Scientific Software

VAM

The Virtual Amoeba Machine is a communication and application environment for distributed measuring and data processing systems. Additionally, it's suitable for network coupled parallel computer clusters used for solving numeric problems. The VAM is executed on the top of already existing operating systems from the UNIX world, for example Linux or FreeBSD. But it can execute on the native VX-Amoeba kernel, too, a micro kernel best suited for distributed program environments.

The VAM system is mostly implemented and progarmmed with the functional programming language OCaML from the france INRIA software institute. The OCaML programming environment consists of a bytecode compiler and a virtual machine executing the operating- and hardware architecture independent byetcode. The virtual machine hides system dependencies and provides a unique execution environment. Therefore, VAM programs are not dependent from the local operating system executing below. VAM is both a distributed operating system environment and a programming environment with additional interactive features.

With VAM, distributed operating system concepts and functional programming is united. Using this approach, abstract and reliable distributed systems can be programmed and realized.

Activity:

More Informations

VAMNET

VAMNET glues the VAM system and the VX-Amoeba kernel to a complete distributed operating system. Currently, VAMNET consists of these components:

• The VAM development and execution environment on the top of UNIX operating systems, providing Amoeba concepts implemented with functional programming,
• a new Amoeba Fileserver: AFS,
• a new Directory- and Nameserver: DNS,
• a boot server,
• the VAM-shell: VASH, which gives access both to the distributed Amoeba and the local UNIX operating system services,
• a graphical filetransfer frontend xafs, which enables transfer between UNIX and Amoeba and vice versa,
• the native VX-Kernel and some system libraries needed to execute native Amoeba programs including a raw version of VAM,
• AMCROSS: an Amoeba crosscompiling environment for UNIX,
• and finally: AMUNIX, which integrates Amoeba concepts inside UNIX process space (VAM uses AMUNIX)

Activity:

More Informations

AMUNIX

The AMUNIX development and execution environment implements Amoeba basic concepts like remote procedure call and Amoeba multithreading in the UNIX user process space. It consists of a protocol stack FLIP, which executes as a pure user process, and a special Amoeba system library adapted to the UNIX environment. The VAM system builts upon this AMUNIX layer.

Activity:

More Informations

AMCROSS

An Amoeba crosscompiling environment for UNIX. With AMCROSS, the VX-Kernel, native Amoeba libraries and programs can be build.

Activity:

More Informations

VX-Amoeba Kernel

The new VX-Kernel is a strongly enhanced and extended Amoeba microkernel with a large number of new features like:

• user process device drivers,
• a new high performance local commnunication system IPC,
• modified and enhanced FLIP protocol stack (also implemented in AMUNIX),
• a fully revised thread and process management with a new priority driven scheduler,
• and many more.

Activity:

More Informations

VAMLAB

VAMLAB is a distributed scientifc data processing and analysis package with several visualization features. It's mainly used for distributed solving of numeric problems and distributed data measuring and acquisition systems. VAMLAB is based on the predecessor program PsiLAB. VAMLAB is implemented mostly with the functional programming language OCaML, which is a member of the ML programming language class.

Activity:

More Informations?

PSILAB

PsiLAB is a scientific data analyis, visualization and numeric package, similar to INRIAs SCILAB free software. PsiLAB is mostly implemented with the functional programming language OCaML.

Activity:

More Informations

HP-Linda

HP-Linda is a high performance implementation of the parllel programmaing extension Linda for the Amoeba operating system using RPC communication. HP-Linda based on a n-dimensional tuple space. A tuple can be an arbitrary data structure, consisting of different data entries and data types, for example integer values, strings, matrices and so on. Tuples can be stored inthe abstract tuple space using the OUT operation, and can be read (by another program) using the IN operation, either as a copy of the data or destructive.

With this simple modell, an abstract process synchronized data transfer can be realized in distributed systems. This tuple space can be treated like a simple implementation of a distributed and shared memory model.

The tuple space can be build from several tuple servers. Each time, an IN operation is performed in this tuple space, the tuple space will be searched for a matching tuple. In the case there is no matching one was found, the requesting thread or process will be blocked untill a matching data tuple was stored by another thread or process in the tuple space using the OUT operation.

Activity:

More Informations?

FSD-AMOEBA

FSD-Amoeba is an enhanced software distribution version of the original Amoeba operating system from the Vrije University in Amsterdam around the work group of Prof. Anrdew Tanenbaum.

Activity:

More Informations