System Design for the 21st Century

Uploaded by niglobal on 07.08.2012


Ladies and gentlemen, please welcome Dr. James Truchard.

We're here to share the latest in system design.
So today I'd like to share the latest in what we're doing in
system design.
Last NI week we talked about the evolution of
instrumentation starting with the vacuum tube that really
launched a modern era of instrumentation.
That moved to the transistor, a new era with the transistor
and the integrated circuit.
And now we are in the era of software based design where we
use software and software ecosystems to build our
It is very clear that software is where it's at.
We see that in the PC and now in the smart mobile devices
where its software that enables these tremendous
ecosystems that we work in to share a lot of technology.
Really that gives us the leverage we need to make these
very, very complex systems work.
We started with virtual instrumentation back when we
introduced LabVIEW in 1986.
This was a combination of hardware and software.
It was a very revolutionary idea at that time where we
talked about how the software is to instrument that's real
focus on the software, and its role in integrating the whole
system giving the user the flexibility to define their
own instrumentation.
We added to that the digitizers and D to A
converters that would allow us to create the same kind of
functionality that traditional instruments had.
Spanning all the way from DC to now very high frequency.
RF, we talked last year about how we acquired technology
about the design space with AWR and high frequency
electronics development capability in PMI.
And these now have been integrated into our
technologies to create better RF solutions.
We also expanded the roll of this virtual instrumentation
to what we now call Graphical System Design.
We continue to focus on Test and Measurement, but we added
this new space of Embedded Design.
And Test and Measurement we--
our goal in 1985 was to do what the spreadsheet had done
for financial analysis, we wanted to do for scientists
and engineers.
I think we can see from all the demonstrations, all the
showroom floor demonstrations that we really have
achieved that goal.
And then we expand it because of the system
design nature of LabVIEW.
Matter of fact, its original name was Graphical Measurement
System. so the concept of Graphical Systems has been
with LabVIEW since the beginning.
We expand it to this new role of Embedded Design where we
said, we want to do for Embedded what
the PC did for desktop.
In other words, create this ecosystem.
So a software based ecosystem that you can integrate just an
incredible array of applications.
Customers could share applications, suppliers could
add in applications to create a very, very large ecosystem
for across both design and test.
With this we are able to create the most advanced
systems in both Test and Measurement and Embedded
Design and I sometimes say for the first time in history, we
can have advanced measurements and advanced control in the
same system.
Now we also have introduced and I gave a talk at the
International Test Conference the year before last that
introduced the role of the V in our view of design.
The V is very conventionally used in many areas like
automotive and aerospace where on the left side you talk
about system level design, and component level design, and
onto to the detailed design of the system.
And then you have a matching corresponding
test on the test side.
So this is a classical way of looking at the design process.
Now a lot has changed over the years.
If you look at a cell phone just a few years ago, it was a
whole lot simpler device.
But now we see cellphones have gotten a lot more complex.
And every day a new feature, cameras, accelerometers, all
these things that add it to the cell phone because it
makes it more useful.
And then there's a whole now the set of applications that
can be built with it.
And oh yes, it still has to make a phone call.
So this complexity is really putting tremendous demands on
the designers and the testers.
And so we need to have higher levels of abstraction and we
need to have a systematic way of integrating those
abstractions into our design and test process.
So we introduced the V, it expanded true to use of higher
level view of where we use the combination of Design and Test
to integrate a new kind of test we call
Cyber Physical Test.
And this has really taken from a term the National Science
Foundation uses and talking about computers used in real
world systems with real world I/O. But we can define systems
at this level and test the system at
the functional level.
Moving down we have hardware in the loop.
This is commonly used for things like testing ECUs in
automobile engines and things.
And then we introduce another concept of
protocol aware test.
This is working at a protocol level where we're
communicating with the device in the protocol it expects.
This simplifies the test, it's much faster than the
traditional bit vector testing.
Then of course we need to be able to dive down whenever
necessary for the bit level to do traditional high scale ATE.
So our goal in this implementation of V from
design to test is to make this a process where an IP that
used the design process can be moved to test process to do
testing at whatever level of abstraction is sufficient.
If necessary then we can dive down.
So this is a very important view of how we view the role
of Graphical System Design in Concurrent Design and Test.
So it's all about abstraction.
You need very, very elegant methods for creating
abstractions that still carry the lower level detail in a
way that if you need to look at those, you can.
So abstraction makes it possible for us to scale to
these much more complex system like modern smart mobile
devices where we continue to deal with the complexity while
meeting in a time to market at the same time.

Now if you look at the history of LabVIEW and our approach to
system design we have to look at the early history of
electronic design itself.
Berkeley introduced SPICE in 1973.
This really kicked off modern electronic design.
And then later their technology that they had
developed went into the forming of
Synopsys and Cadence.
And then in 1986, National Instruments introduced
LabView, the first version, at the same time professor Edward
Lee got his Ph.D. in Synchronous Data Flow.
So for over a quarter centuries both Berkeley and
National Instruments have been focused on data flow as a way
to go to next level up the curve the system level design.
We've had--
LabVIEW is a practical commercial product and we've
since evolved it to include many, many additional features
up to supporting MultiCore and FPGAs.
And we've also recently been over the last years, been
working closely with Berkeley so we understand their
technology as well as what we're doing so we can have the
best capability for this system level design.
We also have introduced in the early release version of
Synchronous Data Flow that allows for high speed
computation on FPGAs.
So this is an exciting area to extend what we call Models of
Computation for the next era of
instrumentation and system design.
We also have the advantage of having integrated
hardware and software.
So we've been able to add Time at all levels from very high
speed back plane timing, to GPS timing,
to timing over cables.
We're working with CERN on White Rabbit, to software Time
in the FPGAs, to timing loops in LabVIEW.
A tremendous innovation for LabVIEW to really bring it
into the real world with Time included.
And then the software constructs like FIFOs and so
forth to complete ability to work with Time in a very
nimble way.
We also look to the ideas about system design.
And this is an image from Sangiovanni-Vincentelli
putting it on one view of how we want to visualize graphical
system design.
What we have is the exploration space at the top
where any scientist or engineer can imagine a problem
they want to solve and have a way of mapping it through
LabVIEW in our case, or a true APIs onto a
choice of hardware platforms.
So that's exactly what we tried to do with our platform
based approach with LabVIEW being able to solve such an
incredible range of problems from very large scale physics
problems to the latest research.
So basically a platform based approach combining a choice of
applications with the technology for signal
processing math and so forth, mapped onto a choice of
hardware, PXI, CompactRIO, Plug-In Boards, USB devices.
All making it possible to quickly implement systems with
this graphical system design approach.
Looking at our investment in the area FPGA, we first showed
LabVIEW FPGA in 1997 in NI week.
And we've been refining it, adding new products
all along the line.
Our Plug-In Boards first, RIO boards we call them.
Then cRIO, CompactRIO, our IF-RIO, then the FlexRIO.
And we have a very, very exciting announcement you will
hear, not from me, but the next speakers where we are
really raising the bar.
And our vision for instrumentation is being taken
to another level.
All the pieces of a puzzle coming together to create an
incredible capability for next generation instrumentation.
Here's an example of one of the, as we call them Edisons,
who are pioneering next generation technology.
In this case, Dr. Fettweis at Dresden
University on 5G Wireless.
And this is our vision.
We want to be in front of the curve, designing and defining
the next generations of standards and technology with
this instrumentation.
We in the past have talked of Edison.
And we want to empower the Edisons of the world to
innovate and discover in a much faster pace.

In this process we want to follow the ideas of scientific
principles first enunciated by Galileo where we have an
experiment, we make a hypothesis, a mechanism to
explain that observation in a real world.
And Edison was an incredible experimentalist doing
literally 10s of thousands of experiments on light bulbs,
batteries, and on and on and on.
And very, very successful with his experimental approach.
Over time folks have used this but also added these
approaches of math theory to go with explaining why it
works, to refine the design.
And then final a demonstration with real world measurements
to see if it works beyond paper.
Now we would also add beyond computer simulation as well to
this statement.
So if you look at one area that National Instruments has
been involved since 1989, it is cold fusion.
This is a very controversial subject and in 1989, we were
looking to working with some of these early Edison's and we
actually offered everybody who wanted to prove that cold
fusion existed a copy of LabVIEW and everybody who
wanted to prove it didn't exist a copy of LabVIEW.
And as far as I know, only the ones trying to prove it
actually took us up on the offer.
So we actually did an app note, how to do cold fusion
with LabVIEW.
So we've been there all along in this very
controversial area.
Most recently a lot has happened since then.
A guy did, Grimshaw did his master's thesis at the LBJ
School of Public Policy, you view that as fairly neutral,
saying he had found 184 examples of positive
experiments in this area.
And he recommended that public policy support
research on this area.
We also have been working with LabVIEW over the years and
many of these Edisons in their basements, in their closets,
in their out back laboratories, in the back of
their yard to continue this research.
We actually have a demonstration of Dr. Cellani's
work from the Italian Nuclear Institute on
the showroom floor.
And I'm actually recommending a different name.
I think that name got a little tarnished.
Maybe we should call that Quantum Reactor.
That's what the name often used in
science fiction movies.
So on the show room floor we have a demo.
And we also have a panel of some of the select
researcher's that have been working this area for the last
20 years, including Professor Hagelstein from MIT who has
worked on some 282 theories about how it works.
His latest one he thinks is his best to one,
so check that out.
So in the experimental process in this case, it's kind of
been stuck at this first step of observation.
So our goal in life is to make sure these Edison's can move
on to the process, get the measurements they need, have
the capability they need, then our graphical system design to
solve these really, really important
problems as we go forward.
So it is about Graphical System Design where the
inventor can choose whatever subject.
We say we don't judge, we measure.
So we basically want to make so the user has a lot of
choice in what they do in their Graphical System Design
for both design and test.
Doing it ever faster, more ably as we see from the many,
many examples we see as we go.
It's still about virtual instrumentation.
VI is still alive and well.
It's the fundamental building blocks of Graphical System
Design, and the instrumentation is in the
software era with virtual instrumentation where the
software is the instrument.
Thank you for your time, thank you for joining us here at--
--NI week and we are looking forward to an incredible