April 2015
6
M
ICRO
TCA
Now is the time to change your VME
and CPCI computing platform to MTCA!
By Vollrath Dirksen, Heiko Körte and Terry Manus,
N.A.T.
For more than a decade VME and
CPCI users have been waiting to see,
which open computing standard
would prevail as the most promising
successor. Now is the time to change:
MTCA provides the necessary
technology building blocks to provide
a smooth migration path for your
existing application. With a little basic
planning, MTCA provides a straight-
forward migration path to address
your application requirements now
and for the future.
e recent obsolescence of a key semicon-
ductor device used on VMEbus products is
causing VME and CPCI users alike to rethink
their future product strategies. For more than
a decade these existing VME and CPCI users
have waited to see, which open computing
standard would prevail as the most promis-
ing successor. And now this recent announce-
ment by IDT
1
and IBM to end-of-life (EOL)
a key component for VME board designs is
a clear sign that semiconductor manufac-
turers will not provide on-going support for
products that do not meet their economic
guidelines. Of course being le without a
compatible replacement device will in turn
force board vendors to extend product lead-
times, increase minimum order quantities
(MOQ) and ultimately EOL their own board
products as well. So as OEMs and End Users
of these board products focus on the short-
term Band-Aid
2
to this EOL problem, the near
term solution will ultimately be de ned by the
computing architecture they choose to move
forward with. is article will attempt to put
together a framework to rationalize what an
OEM and End User should consider when
making a change in their computing platform.
In today’s technical world it would appear that
the only constant is change. Based on numer-
ous examples in life we know that when new
demands can no longer be satis ed with
an existing solution they become the hard
requirements that drive a new one. ere are
many academic theories and laws dealing with
this position, but the most famous is Moore’s
Law. Over time Moore’s Law has been proven
time and time again, and it is fair to state that
his basic assumptions and conclusions can
be transferred to other industries or markets
as well - the most obvious reason being that
these industries and markets are massively
impacted by the semiconductor industry as
well. is position can be observed in the
history of computing architectures which will
undoubtedly continue into the future.
Although these standards all continued to
provide additional years of success beyond
its follow on replacement, there must be a
good reason why another standard is created
although an existing standard still contin-
ues to be used. is position then poses two
distinct questions:
1) What makes a computing standard
successful?
2) What are the forces that drive a new
computing standard?
To properly address these questions requires
at least a short historical review. A er all, it is
the historical record showing the test of time
that ultimately de nes a successful computing
standard.
VME –
From this perspective, even a short his-
torical review must start with VME as it is the
most successful computing standard used by
the market thus far. Remarkably VME has pro-
vided more than 30 years of success. And the
reason for its success is based on the fact that it
has been able to provide multiple sources at the
chip, board and system level. It used state-of-
the-art technology at the time the standard was
developed and in turn was accepted by multi-
ple markets for various reasons including its
modular design, open system standard, mul-
tiprocessor support and robust environmental
capabilities. e architecture allowed users to
combine high performance boards such as pro-
cessors with low performance IO so a system
could easily scale in terms of size and capability.
CPCI and PCI –
Continuing with the historical
record and as a consequence of PCs becoming
a consumer product, industry started to look
for ways to leverage this technology into indus-
trial use cases. In fact VME provided su cient
proof that PCI was a suitable approach as PCI
solutions were already being used in the formof
PCI Mezzanine Cards (PMC). As a result of this
commercial in uence the open standard Com-
pactPCI (CPCI) was released in 1999. Since its
inception CPCI vendors have always looked at
CPCI as being the successor to VME - mainly
because CPCI was based on the commercial
and technological advantages of PCI while
