ATCA Summit Report

Fractal Realms series. Backdrop of fractal elements, grids and symbols on the subject of education, science and technology





Untitled Document

INTRODUCTION

Advanced Telecom Computing Architecture (AdvancedTCA) has been touted as the architecture of the future for telecom. The PCI Industrial Computer Manufacturers Group (PICMG) first released the ATCA specification in Dec of 2002 as PICMG 3.0. Progressively, it incorporated Ethernet, PCI Express, Infiniband and others. Additionally, the MicroTCA design was introduced to address access point concerns which required smaller foot-prints and lower costs. (See MicroTCA Complements ATCA/AdvancedMC for Low- to Mid-Range Telecom Apps).

The product food chain includes silicon, hardware platforms (single-board  computers, packet processing blades, chassis and chassis management modules), middleware, operating systems and application software.

The architecture has been increasingly backed by indutry heavy weights such as Intel, Motorola, Freescale, HP, Lucent, Nokia, Samsung and the like.

ATCA has been gaining momentum and is expected to excel in the upcoming years in USA, Europe and, perhaps most interesting, in China.

THE ATCA SUMMIT

Anthony Abrose

Anthony Ambrose, General Manager, Modular Communications Platform Division, Intel Communications Infrastructure Group

In December, 2005, the ATCA Summit in San Jose, California featured four keynote speakers presented by Intel, Motorola, Freescale and Wind River. Leading off the keynote presentations was Intel’s Anthony Ambrose, General Manager of Modular Communications Platform Division. In the keynote, Intel promoted ATCA as a Modular Communication Architecture. The main goal for Intel, naturally, is to sell silicon, and together with its Intel Communication Alliance (made up mostly of hardware vendors) Intel is looking to strongly promote ATCA.

According to Intel, the key benefits of ATCA included shorter time-to-market for developers and reduced R&D costs. With ATCA standardized system components, developers do not need proprietary designs. Furthermore, these standard components would be readily available off-the-shelf.

Citing In-Stat, Intel predicted the worldwide revenue for ATCA would reach $7.9 billion in 2009. Motorola had a slightly lower prediction, estimating ATCA revenue would hit a range of $1.6 billion to $3.7 billion by 2007. However, both companies felt that the real market growth would start to ramp up in 2006.

Intel also suggested that the driving force for future ATCA growth would come from the IP Multimedia Subsystem (IMS) which was designed to support the telecom operators with both mobile and fixed multimedia services. Packet-switching and the existing circuit-switching phone systems would be also supported. Ambrose stressed that interoperability was very important to the success of ATCA. To facilitate that, Intel would provide a design guide book to show best practice approaches to design, mainly because the PICMG specification has left a lot of leeway in design choices. In this guide, backplane choices and software interactions with hardware will be described. In his conclusion, Ambrose’s call to action was for companies to work together to reduce the interoperability problem. By doing so, the market would grow much faster benefiting all companies involved.

Freescale, a competitor of Intel, is another silicon company supporting the growth of ATCA. While Intel is working hard on the ATCA server technology, Freescale is actively developing silicon focusing on a broad range of technologies including RapidIO, Gigabit Ethernet and PCI Express. Their overall goal is to target a wide variety of industry applications such as wireless infrastructure, enterprise routing/switching, telecom switching and media gateways.

To help stimulate ATCA market growth, Motorola has teamed up with Intel to demonstrate a proof of concept for the IP Multimedia Subsystem (IMS) network functions based on Motorola’s AdvancedTCA Avantellis communications server.

John Fryer

John Fryer, Director of Switched Platforms Operation Marketing, Motorola

The system would consist of an Open Application-Enabling Platform from Motorola, including AdvancedTCA blades with Dual Low Voltage Intel Xeon processors and Intel IXP2850 network processors, Carrier Grade Linux and Motorola’s NetPlane high service availability software, validated with IMS software from independent vendors.

In his keynote address, John Fryer of Motorola suggested that ATCA would benefit everyone involved including the medium and small-sized companies from the cost saving standard. In the past, a few large companies monopolized the market with proprietary designs. Motorola’s strategy was to be a total solution provider, manufacturing systems to meet the requirement of companies such as Nortel and Alcatel. When asked about what the market needed to excel the growth ATCA, Fryer commented that there was a need for continued education, both in terms of the benefits of ATCA as well as general know-how.

John Round

John Round, Fellow of Technical Staff and Manager of EMEA Platforms Team, Freescale Semiconductor

Kontron, a competitor of Motorola, provides application-ready ATCA systems ideal for IMS elements deployed for next generation 3G wireless and wireline networks. "Kontron’s strategy is to incorporate and validate all of the needed components including middleware, which is currently being offered by a choice of leading suppliers such as Enea, GoAhead and Solid," stated Peter Matz, Product Marketing Manager.

Component vendors such as Philips are also benefiting from the growth of the ATCA market. According to Phillip K. Luu, Technical Marketing Manager of Philips Semiconductors, “We are glad to see the high growth prediction of ATCA in 2006 as we manufacture ATCA components including I2C hot-swap buffers, bus repeaters, bus extenders, temperature sensors, I2C multiplexers and switches.”

On the software side, operating systems (OS) and middleware were discussed. By far, Linux was the leading operating system supported by Intel’s Modular Communication Architecture. Other ATCA OSs are provided by Wind River and Enea. Additionally, Enea also provids middleware. To expand their market, both Wind River’s and Enea’s products incorporate Linux in their product offering and support, with Wind River also supporting expanded Carrier Grade Linux (CGL).

THE FUTURE

John Fanelli

John Fanelli , VP of Product Planning and Management, Wind River

In telecommunications, down time for either voice or data is not tolerable. To address this issue, the Service Availability Forum was formed by a group of leading communications and computing companies. They worked together to define a specification to support uninterrupted service with high availability (99.999 percent uptime). Middleware managed all the critical tasks to keep the service highly available. The forum further promoted The Hardware Platform Interface (HPI) specification which separates the hardware from middleware and helps hardware vendors design more plug-compatible equipment. This means that any middleware designed to the HPI specification would be able to run across all the ATCA hardware platforms made by any hardware vendor.

ATCA is gaining momentum everywhere. PICMG USA, Europe, Japan and China are all promoting ATCA. Of particular note is China’s increasing demand for telecommunication services. Recently, China topped the US in mobile phone consumption. Its interest in ATCA naturally follows. The annual PICMG Tech Forum (held in Shanghai or Beijing), featuring ATCA the past two years has been well received. “We see the market heating up in 2006,” according to Liu Hui, deputy secretary general of PICMG PRC in China, "ATCA is a perfect solution for the hardware vendors in China. Not only TEMs like Huawei and ZTE take advantage of this communication platform, new comers like EmbedWay developed and tested ATCA switch cards."

The open-standard ATCA is a great idea but is not without its challenges. As mentioned above, interoperability to ensure systems made by various vendors work seamlessly is a key element to its success. For all the companies involved to work together to achieve the interoperability goal will take time, but the benefits will certainly outweigh the cost. Using standards reduces development costs, shortens the time-to-market, and makes training and support on products much easier. Ultimately this will lead to the growth hardware and software in volume.