Talk:High Level Architecture

What is the area where this definition applies? For instance, would it be correct to start the article "In computing, ..." ? Notinasnaid 10:51, 3 Nov 2004 (UTC)

Interoperability Message Being Changed to Reflect a Different View
The original section titled “Interoperability Limitation” was almost completely rewritten to reflect a different view. I put it back as it was, but changed it to reflect only the raw facts. Too many people have found out the hard way that the HLA has no network standard and any given RTI implantation will not interoperate with any other, even open source ones. There is an expectation that these RTIs will in fact connect with each other by the Distributed Modeling and Simulation (DM&S) public, but alas, it’s not true and my original purpose for this section was to make that point clear. It is my opinion, watering down this meaning is wrong technical and ethically.

Closing architectures rarely benefit anyone in the long run. Researchers outside of DM&S have shown this to be true. The idea that the HLA architecture was closed in order to provide for technical innovation seems doubtful since anyone would be free to innovate regardless of the open or closed nature of the network transport. It’s a business model just like any other proprietary network transport that traded openness for market share. In the relatively small DM&S community, this seems to have been missed. — Preceding unsigned comment added by LVCTaylor (talk • contribs) 23:31, 1 January 2013 (UTC)

Once again, putting the message back in after deletion
The last editor deleted the controversial section altogether and uses this reasoning: ''Section deleted since it was irrelevant. HLA standard is open on the API level and is not intended to be locked into any specific way of physical data distribution The authors point, Section deleted since it was irrelevant'' could not be further from reality, as this point is the single greatest disadvantage to using the HLA as it is the expectation. The second part, locked into any specific way of physical data distribution makes little sense, because any RTI author would still be free to do as they wished regardless if a wire standard existed or not. Similar circumstances existed in other industries and these excuses for closed architectures were used there as well until well thought out and mature standards emerged. Case in point is the Real Time Protocol (RTP). Attempts to hide these facts only serve to highlight this issue. Other interconnect mechanisms also have “disadvantage” sections on their wiki's. This section is headed to dispute resolution. — Preceding unsigned comment added by LVCTaylor (talk • contribs) 08:22, 15 January 2013 (UTC)

Concept of interoperability
The concept of interoperability does not exclude the use of an API standard to support interoperability. To state that a wire protocol is necessary in order to achieve interoperability is misguiding. — Preceding unsigned comment added by SpaceRay (talk • contribs) 20:01, 15 January 2013 (UTC)

Disagree
API’s are everywhere and thousands are standardized and serve only to describe one side of the data flow in this case. In the case of HLA, the standardized API connects applications to the proprietary RTI interconnect mechanism which requires all to use the same proprietary RTI, a fine example of the definition of non-interoperability among participants. The author that is trying to tell us that HLA is somehow interoperable is just plain wrong. It might be deemed data interoperable, but even that’s a stretch because the data has to flow through a proprietary network to get to the other side. If everyone on the planet used the same RTI and it was the only one, a wishful though indeed for some, we might be able to address HLA this way, but for now, we cannot and should not, else we would be misguiding indeed. The long section discussing the API standard versus a protocol standard is an opinion and is not here in support of the definition of HLA. It should be deleted or moved to the talk section. — Preceding unsigned comment added by Realtme101 (talk • contribs) 05:59, 16 January 2013 (UTC)

Protocol vs API standard
Whether an API standard is an advantage or a disadvantage has to judged by users of the standard and not by any individuals. The description that HLA is an API standard and does not specify any data exchange protocols is enough.

Moved into talk section: HLA defines an API standard, which means that the standard is based on achieving interoperability between simulation models (federates) through the use of an open API. The standard does not specify how data and services is distributed in the RTI or what physical mean of data distribution that is used. This allows the underlying infrastructure to be improved or replaced without affecting the API or the simulation application. HLA compliant implementations is therefore not locked to a certain networking standard and can use other methods of data distribution such as reflective memory buses etc. However, today most, if not all, RTI implementations use standard IP network. The HLA API establish an API between the simulation application and the RTI and the specific method of data distribution may be proprietary to a specific RTI vendor. Today's modern high performance RTI implementations use a complex distributed algorithm that can be optimized to limit the amount of data that is sent when executing very large simulation in terms of object and interactions and/or number of participating simulation applications or any other factor that is important to the RTI implementor. Local RTI components from different sources implement these algorithms in different ways in order to find optimized solutions and can therefore not normally be mixed with each other in the same federation and exchange data with each other. This is not normally a problem since a single RTI implementation can be chosen to support a simulation event and local RTI components can be replaced without the need to recompile or link the participating simulation applications. In the HLA standard the possibility to optimize data exchange performance and avoid locking simulation applications to any physical mean of data distribution have been selected over the use of a protocol standard where data is exchanged between simulations applications through an agreed protocol. (Editorial opinion, delete, change or move to talk page) — Preceding unsigned comment added by SpaceRay (talk • contribs) 08:27, 16 January 2013 (UTC)

SpaceRay (talk) 20:09, 15 January 2013 (UTC)SpaceRay

Factual deletions are not ethical
To say that interoperability is the result of proprietary communications is true only in the literal sense. I see a desperate attempt to justify this idea that closed architectures are somehow characterized as interoperable. Closed architectures are by definition non-interoperable. I am not purporting that the HLA has no value, only that is not be sold to the lay public as an interoperable mechanism. I have personally known several instances where national labs used the HLA, and were surprised when that lab could not interconnect to others using the HLA because the other side used a different HLA implementation. By the way, the lay public also does not realize that the cost of a typical commercial HLA RTI middleware package is around $1,500 US. A lab with twenty workstations would have to fork over $30,000 for the privilege, only to find out that they cannot interoperate with other RTI implementations. This is the single message you continue to delete and follow with less than factual stuff that is at best an opinion. My section was factual. It is a fact, and you keep deleting it. — Preceding unsigned comment added by LVCTaylor (talk • contribs) 22:41, 15 January 2013 (UTC)

Personal crusades are not ethical
The definition of interoperability clearly states something else here on Wikipedia than what is claimed by LVCTaylor in the above section. LVCTaylors attempt to mislead readers with personal opinions is depressing. — Preceding unsigned comment added by SpaceRay (talk • contribs) 07:53, 16 January 2013 (UTC)

Unfortunate
SpaceRay’s deleting message after message of anything that mentioned or even hinted that the HLA was not an interoperable mechanism except of course, between different RTI’s of the same type and version shows a clear intent. I expected a push back from that sector of the market, but didn’t expect it to that degree. I also didn’t expect the redefinition of the word interoperability. RTIs are not interoperable from different vendors. That is the key message and was the intent of this effort to show that message is actively being suppressed. We will all benefit from opening architectures and closing them is not only sticking it to the taxpayers, it is done so with selfish intent to corner markets despite the plee that this was done to help us. Maybe someone else will pick this ujp now. — Preceding unsigned comment added by 70.199.197.230 (talk) 08:30, 19 January 2013 (UTC)

RTI to RTI communication
A factual description that RTI to RTI communication normally isn't possible without the use of a HLA to HLA (federation to federation) bridge should be included. Whether this is an advantage or disadvantage must be judged by the reader. Calling this "Lack of interoperability" is misleading. The interoperability between federates (i.e. connected simulation applications) works perfectly well and according to the definition of interoperability.

This description is better placed under the section Run-Time Infrastructure and is already there:

"Due to this, interoperability between RTI products and often, RTI versions, should not be assumed unless the vendor specifies interoperability with other products or versions."

— Preceding unsigned comment added by SpaceRay (talk • contribs) 08:39, 16 January 2013 (UTC)

What is going on here?
The word interoperability is being abused here. If RTI software, whatever that is, requires a bridge to interoperate with another RTI then the RTIs are not interoperable, simple as that. There is nothing misleading about it, because it is the simple truth. — Preceding unsigned comment added by 205.168.20.2 (talk) 04:31, 21 January 2013 (UTC)

Realtime101
What is going on here is a single individual, SpaceRay, is on what is claimed that others are on, a personal crusade, except SpaceRay’s is to delete any information that might make the reader aware that the HLA is not open or interoperable. While it’s true that the HLA standards don’t provide a RTI interconnect standard, most are not aware the RTIs are proprietary. Many have expectations that the HLA is interoperable across the vendors of the RTIs and found out too late and don’t want to look like they are less informed. The fact is, this fact, is hardly mentioned anywhere, even on web sites of HLA RTI vendors. Of course it is known by HLA experts, as this is their job to know. RTIs are proprietary by design and provide no standardized mean to connect one HLA application to any other outside of that HLA vendor’s implementation of which there are many. This is a fact and hardly irrelevant as SpaceRay would have us to believe.

HLA is not used in any other industry that I can find, and for good reason. It is closed. No one wants to buy expensive middleware simply to connect real time data to other applications on other computers, although it seems that this industry does. SpaceRay also tells us that this is a necessary condition so that these RTI vendors can give us better and faster performance, but once again, fails to see that non interoperable RTIs are interoperability killers, requiring a bridge to connect them. Not only is the bridge custom and extremely expensive, but only connects two dissimilar RTIs, and when a new one is introduced, the bridge must be modified to include yet another vendors RTI and more custom expensive code.

HLA is a closed architecture. I cannot see how it could be described in any other way while being completely honest. I have read and understood all of SpaceRays arguments to the contrary, but find them lacking of any meaningful logic and certainly vision when these are analyzed against other architectures and systems. This is not the first time that a vendor has attempted to build a market by closing a networked architecture. (Empire Building)  Only a few have made it work, and the public suffered as a result.

Again, this is a closed architecture and closed architectures are not in vogue now anyway. Just try to imagine for a moment how non interoperable the planet would be if this idea of proprietary middleware was allowed to foster and survive in other industries. This hyper conservative thinking continues to create closed systems that serve only the middleware vendors and will never promote simulation interoperability as we now begin to build bridges to make thee non interoperable things work together out of necessity. — Preceding unsigned comment added by Realtme101 (talk • contribs) 06:45, 26 January 2013 (UTC)

Lack of understanding
I am surprised that individuals who clearly don't understand the principle of HLA take it upon themselves to explain to others what it is.

If somebody called IP, the internet protocol, as non-interoperable or having "lack of interoperability" because it doesn't state what physical media it should be transported on and that it requires proprietary hardware to route between different physical media types I don't think anybody would agree.

The open HLA standard defines an API between the simulation model and the a run-time infrastructure that allows simulation applications to exchange data, syncronize time etc. without requiring any knowledge about how this is done behind the API. Instead of having the TP connector on the back of the computer, the connector has moved into the computer and sits closer to the simulation application and is called a software API. The API allows the simulation application developer to use all the HLA services without having to have any knowledge on how to distribute data and syncronize time between simulation applications. HLA allows simulation applications to avoid a technical lock-in to a certain data distribution mechanism and allows for the the physical network to be replaced with another mean of data distribution without having to modify the simulation application. The standard defines the API and allows for the RTI itself to be replaced with any other proprietary or open source RTI without the need of any relink or recompilation so I have a hard time understanding how it could represent a commercial lock-in situation. Everybody has the same opportunity develop an HLA compliant RTI, the interface specification is an open standard and every developer can choose to publish a wire protocol if it is desirable.

If you go to the "RTI implementation" section of this document there is a number of open source implementations that allows you to run any number of connected simulation applications in the same federation without having to invest a single dime and allows you to have full control over every data package that is sent (by analyzing the code). To claim that the HLA standard itself is expensive and/or proprietary is simply wrong since it can only apply to how the RTI is implemented and if some of the RTI implementations is to expensive or lack any highly desirable features such as a "wire-protocol" I guess the law of supply and demand will take care of that in due time. --SpaceRay (talk) 23:31, 26 January 2013 (UTC)

Here we go again
Lack of understanding? We fully understand HLA principles as HLA is not rocket science. It is a simple mechanism, and we get that. It is you, SpaceRay, that does not get it. This is a matter that requires openness to experience, nuanced thinking, and most of all avoidance of motivated reasoning. There is no doubt that SpaceRay believes what he is saying. However, despite his overwhelming confidence he does not understand what it is I am saying and others have said here. This is an age old argument not only for HLA, but one that has been bantered back and forth for years for all kinds of closed systems. Thankfully the planet didn’t embrace most of these where vendors vied for dominance. In the end, it was open and standardized mechanisms won out. There are still holdouts, such as Skype and a few others that closed off the UDP/IP layer. In fact, I am surprised that a few RTI vendors didn’t try that route and create a new transport layer. This is interesting, because even the RTI vendors utilize existing Ethernet stacks. I fully understand what is it that SpaceRay wants us to think. However, that thinking leaves out the most important points and focuses on only the parts that justify SpaceRay’s view. That view would be normal if this were 1986 when closed architectures were common, as network transports were not yet developed and standardized. I am not asking that SpaceRay understand this, only that he stop deleting what others have placed here, and begin the process of peer review and discussion before he simply deletes existing facts to suit his needs, especially when he is financially motivated to remain closed. I am not the only one watching this mess. Look at what others are saying. Look at the ratings for this page. Every month or so I meet a colleague that I respect, that has not yet worked with HLA, but has excellent networking systems knowledge, that is amazed to hear the HLA RTIs are proprietary and cannot interoperate. By the way, that word “interoperate” drives HLA folks nuts. This is because they only need to see a standardized API between the simulation application and their RTI to call the HLA apparatus as interoperable. In fact, HLA is not an interoperable mechanism. LVCTaylor (talk) 05:45, 29 January 2013 (UTC)

Really
If LVCTaylor really understands HLA and yet claim it to be non-interoperable how can the following be explained:


 * The US DOD through its M&S office (DMSO, today MSCO) release the first version, 1.3, of HLA as the preferred simulation interoperability standard
 * IEEE through SISO and using an open balloting process defines the HLA Interoperability Standard and releases it in 2000 and updates it in 2010
 * NATO nominates HLA as STANAG 4603; Modelling And Simulation Architecture Standards For Technical Interoperability
 * All distributed simulation user who together makes HLA the widest spread interoperability standard for distributed simulation
 * Other initiatives, such as the US DoD TENA, Test and Training Enabling Architecture, has also selected an API standard in favour for a wire protocol (without the influence of commercial companies)

Have all people behind these events and facts been bribed, hypnotized, fallen victim to the biggest scam in modern time or are they just plain stupid?

It seems to me that LVCTaylor just represents himself and his own crusade. SpaceRay (talk) 20:53, 29 January 2013 (UTC)

Easy Now
DM&S is a very small community. No, I do not think anyone was hypnotized. If it was a much larger community, such as VoIP, or industrial I/O the HLA would never have made it as those industries have open international standards that have wire protocols to transport the data. There are few if any closed systems there. I have never said HLA is bad. In fact, HLA has solved some critical problems and has been put to good use in many situations. In fact, it is being used on one of the largest training systems in the US that is currently under development. In that case, it is a perfect fit as the training devices have no need to interoperate outside of their local environment. In that case, multiple processors use the HLA to communicate, solving a problem that would have required a larger code effort. However, the project has no need to communicate outside of its WAN. If it did, then it would be a different story. I am merely trying to mention one negative aspect of the HLA that few seem to realize.LVCTaylor (talk) 22:10, 29 January 2013 (UTC)

That last edit is good enough, although it was my intent to reach the lay person, who does not know what these things mean exactly; API, wire protocol, C++, etc. Of course we expect that an RTI vendor and standards developers to be aware of these facts, but the lay person is not clear on this and has an expectation that HLA is an interoperable mechanism that can connect all HLA software applications together regardless of the RTI. This is what has been wrong with the message up to this point. The following statement is factually correct: “HLA applications are not interoperable across different vendors of HLA software.” This is easy to understand and designers having never used HLA will know what that means. I meet people all the time that have an expectation of interoperability regardless of the RTI vendor. The public deserves to know this. LVCTaylor (talk) 11:54, 2 February 2013 (UTC)

Definition of terms: standardized protocols support interoperable data exchange
Standards are designed to achieve interoperability. Indeed the very name of Simulation Interoperability Standards Organization (SISO) includes "Interoperability" in its title. There are three definitions given for interoperability on Wiktionary:
 * 1) The capability of a product or system to interact and function with others.
 * 2) (military) The capacity for a service, piece of equipment etc., to be operated by different forces or groups.
 * 3) (computing) The ability of software systems that may be running under different operating systems and hardware to exchange information through compliance with technical interoperability specifications, which typically define how different file formats and messaging protocols can work together.

Similarly in Wiktionary the ninth (computing) definition for protocol states "A set of formal rules describing how to transmit or exchange data, especially across a network."

Contrary definitions regarding RTI have caused a lot of misunderstanding. It is simply contradictory to claim that an RTI implementation supports a standardized protocol if interoperable data exchange is not provided.

Don Brutzman (talk) 20:24, 14 February 2013 (UTC)

Indeed
The concept of “Interoperability” has never been so mangled and redefined as it has been under banner of HLA. SISO has seen more than one attempt to change its name to remove or replace the word “Interoperability” to avoid such a prominent display of SISO’s purpose. I know of at least two excellent systems architects who left SISO partly for this very reason who know fully well what the HLA is and why it should never be used where the word “interoperable” is important.

One has to ask a simple question when designing DM&S applications: Do I expect anyone to interconnect to this application in the future that might not use the same RTI? If the answer is yes, then don’t force the other sides to PURCHASE the same middleware, or for that matter, don’t use it at all and find another means to interconnect the applications. LVCTaylor (talk) 22:44, 16 February 2013 (UTC)

Wrong name? Should be High Level Architecture
The standard says it: https://standards.ieee.org/findstds/standard/1516-2010.html — Preceding unsigned comment added by 195.37.177.69 (talk) 07:55, 8 August 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on High-level architecture. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20050113190226/http://www.scs.org/confernc/hsc/hsc02/hsc/papers/hsc017.pdf to http://www.scs.org/confernc/hsc/hsc02/hsc/papers/hsc017.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 03:26, 2 April 2017 (UTC)

Requested move 25 April 2019

 * The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion. 

The result of the move request was: Page moved. (closed by non-admin page mover) --  Dane talk  01:04, 4 May 2019 (UTC)

High-level architecture → High Level Architecture – The original name of the standard – Danmichaelo (talk) 17:19, 25 April 2019 (UTC)

Good move, both the name of the original and the most recent version of this standard is High Level Architecture. 61cc (talk) 13:48, 30 April 2019 (UTC)


 * The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.