Network integration of simulations

[BaLrOg_70]

https://forums.mudspike.com/t/vbs3-steel-beasts-motion-simulation-flight-sim-combo/3757

 

last year this happend?

[Gibsonm]

Sure, but there is a lot of work involved forming a federation of simulations.

 

Making sure the terrain is the same, entities have the same enumeration data, the DIS / HLA bridge is set up correctly, ...

 

Its not just "plug and play".

 

It can be setup for say I/ITSEC or ITEC (a controlled environment where you can test / rehearse for months prior) but certainly not straight forward for people in different parts of the world to hook up for TGIF.

[BaLrOg_70]

I think if we did easy we wouldn't be here  but yes agree it takes a degree of smarts but one thing that is clear is the technology is converging   

[Ssnake]

Not sure if I would call this a "convergence" of technologies, rather a topic of interoperability and integration. These may sound like buzzphrases but they are not. Interoperability is what you get when you have progressed far enough with the integration effort. Integration requires effort from many (if not all) of the parties involved, and it must be coordinated. This is probably the bigger problem than the technological challenges.

But even then,

• you need correlated terrain across all simulation domains,
• you need correlated entities (down to the individual bullet, literally) - an often overlooked aspect,
• you probably want to have correlated animations and postures, and possibly even correlated lifeforms. A tank is one thing, if you add wildlife you invite a whole host of new issues.

 

Then there's the problem of synchronization, network bandwidth, efficiency of network protocols, lag, ... It's very easy to dismiss these topics as "minor issues" but they can have a very noticeable effect on user acceptance.

 

How easy is it to set up a network exercise involving five different simulations?

We successfully demonstrated that two years ago with the exercise "Danish Thunder", but in order to get it going for one afternoon we had to fly in five software engineers from two or three continents, and to prepare the whole thing for days. At which point the average army instructors says "nice toys, but I can't work with this" - and he would be right. What's needed is something that people will perceive as effectively "one product" where you have a controlled procedure that lets you hook all the simulations that you need into the same exercise. And to get that working requires a lot more effort and dedication (and continuity of vision by the parties involved) than what can typically be observed.

We are, after all, talking about commercial software development. This means that if the developers and integrators can't make a profit during the experimentation phase (because a customer with very deep pockets is willing to pay for the horrendous bills, for years), it will only happen if the integration effort can be done "on the side", and/or if the companies involved see a tremendous business opportunity (in which case it's called an "investment"). Also, you can't partner up with companies who want to rule the world with their own product.

[Red2112]

Without even going as far as "interoperabillity" (I think Buzz is a hero ), you can find some of these issues, and not so minor, with one single title/product!

 

In this example, what you see is not always what you get...

 

 

and it´s a known issue that even triple A titles can have.

https://krisredbeard.wordpress.com/articles/true-first-person-viewpoint/

 

Red

 

[Gibsonm]

On 11/25/2018 at 7:45 AM, BaLrOg_70 said:

I think if we did easy we wouldn't be here  but yes agree it takes a degree of smarts but one thing that is clear is the technology is converging   

 

I've attached a UNCLAS PDF detailing how "not simple" it can be:

 

The image on the left hand side of slide 8 gives an idea of the possible partners in the federation and how they are meant to share information.

 

 

 

SB Pro, VBS, DCS, etc. live in the "Gaming" oval.

 

JSJ7_Weter.pdf

Edited November 25, 2018 by Gibsonm

[Apocalypse 31]

6 hours ago, Gibsonm said:

 

I've attached a UNCLAS PDF detailing how "not simple" it can be:

 

 

It's never simple when there's purple.

[Ssnake]

I thought this was interesting enough to warrant a thread of its own.

[BaLrOg_70]

Death by power point is always a good way to seal the purse strings  

 

Thanks Ssnake I wish I had noticed the thread before I replied to Gibson's pm but I will offer up some bullet points that may well be worth considering as a future tech 

 

Thin Client 

 

Virtualisation of Servers 

 

Cloud solutions such as Azure / Lumberyard 

 

Why buy a cloud when you can host your own (from a government perspective that is)  

 

https://shadow.tech/gben 

 

I would offer my referal code but I probably breach some terms of usage and its black Friday so if you jump in now you might get a good deal (be aware it takes them time to spin up your machine) and please read the small print 

 

With regards to the next iteration of SB and the demands the new terrain model has it might well be the only way the client can keep up with the technological requirement?

Edited November 26, 2018 by BaLrOg_70

[Gibsonm]

Cloud solutions are not viable due to the security classification.

 

We already use VMs

 

You can't use thin clients - try running 300+ machines generating 60+ FPS and 1080p running SB or VBS over a thin client architecture.

 

Anyway I'm leaving it there.

[BaLrOg_70]

well you can lead a horse .....  

[Gibsonm]

9 minutes ago, BaLrOg_70 said:

well you can lead a horse .....  

Mate, don't paint the situation as if the Military is ignoring the genius in the room.

 

Lots of people are working on these nirvana type solutions but they just don't work at the moment.

[BaLrOg_70]

Quite the contrary mate I am sure its not the 'military' that are the issue