Projects

Firebird

Firebird is a relational database offering many ANSI/ISO SQL standard features that runs on Linux, Windows, and a variety of Unix platforms. Firebird offers excellent concurrency, high performance, and powerful language support for stored procedures and triggers. It has been used in production ... [More] systems, under a variety of names, since 1981. Firebird is a commercially independent project of C++ programmers, technical advisors and supporters developing and enhancing a multi-platform relational database management system based on the source code released by Inprise Corp (now known as Borland Software Corp) on 25 July, 2000. Join the Firebird Foundation The software is completely free of any registration, licensing or deployment fees. It may be deployed freely for use with any third-party [Less]

Neo4j - the graph database

Neo4j is a graph database. It is an embedded, disk-based, fully transactional Java persistence engine that stores data structured in graphs rather than in tables. A graph (mathematical lingo for a network) is a flexible data structure that allows a more agile and rapid style of development.

InfoGrid.org

Internet graph database. Represents information as nodes and edges, which may carry properties and may be typed. Selection of storage backends including MySQL, Postgresql, native files and NoSQL. Optional components map URLs directly to objects in the graph database with custom rendering via the ... [More] Viewlet framework. Very easy to access external data sources as always-up-to-date object graphs within InfoGrid. And more ... [Less]

Ganymede

GANYMEDE is a portable and customizable network directory management system, released under the GNU General Public License. Ganymede is written in Java, and requires Java 5 or better. The Ganymede server is designed to run on a UNIX-type system, but can be run on almost any Java-supporting operating systems with the proper support scripts.

txmonitor

The purpose of this project is to create an end-to-end transactional monitor application with following capabilities: Capturing end-to-end transactions across distributed layers of an application architecture. Exporting relevant transaction information to the end-user in a form of a graphical ... [More] visualization (tree like structure for example). The following information about transactions should be made available: execution times and transaction context (parent/child relationship). It is also the aim of the project, to differentiate from other profiling tools by supporting transactional profiling on a SOA systems. Possible techniques can be applied: AOP; Dynamic Bytecode Instrumentation (with the use of BCEL) Notes: (11/12/2007) No source has been made available. Only the creation of the project hosting for now. [Less]

transmem

The aim of this project is to provide a framework for investigating hardware transactional memory (HTM) within the simics full-system simulator, using the x86 architecture. The project aims to achieve the following: Provide a simple module to start working with transactional memory. Ability to ... [More] send events to an external event monitor/simulator. Regression status: STAMP passes (with caveat noted in issue 1) The code here is very early alpha if not pre-alpha... you've been warned If you're still keen to try it, please follow the GettingStarted guide to learn how to use transmem. If you do use transmem please let us know, thanks. [Less]

nbds

C implementations of several scalable non-blocking data structures for x86 and x86-64. StatusOverall the package is pre-alpha. It is still under development. The skiplist is beta. There are no known bugs and it is feature complete. The hashtable is alpha. It has not been tuned for memory usage. ... [More] There are no known bugs and it is feature complete. The lock-free memory management and runtime support is pre-alpha. It is not feature complete. The transactional system is a toy implementation for now. Developers will find it mostly interesting for the concepts. Tested under Linux with gcc 4.1-4.3 with Ubuntu 8.10. Probably works on Mac OS 10.5 with gcc 4.1-4.2. NewsVersion 0.4.3 - 4/7/09 - bug fix in hash table iterator. bug fix for OS X in memory manager. add performance tests to measure steady state behavior. major performance tuning for skiplist. Version 0.4.2 - 1/18/09 - Linux support, gcc 4.1-4.3. Fixed major bug in skiplist. Memory manager re-write. Version 0.4.1 - 12/16/08 - support for 32 bit x86. iterators for hashtable, list, and skiplist Version 0.4.0 - 12/10/08 - lock-free transactional map Version 0.3.2 - 12/3/08 - support for arbitrary key types, with a fast path for integer keys. Version 0.3.1 - 11/29/08 - list and skiplist now support update operations and string type keys Version 0.3.0 - 11/23/08 - lock-free skiplist Version 0.2.1 - 11/15/08 - Started work on making the hash table transactional. Bug fixes too. Version 0.2.0 - 11/07/08 - Initial public release What's InsideLock-Free Transactional MapLock-free transactional key-value store. The transactional semantics follow database-style snapshot isolation using multi-version concurrency control. With snapshot isolation updates are versioned. Transactions are isolated from changes after they begin. They always see a consistent view of the structure. Conversely, a transaction's updates are invisible to all other transactions until it commits. When a transaction tries to commit its writes are checked for conflicts against any committed concurrent transactions. If the system detects a conflict then the committing transaction rolls back. Otherwise its updates become visible to future transactions. Lock-Free SkiplistThe lock-free skiplist data structure created by Maurice Herlihy, Yossi Lev, and Nir Shavit. See Herlihy's and Shavit's book "The Art of Multiprocessor Programming" and Kir Fraser's dissertation "Practical Lock Freedom" http://www.amazon.com/Art-Multiprocessor-Programming-Maurice-Herlihy/dp/0123705916/ http://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-579.pdf I've generalized the data structure to support update operations like set() and CAS() in addition to the normal add() and remove() operations, while preserving the lock-free property and linearizability. Lock-Free HashtableCliff Click's lock-free hash table from http://www.azulsystems.com/events/javaone_2008/2008_CodingNonBlock.pdf http://sourceforge.net/projects/high-scale-lib LicenseAll code in this project is public domain, except the unit test framework (which was not written by me). The hash function (murmur hash 2.0) is not written by me, but is public domain. ContactFeel free to email me with comments or questions: jdybnis at gmail. [Less]