April 2015
8
M
ICRO
TCA
maintaining the open and modular system
architecture of VME. But as history has shown,
and even though the CPCI standard was devel-
oped to meet the market’s requirements, it did
not provide enough sway to grab the purchase
orders of large main street decision makers
that were still wanting to see market validation.
ey simply wanted to wait and make sure their
next generation product would be based on a
future proven and successful technology that
would survive the test of time as did VME.
What makes a computing standard
successful?
With the aforementioned computing stan-
dards in mind, the answer to the question
posed earlier - What makes a computing stan-
dard successful? - becomes clear. A standard
is successful in the respect of a broad adop-
tion by di erent markets only if and when it
provides:
An open speci cation that necessarily allows
anyone to have access to it.
Architectural control by an independent
organization as opposed to a single vendor.
A healthy ecosystem o ering a broad range
of needed infrastructure and processing
capability using devices such as processors,
IO, chassis, power supplies, etc.
Vertical market engagements that require
longevity.
A clear commitment of leading semiconduc-
tor companies to support these markets.
A standard that allows the exibility to adapt
to further changes in technology.
Market acceptance o ering solid monetary
support.
What are the forces that drive a new
computing standard?
Based on the historical review presented ear-
lier the forces that drive a new standard are
typically disruptive in their nature. In this
speci c scenario the disruption would either
be a technical or economic capability that was
not previously available. And although a suc-
cessful standard should allow for evolution-
ary changes to occur, there will be some point
in time where a disruption is inevitable. e
point being that these changes are ultimately
necessary to meet the application require-
ments for tomorrow. An example of this
might be moving from a parallel to serial data
paths. Other examples could be referenced by
Ethernet, PCIexpress, storage (SATA, SAS),
RapidIO and USB.
ATCA and AMCs –
In 2000 a few large cus-
tomers in the communications market worked
with board, chassis and system vendors to
de ne a new computing standard that would
meet the computing requirements for the next
decades.EstablishedasaPICMGstandardization
committee this group worked to develop a
speci cation that initially targeted the tele-
com market and then with a bigger scope to
replace an aged installed base of computing
platforms running on VME and other pro-
prietary form factors. e result of this work
was the development of an open standard
de ned as the Advanced Telecom Computing
Architecture (ATCA).
e target market ATCA was developed for
made it a natural architecture to consider
as the successor of existing VME industrial
applications. is is underlined by the mez-
zanine concept that accompanies ATCA. e
fact that PMC modules used with VME and
CPCI did not o er hot-plug and hot-swap
functionality and provided a limited data
interface made an alternative mezzanine
architecture necessary. With this being the
case a new mezzanine card architecture for
ATCA was developed – the Advance Mezza-
nine Card (AMC). AMCs use an enhanced
architecture based on PCIexpress, SATA/SAS
for storage, both 1/10 GbE, Serial RapidIO
(the successor of Parallel RapidIO), IPMI for
system management and diagnostics, and
user IO to not only address today’s require-
ments, but tomorrow’s as well.
Today ATCA is being used in various high-
end applications that include the core network,
semiconductor fabrication and military/aero-
space. e research and test markets have also
used ATCA for applications involving exper-
imentation as well. Generally speaking appli-
cations in the standard embedded computing
market nd ATCA as platform that exceeds
their requirements for cost, power consump-
tion and size.
MTCA and AMCs –
In 2006 the rich set of
possible interfaces and features of AMCs
originally derived for ATCA paved the way
for a smaller more cost e ective and scalable
ATCA architecture known as MTCA.
On its own, MTCA systems can scale from
a single board to a 9U rack-mount applica-
tion o ering high performance capability in a
small package size. e carrier functions and
all other system infrastructure components of
ATCA, for example Ethernet switch and sys-
tem management boards, are all minimized
as optional sub modules and serviced by one
Release dates of popular Computing Standards
Time to change to MTCA -
User Statements
DESY - Control of XFEL accelerator:
For complex, high performance and
large installations at DESY a scalable
and modular standard is mandatory.
Furthermore, redundancy of fans
and power supplies are required to
assure 24/7 operations. With MTCA.4
we save a lot of investment since
the modularity allows reusing hard-
ware, so ware, and rmware and the
integrated remote management is a
unique feature to maintain large dis-
tributed systems. e excellent ana-
log performance and the high speed
data transfers are further arguments to
use MicroTCA as a successor of VME.
(DESY, Kay Rehlich, Head of the XFEL
accelerator controls)
Lockheed Martin Space Systems:
3 major reasons to use MTCA for a
new Test-Equipment and Instrument
Veri cation Facility for High speed
data processing and Protocol support:
COST!
Native XAUI and SRIO protocol
communication without FPGA’s
Simple backplane compared to com-
plex VPX con gurable backplane
RUAG Space (RSA):
3 major reasons to use MTCA for a
new Test-Equipment and Instrument
Veri cation Facility for RSA GPS-Re-
ceiver for Precise Orbit Determination:
Matured Ecosystem
Scalable architecture
Legacy IO could be re-used
GDP Space system:
3 major reasons to use MTCA to create
a Telemetry Network Appliance:
Hot Swap modules
Redundancy for High Availability
Channel Modules (PCM, Video,
IRIG, T1, etc) and Link Modules
(ATM, Ethernet, PCM, OC3, OC12,
etc)
Varian Medical Systems:
3 major reasons to use MTCA in Test
& Measurement Equipment used for
producing sensitive components used
in medical imaging equipment:
PCIe support
Fast data paths available (SATA 2,
PCIe (Gen.3) and GbE)
Scalable architecture
