Breaking News
Industrial Control DesignLine Blog

The Problem with IIoT Design

View Comments: Newest First | Oldest First | Threaded View
User Rank
The Bigger Problem
jring614   1/9/2017 5:10:39 PM

Thanks for the excellent perspective re: IIoT.

FWIW, the layer approach is not the more efficient and secure architecture. However, it may make life easier for programmers.

A far more important problem is lurking in the move to the IIoT, notably, "an avalanche of bugs.' Used to be that a bug might interfere with one program. With the IIoT a bug can interfere with thousands. And, strangely enough, it doesn't take a bug to start the avalanche. Two programs, each bug free (proven correct) when caused to interoperate may cause a logic, arithmetic or semantic inconsistency to appear.

Further, the lemming reliance on testing ignores the fact that even the best tested software suite still contains latent bugs when deployed to unfortunate users. This, because as Prof. E.W. Dijkstra warned us 40 years ago, "Testing shows the presence, not the absence of bugs." 

To my knowledge the OMG is not focusing on Zero-fault software. The good news is that we can now a) diagnose user-encountered errors in seconds, b) clarify the situation for those who have to generate a fix, and c) confirm that the fix is adequate and causes no collateral problems --- all without test beds, test cases and regression testing. And all this can be done for less than 20% of what people are spending now. Further, we do not have to wait until a user encounters a latent fault. If we adopt a discipline of specifying the purpose of a program then its viable post-condition(s) can be specified then used to determine the weakest pre-condition thus any faults. All this can be done to hasten system integration and even development. Software budgets can be reduced by as much as 50% if we all pursue Zero-fault code. Perhaps more important system vulnerabilities exploitable by hackers can be reduced to near zerol

If you want to dig deeper, pls see    

Jonas Berge
User Rank
layered approach connectivity model for the IIoT
Jonas Berge   1/5/2017 9:33:18 AM
I personally agree with a layered approach connectivity model for the IIoT. An additional advantage is that it makes changes to any layer easier, so the system can be kept up to date as technologies and standards evolve. You can change one layer without changing the entire system. See this essay on how it is being done:

User Rank
Re: Correction
ewertz   1/4/2017 10:14:39 PM
No problem -- exaggeration is the new norm.  My only criticism is that you didn't exaggerate nearly enough, by contemporary standards

It would be better stated that it's been 300 years, and that the work that the OMG is does is "amazing, and the best ever" and "trust me, you wouldn't believe the work that they do!".

The existing hacked systems are destroying our livelihoods and our country -- thank you for helping us to make our systems great again.  Unfair!


User Rank
RichQ   1/3/2017 10:40:38 PM
The OMG was founded in 1989, so the 30+ year age originally stated was a slight exaggeration. The text will be corrected to reflect a 25+ year age.

Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed