swarm-modeling
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Swarm-Modelling] Multi-Simulator Interface. ??


From: Scott Christley
Subject: Re: [Swarm-Modelling] Multi-Simulator Interface. ??
Date: Wed, 16 Apr 2003 18:36:11 -0700

Dear John,

I followed your link regarding Syncer, and I was shocked to find out that this concept was patented. Do you know if this patent been released under a public license, specifically GPL compatible? I admit that I haven't read the patent itself, just parts of your SwarmFest paper, but I worry that the patent was awarded a broad definition.

I have been thinking about implementing distributed swarm capability with the help of GNUstep; that is why I have been working hard on the Swarm code these past weeks to get it to work with GNUstep. GNUstep has a Distributed Objects (DO) system, which is of course based upon the Proxy design pattern; something that I think NeXT patented a long time. I suppose there is a point that I'm doing GNUstep while you are doing Java, but still I'm paranoid about patents because Swarm is GPL code; I don't want to get SDG sued!

thanks
Scott

On Monday, April 14, 2003, at 05:43 AM, address@hidden wrote:


HLA is an IEEE standard that grew out of the military distributed wargaming
simulation environment. As such it is heavily loaded with a lot a
facilities necessary to manage real-time distributed wargaming scenarios
and makes assumptions about the communications model that is sometimes
quite limiting. Swee Leong at NIST has done some work to develop an HLA
subset and a set of conventions for using HLA to run distributed simulation
models. Their focus is on factory simulation and they have built HLA
interfaces for most of the discrete event simulation packages typically
used in factory simulation (Arena, ProModel, etc.).

Despite the fact that it is an IEEE standard I would not suggest HLA for connecting Swarm simulations together. There is a lot of overhead involved and the time synchronization approach is very primitive resulting in poor performance (remember their focus is on real-time simulations). In other words, your distributed simulations can often run more slowly than if you
ran them all on a single computer.

I'm not familiar with the LM-ATL MSI and what it is capable of. Hopefully
you are aware of our own work in this area called "Syncer" that we
presented at SwarmFest 2001 (so it's already integrated with Swarm). See http://www.erim.org/swarm for more information. It is a highly optimized algorithm for connecting distributed Swarms together. We have been tuning
the algorithm for its performance for a number of years. Our original
implementation just relied on sockets, and ran wonderfully fast, but lacked
the facilities for registration and marshalling that the standard
infrastructures give you.The latest incarnation for Swarm, used the Java RMI and CORBA infrastructure for supporting the distributed messaging part
of the protocol. Unfortunately, we found performance problems in those
implementations.

We can certainly talk more about this. Hopefully we can help you get to
where you need to go more quickly.

John A. Sauter
Group Leader, Emerging Markets
ph: 734.302.4682 fax: 734.302.4991
Enterprise Solutions Division
Innovative Solutions in Program Management, Analysis and Logistics
Altarum (http://www.altarum.org)
Street: 3520 Green Ct, Ann Arbor, MI 48105
Mail: PO Box 134001, Ann Arbor, MI 48113-4001




                      Paul E Johnson
<address@hidden To: <address@hidden>
                       cc:
Sent by: Subject: [Swarm-Modelling] Multi-Simulator Interface. ??
                      <modelling-admin
                      @swarm.org>


                      04/10/2003 12:21
                      PM
                      Please respond
                      to modelling





I get the freshmeat email newsletter every day and don't know what most
things are for. This one is interesting and it uses lots of the same
words that we use when we talk about Swarm and it speaks to our desire
to run separate simulations that are synced.  But I don't know what HLA
is and thought I better ask you scientific types.

It is from lockheed martin folk:

http://www.atl.lmco.com/proj/csim/msi/

The Multi-Simulator Interface (MSI) is a simulation interconnection
engine. In other words it is a program that connects simulations
together by synchronizing their clocks and data. The MSI serves the same
purpose as HLA and supports most of HLA's functionality (and more). In
that respect it can be considered an HLA alternative. The major
motivating factors in the design of the MSI are speed, interoperability,
and ease of use.

They add this:
     * Time synchronization that allows for proper causality when used
with discrete event simulators.

So, like war, what is it good for?

--
Paul E. Johnson                       email: address@hidden
Dept. of Political Science            http://lark.cc.ku.edu/~pauljohn
1541 Lilac Lane, Rm 504
University of Kansas                  Office: (785) 864-9086
Lawrence, Kansas 66044-3177           FAX: (785) 864-5700

_______________________________________________
Modelling mailing list
address@hidden
http://www.swarm.org/mailman/listinfo/modelling





_______________________________________________
Modelling mailing list
address@hidden
http://www.swarm.org/mailman/listinfo/modelling



reply via email to

[Prev in Thread] Current Thread [Next in Thread]