes others to join in the development of the framework and the project website. |
| Top... |
