It is mid-2010 and ATCA as a technology is about to reach its next milestone by becoming the only standards-based hardware architecture with 40GbE support with an all-IP platform architecture.

Since IEEE has standardized IEEE 802ap in May 2007 and IEEE 802.3BA in June 2009, it is becoming a natural fit for ATCA. PICMG is currently in the process of creating the PICMG specification under PICMG 3.1 to implement these standards.

The xTCA ecosystem has gone along with this direction and we can see products with this new implementation becoming available in the market during this year.

However, there have been several open questions from customers and integrators when planning ahead with 40GbE components and systems.

At the recent ATCA Virtual Summit, vendors and customers outlined how 40GbE ATCA-based platforms can be used to create dedicated Network Elements for the most promising network architectures like LTE – but also how ATCA as a technology is moving toward the military and aerospace marketplace.

The new multicore processing bladed architecture is increasing the demand for hardware platforms with support beyond 10GbE. With that in mind, the ATCA infrastructure needs to be able to support this – and with high memory density and multiple processors on an ATCA form factor, it is pushing the power budget and, with that, the required thermal profile for these platforms.

In the early days of ATCA, 200 watts was the overall guideline for ATCA-based components and CP-TA has closed several gaps to ensure that there are generic guidelines regarding interoperability for vendors and customers. As a result, the Interoperability Compliance Document (ICD) and Test Procedure Manual (TPM) are clearly defining the test methodologies for these platforms and a result of multi-vendor cooperation to ensure products will work together without major challenges to integrate them. At the same time, test tools became available to perform these tests during the design and integration stage; the most common guideline is the CP-TA B.4 recommendation in the current TPM.

As next generation ATCA blades are becoming available with higher memory densities, higher compute performance and multiple options for the fabric interface, integration of all of this will need to get to the next level of interoperability guidance. The first challenge is the power and platform profile to host blades with a power envelope beyond 200W and 300W per blade. As the chassis and switch blades are considered to be “infrastructure” and node blades are “revenue-generating” components in the platform, a long-term approach needs to be considered when developing new network elements as node blades are typically exchanged more frequently than the “infrastructure” portion of the platform.

Backward compatibility needs to be supported by the blades and platform and is becoming more crucial in 10GbE-KR-based platforms.

On top of that, new applications intended to use 40GbE standards-based platforms require new switching features and will potentially be supporting an all-IP infrastructure. Sync-E and IEEE1588v2 are covering new ground in ATCA to meet this requirement.  Data transport and synchronization will become more complex in 40GbE ATCA and interoperability guidelines and test tools need to be created around this new platform. The member companies of CP-TA are working together to help make all of this happen.