Visit the Micro-Robotics web site
Click on the advert above to visit the company web site

Product category: CompactPCI Boards and Assemblies
News Release from: Adlink Technology | Subject: ATCA products
Edited by the Electronicstalk Editorial Team on 07 February 2005

The future of AdvancedTCA

Jeff Munch, Chair of the PCI Industrial Computer Manufacturers Group AdvancedTCA Subcommittee, examines whether ATCA has delivered on its promises.

The AdvancedTCA specification was ratified by PCI Industrial Computer Manufacturers Group (PICMG) in December 2002 At that time, ATCA promised to revolutionise the telecomms infrastructure by providing high performance, cost effective, standards based platforms that could be provided by multiple vendors

The goal of ATCA was to eliminate the proprietary platforms that existed, reduce development time and reduce development costs.

Has ATCA delivered on its promise? What are the lessons learned over the last two years? And what does the future hold for ATCA?.

Since the introduction of ATCA, we have seen several carriers announce support for ATCA in next generation platforms.

Telecommunications equipment manufacturers (TEMs) are now seeing request for quotes from carriers that specify ATCA.

In one recent quote in the area of wireless infrastructure, four of the five responses were based on ATCA and the fifth was based on PICMG 2.16 CompactPCI.

The company that quoted 2.16 was asked to explain why they did not offer an ATCA based solution.

This is proof that carriers are serious in their use of ATCA for next generation platforms.

Although ATCA was architected with telecommunications requirements in mind, we have started to see some movement in non-telecommunications applications for ATCA.

Hewlett Packard recently announced a data centre product based on ATCA.

We expect to see other market segments adopt ATCA as well.

AdvancedTCA was part of the modular compute platform (MCP) concept.

MCP defined the layers of hardware and software in typical telecommunications platforms, and most importantly the interface between the layers.

These layers were called building blocks.

By modularising and standardising the platform, multiple vendors are able to provide compatible building blocks for a given layer.

AdvancedTCA building blocks reside in the blade and ATCA chassis/switch layers.

The blade can be network processor, x86 processor, or general purpose processor based.

The blade, switch and chassis are the basic ingredients of an ATCA platform.

In ATCA platforms the switch takes on a very important role as all blade to blade communication pass through the switch.

The AdvancedTCA specification provides a significant amount of flexibility.

It is a specification designed by engineers and is best understood by engineers.

Blades can chose from multiple fabric interfaces including PICMG 3.1 Ethernet and Fibre Channel, PICMG 3.2 InfiniBand, PICMG 3.3 StarFabric, and 3.4 PCI Express.

Within each of the Fabric Interfaces, multiple options are allowed.

As an example PICMG 3.1 Ethernet/Fibre Channel specification contains nine different interface combinations such as: one Gigabit Ethernet; two Gigabit Ethernet; four Gigabit Ethernet; one Gigabit Ethernet and one Fibre Channel; one Gigabit Ethernet and two Fibre Channel; two Gigabit Ethernet and two Fibre Channel; one Fibre Channel; two Fibre Channel; and one 10 Gigabit Ethernet - these are just the PICMG 3.1 options.

Blades can also choose the type of topology used by the fabric interface such as none, dual star, dual-dual star and full mesh.

The same levels of options are available for Switch boards as well.

This can make integration of an ATCA platform a challenge.

It requires that the integrator understand the options allowed and the expected compatibility.

As an example, a PICMG 3.1 blade that implements one Gigabit Ethernet will not work with a PICMG 3.1 switch that implements one Fibre Channel interface.

In this example, both components are PICMG 3.1 compliant but not compatible.

All this flexibility can combine into a challenge for the users of AdvancedTCA platforms.

Is it a good use of the end user's time to have to understand all the flexibility inherent in AdvancedTCA and then select compatible products? This is not where users of ATCA add value.

This is a detractor of value.

The burden of understanding and configuring AdvancedTCA systems lies with the manufacturer and integrators, not the users.

We are now seeing traditional board level suppliers playing a roll as integrator as well.

After all, the board and chassis vendor are best suited to understand and solve the compatibility issues.

As vendors of ATCA products move towards integration we are starting to see a new concept occur.

Instead of users purchasing and integrating blade level building blocks we see the emergence of application ready platforms.

These platforms are built on standards based off the shelf building blocks.

From a user's standpoint, the platform purchased is ready for application specific software - no need to worry about blade/backplane/switch compatibility.

There is one vendor to handle support calls.

One of the major opportunities in the future will be the integration of ATCA building blocks into application enabled platforms.

AdvancedTCA is well on its way to fulfilling its promise.

Deployment of ATCA in telecommunications applications is happening now and announcements of the penetration of ATCA into the data centre are now occurring.

It is expected that military applications might also benefit from ATCA.

We see technology improvements that will result in a 10x increase in the data throughput of the fabric interface.

The evolution of ATCA will keep it in step with the market requirements.

With all the technology it is important that users of ATCA focus on their value add and leaving the integration of ATCA platforms to companies that specialise in the manufacturing, customisation, and integration of ATCA platforms.

The ability to stay focused will be a key success factor for companies of the not to distant future.

Jeff Munch is Chair of AdvancedTCA subcommittee, PICMG, and also CTO of Adlink Technology.

He has more than 20 years of experience in hardware design, software development and engineering resource management.

Before joining Adlink, Munch spent five years at Motorola Computer Group as Director of Engineering, and previously, 9 years as Vice President of Engineering at Pro-Log. Request a free brochure from Adlink Technology ...

Adlink Technology: contact details and other news
Email this article to a colleague
Register for the free Electronicstalk email newsletter
Electronicstalk Home Page

Search the Pro-Talk network of sites

Visit the Micro-Robotics web site