We are looking for industrial sponsors to help us finish RTZen -
a Real-time CORBA ORB for RTSJ (real--time Java).
Here are some FAQ related to funding this effort:
Similar to the TAO license - truely open-source and free to use
without strings attached. RTZen will be open-source, free to use
in any commercial application free without royalty or fee. RTZen
will remain open-source, but companies can take the open-source RT
ZEN and add value to make commercial version based on the open-source
version without fee.
Yes. Initially, our research group will support RTZen, but since
RTZen is open-source, I am negotiating commercial support with a
company that is currently providing commercial support for other
open-source ORBs (e.g., TAO and JacORB).
Many have expressed interest in having Value-Types and Portable
Interceptors. Portable Intercepters and Value-types are already
implemented in standard ZEN. Both will be available in the
December '04 release of RTZen.
We have a few basic services coded for standard ZEN, such as Naming
and Implementation Repository, Time, Event Service, which we will
move to RTZen as we get to them. We can expidite any service you
want if you are funding this work.
Currently it provides real-time predictability on TimeSys Linix
compiled by TimeSys's RI, Purdue's Open Virtual Machine (OVM) which has both an interpreter and an ahead-of-time compiler
, and Washington University's jRate. If you are donating and you are interested
in a specific platform, we can make it a priority. VXWorks may be
a bit of a problem since we don't have any licenses, but something
can be worked out.
Our goal is for RTZen to be usable in simple RT applications
by December 2004. After that, we will spend about three years
fleshing it out to provide the full RT CORBA specification guided
by configuration to keep memory footprint appropriate to the
application's needs. We will have new RTZen releases every 3 to 6
months as appropriate. The current bleeding version (source and
executable) will be available via our web CVS repository.
We will give you acknowledgement for your donation as well as
priority over what RT CORBA features and services we implement
first.
Funding for a graduate student researcher costs $50k per year.
I currently have a team of five graduate students,
one professional programmer (Peter), and me working on RTZen.
We plan to work hard for the next three years.
The total cost is about $333,400 per year multiplied for three years
totaling about $1,000,000. for the three year effort.
We have a proven ZEN ORB core that is predictable when run on TimeSys
Linux and jTime JVM (although the jTime JVM is currently very slow).
We have also designed ZEN so that the application developer is
shielded from most of the complexity of using RTSJ (real-time Java).
We have also developed techniques for configuration and feature
subsetting that enable shrinking the memory footprint of ZEN to be
proportional to the features used by the application.
We plan regular (6 month) releases along the way which will each
provide a usable subset of RT CORBA specification.
Our goal is to implement the full specification for
Real-time CORBA in RTSJ along with real-time versions of the CORBA
services.
Our funding for RTZen ended December 31, 2004.
At this time we are still seeking funding for RTZen.