Update as of 1/27/2021: MAK Legion won the DisTec People's Choice Award! We are so grateful for your votes and can't wait to see how Legion disrupts technology!
We're proud that MAK has been shortlisted in ITEC's Disruptive Technology (DisTec) Challenge, a competition showcasing solutions that have the potential to disrupt training and simulation as we know it. Our submission highlights our next-generation scalability and communications framework, MAK Legion, to manage and deliver millions of entities. Take a look at our submission video on the DisTec Challenge site or check out the transcript of my interview below with Len Granowetter, MAK's CTO, as he outlines the how's, why's and, so what's about our new Legion technology. (And while you're learning about MAK Legion, vote for us to win the DisTec People's Choice Award!)
Dan: Len, tell us about the Legion Scalability framework. Is it a disruptive technology?
Len: It is. The Legion technology is one we're most excited about at MAK Technologies. It's a technology we've been developing to solve a problem with scalability and performance for very large-scale simulation exercises – talking about doing millions of entities for a number of different reasons.
Dan: Like bringing the whole entire Air Force to war; having Pattern of Life of civil, maritime; all the airplanes in the entire world all at the same time; simulation in mega cities, things like that?
Len: Yeah, in part, it came out of the US Army's STE program, so that particular customer had a strong desire for scalability. But it's also something we've been talking about in a number of other programs and a number of other applications with customers.
Dan: So Len, tell me – what is Legion?
Len: Legion is a scalability and interoperability communications framework for exchanging information, exchanging state and events between different simulation applications - between virtual simulators, constructive simulations, and other tools that go into a simulation system.
Dan: Sounds like DIS and HLA?
Len: It is similar in terms of where it fits in an architecture diagram, but in some ways what we're doing here is, as a company that's been involved with DIS and HLA for a long time – and obviously still very supportive of those protocols – we asked ourselves, if we had to do it over again based on some of the things that have changed over the last 25 years – cloud technology, large multi-core machines, larger expectations in terms of entity count, high-performance graphics cards, raster graphic networks – we asked ourselves – given those advances in technology, what would that change if we had to design something like HLA over again? We also broadened the problem a bit and said ok, part of the job of scalability, the problem of scalability, isn't just about the network communication. It's about thinking about scalability end-to-end from how do I build multi-thread CGF or SAF applications that can simulate tens of thousands of entities in a single application, and how do they store and access data efficiently from within an application, all the way to how do we get it on the network, how do we communicate with other client applications, maybe thin clients, when you have all the simulations running in the cloud. Those are some of the problems that Legion was designed to solve.
Dan: And I'll point out that we're talking about applications that some generate tens of thousands of entities and some generate just a single entity but it's super high frame rate, like a helicopter simulator. Both of those have to coexist in an architecture that allows them to play in a world with millions of other entities.
Len: In terms of some of the feasibility of this, we regularly do demos with a million Pattern of Life entities. We tend to host a fairly large number of our VR-Forces simulation engines in the cloud on Amazon Web Services (AWS) cloud network, and then we connect any client we want, whether it's a client we build ourselves or one from another vendor, and we tie those together and have any application be able to participate in that million-entity scenario.
Dan: Have any other vendors actually used this technology so far?
Len: The US Army STE program, on which we're the prime. We've already done an integration of the Legion framework with VBS4 from Bohemia, we're in the middle of integrating it with ONESAF through our partners at Leidos, and we're also in the middle of integrating it with one of the training management tools called Exonaut from a partner called 4C. There are a couple of other ones out there that we're not ready to announce yet but we're working with several different vendors to integrate with the Legion protocol.
Dan: I think there are several technologies that are similar to this - or at least, let's just say, there's a lot of people working at the problem of getting to really large-scale simulations. What's unique and innovative about our approach?
Len: One is just the focus on the modeling and simulation problem. The simulation problem has a few unique challenges, doing things like ownership transfer - transferring the ownership of the simulation or the authorization to publish an entity from one simulation to another, interest management based on whole-earth geography, the ability to have object models that are extensible so that people can simulate whatever it is they're simulating - whatever domain they want to use whether it's first-person shooters, ground vehicles, aircraft, space simulations or medical simulations. And like I said earlier, thinking of the problem holistically is sort of a big difference. There are lots of people thinking about the problem of communication, but by taking on the problem of how to access data efficiently and quickly within a simulation and tie that together with the idea of doing distributed simulation so you can publish, subscribe, and exchange information about those things - by designing those two halves together, we get a lot of advantages that you can't really get any other way. The main advantage is that you don't have to spend a lot of CPU time copying data from your internal representation to the network representation, which can become a real bottleneck when you're trying to do millions of entities.
Dan: Is this going to be viable in the market? Are our customers going to be able to adopt it?
Len: I think so. That's always been one of our hallmarks as a company - try to do things that are revolutionary in an evolutionary way. To try to make things as compatible as possible. So we're trying to stand on and leverage all the good work that's already been done over the past 25 years on DIS, HLA, RPR FOM, and some other elements worked on by SISO and other organizations, so we're reusing a lot of the same enumerations and a lot of the same data types, the same coordinate systems, the same concepts like dead reckoning. But we're trying to exchange that data in a more efficient, scalable, and more cloud-friendly way. So as far as viability though, the fact that we've proven things out and doing successful demonstrations of this in the past 6-12 months on the US Army STE program helps to prove that viability. And again, as far as integration with existing systems, I mentioned a few of them earlier, we specifically designed the legion framework much like HLA and other infrastructures to be independent of what the application is. So you don't have to be using simulation engines or visualization systems from MAK or any particular vendor to play with this architecture.
Dan: Is this the kind of thing that goes to SISO and becomes a standard at some point?
Len: That would be our intent. We're still maturing the technology a little bit but our history generally is, when we build something like this, we try to make it an open standard. That's part of our ethos as a company. We always want things to be flexible, open and we try to prototype something to get to the point where it can become the basis for a viable standard so when we bring it to a community like SISO, we already have something that's proven and working, and the community can give us feedback and mature that to the level where it can become a true open standard.
Dan: I think we've proven that in the market over the last 30 years.