AdvancedMCs cowboy up
AdvancedMCs have evolved to take on tough chores such as packet inspection, load balancing, and content-aware applications.
Some rodeo feats are harder to take on than others. That’s when it’s time to tell riders to “cowboy up,” a rallying cry that has also made its way into major league baseball. But baseball is a game, and rodeo an entertainment. Telecom Equipment Manufacturers (TEMs), on the other hand, face real-world challenges.
And it can be argued that to meet those challenges AdvancedMC modules have cowboyed up, bringing a modular approach to systems design in a flexible mezzanine form factor. TEMs have gained an open-standard architecture while being able to retain their proprietary base architecture. And the telecom guys aren’t the only ones who must take on cost and resource challenges. Designers in everything from enterprise to commercial to mil-aero are realizing AdvancedMC versatility can go beyond AdvancedTCA-based telecommunications. AdvancedMCs have been sought out for their “can do” processing approach and flexibility to be used as a main controller, data server, traffic processor, and media processor.
Economies of scale
Different usage models for AdvancedMC modules have broadened their use and economies of scale, enabling equipment manufacturers to free up valuable AdvancedTCA slots by combining general-purpose processing with packet processing functions using AdvancedMC slots with PCI Express, GbE, and even 10 GbE on a single AdvancedTCA blade. Plus, manufacturers have found the scalability and flexibility of AdvancedMC modules indispensable to help them manage system upgrades while protecting their initial investment. Standards-based, AdvancedMCs can be reused in either AdvancedTCA or MicroTCA, broadening use and resulting in a larger return on investment for AdvancedMC-based designs.
Which apps will be lassoed next?
The first and most traditional way of using AdvancedMC modules is with AdvancedTCA carrier boards or AdvancedMC slots in a general-purpose processor CPU or switch blade. In this type of application, the AdvancedMC provides hot-swap and redundant system management functions in an AdvancedTCA-based system or makes co-processing on a single AdvancedTCA node or switch blade possible.
A second application trend is in redundant and nonredundant MicroTCA platforms where the form factor can vary between single- or double-wide AdvancedMCs and MicroTCA platforms. These trends enable telecom- and non-telecom-related applications to establish a full-scale PCI Express, GbE, 10 GbE, or Serial RapidIO infrastructure.
A third application trend employs AdvancedMC modules in low-cost backplanes and enclosures. This trend corrals the basic infrastructure to support a pay only for what you need approach. This type of integration is typically seen in non-telecom, nonredundant applications such as test and measurement or military applications. The advantage here is that the cost of the overall platform can be reduced to a level that is still standard AdvancedMC, yet the infrastructure in the platform is used only as needed. It can also reduce the overall platform footprint. Typical implementation for this type of platform is in rugged applications with a small number of interconnected AdvancedMCs. Such an implementation could be, for example, a MicroTCA platform in which AdvancedMCs are used.
Reining in costs
The platform would consist primarily of single-wide or double-wide AdvancedMC modules as the system component building blocks. PICMG’s MicroTCA system specification scales from 4 nonredundant to up to 12 AdvancedMC modules with redundancy when configured with the MicroTCA backplane, incorporating significant flexibility in a smaller footprint. This feature is crucial for MicroTCA-based designs. Furthermore, the low-cost AdvancedMC system implementation with direct interconnectivity between AdvancedMC modules via the backplane, as with the Kontron OM5080, for example (Figure 1) allows developers to narrow costs to only what is needed. Two to four AdvancedMCs provide flexibility in cost and functionality.
Greenhorns contend with legacy architecture
A new generation of intelligent networks is being deployed in enterprise, data center, broadband, and access networks that have to contend with highly distributed, service-related, or legacy architecture applications. This is an atmosphere in which AdvancedMC modules have become an attractive alternative for tackling deep packet inspection, TCP/IP packet processing, load balancing, content-aware applications, security processing, compression/decompression, and other services. Deep packet inspection, or DPI, is essential for monitoring and modifying network activities. DPI informs administrators about typical activities such as traffic profiles, users, sources, or destination. Armed with this information, carriers and ISPs can introduce tiered services, Quality of Service (QoS), and enhanced network efficiencies that spur new revenue opportunities and more efficient operations.
AdvancedMCs partner with multicore processors to handle various packet and security processing functions, including forwarding, load balancing, traffic management, and IPSec in an AdvamcedTCA or MicroTCA environment. Newly available processor architecture allows the AdvancedMC design to be used for higher bandwidth applications with 10 GbE interconnectivity. 10 GbE is becoming a more and more standard platform design requirement, and consequently, putting the industry on the trail to 40 GbE and beyond.
AdvancedMC packet processing modules have quickly evolved and are optimized for layer 4 to layer 7 data and security processing. Designers of 3G/4G BTS, RNC, xGSN, LTE, and Media Gateways are using the modules. Combining next-generation packet processors with 12 64-bit MIPS cores that feature up to 14.4 billion MIPS64 instructions per second (14.4 GOPS), these new modules offer high-density, high-bandwidth serial I/O technology – a leap ahead of legacy parallel I/O technology.
Rustling up new designs
New design opportunities are possible when Kontron AM42xx AdvancedMCs (Figure 2) are integrated with a Kontron AT8904 AdvancedTCA 10 GbE switch blade. Combined, the two add functionality such as load balancing for Web server, SIP server, SSL offload, and content- or application-aware processing applications on a single AdvancedTCA switch blade, freeing up AdvancedTCA slots in the system with the shortest path of traffic. In addition, they can support QoS over Ethernet for IPTV, video on demand, and other broadband media services. When integrated with a general-purpose AdvancedTCA processor node, such as the Kontron AT8020 or AT8050, the Intelligent I/O module also increases functionality for LTE applications. In either case, Network Equipment Processor clients achieve the highest density platform design without the loss of a valuable AdvancedTCA slot.
The other advantage of using multicore-based architecture in an AdvancedMC form factor is that it can reduce the platform footprint significantly. And it enables the infrastructure to have dedicated cores for specific applications. Combined with virtualization, multicore-based architecture creates a unified processing entity for space- and cost-constrained applications.
Telecom equipment manufacturers are quickly realizing the need for higher-bandwidth solutions beyond 1 and 10 GbE. Next generation multicore processors featured on AdvancedMC modules significantly increase flexibility for next-generation xTCA platforms. Implementing multiple 10 GbE I/O capability with DDR3 memory bandwidth enables system designs for LTE and other core network elements. Multicore processing also produces a wealth of preintegrated platform designs. OEMs can build a broader range of dedicated security, load-balancing, and content- or application-aware processing applications.
As network traffic increases so does the need for application awareness, content inspection, and security processing, which increases the performance required for network I/O processing. In addition, hyper networks that serve enterprise, data center, access, and Service Providers need support for converged data, voice, and video. For this reason, hyper networking support and coprocessor packet filtering have become important features handled by AdvancedMC modules. Such handling has the advantage of integrating the system management functionality for the overall system. Also, the building blocks that affect system traffic can be cut away from the herd, that is, separated from the management building blocks. AdvancedMCs are hot swappable, so they can be replaced if needed to maintain full redundancy without any traffic impact to the system.
On the COTS open range
Profitability, of huge interest in today’s economy, is closely tied to time to market. The availability of COTS products helps speed time to market and is a cost-effective solution to free up engineering resources. It also ensures interoperability and satisfies market demand for open-standard equipment. COTS extends beyond hardware to software and middleware to offer enhanced reliability, longevity, and scalability for applications. To accelerate the development of new applications, designers, too, are seeing real value to working in a common, modular platform in order to maximize their resources to build or reuse multiple network or system elements with distinct differentiation.
Key industry organizations and working groups are also influential in ensuring that current and future standards meet system requirements and interoperability. One such organization is theRather than defining its own specifications and standards, the SCOPE Alliance relies on its members to define application profiles or best practices for applying AdvancedTCA, AdvancedMC, and MicroTCA technology as solutions for TEMs.
Growing the AdvancedMC ecosystem
From processors to operating systems to middleware and protocol stacks, a strong prevalidated ecosystem has been built to support AdvancedMCs. Today, that ecosystem also includes storage, tool support, reference designs, and other services. Ecosystem partners have delivered optimized multicore solutions across operating systems, development tools, applications, middleware, and hardware platforms. The ecosystem for AdvancedMC modules has strong and multiple vendor support that can help designers innovate new products and shorten product development time.
No one single vendor has an xTCA portfolio to create the best suitable xTCA platform for a specific application. Working with an ecosystem means “freedom of choice” for the best-of-breed components, but also brings interoperable solutions to market, including TDM, ATM, or DSP functions in combination with general compute processing or network processing.
Multicore processors have been an important addition for their low power and high processing performance that provide deep packet inspection for security and data content, co-processors for TCP. New MIPs-based processors from(Figure 3 shows a block diagram for the Cavium OCTEON processor) integrate 1 to 32 MIPS64 cores. The processor can include up to 75 Application Acceleration Engines with up to 400 Gbps of DDR3 memory bandwidth, and up to 100 Gbps of network connectivity while consuming only 2 to 60 W.
When integrated with a 10 GbE switch, packet processor modules can be used for load balancing for Web Server, SIP Server, SSL Offload, and content- or application-aware processing applications. Likewise, they can support QoS over Ethernet for IPTV, video on demand, and other broadband media services.
When integrated with a MicroTCA 2U platform, AdvancedMC modules can be configured, depending on the application, with other processor and storage AdvancedMC modules for various integrated security services.
Intelligent I/O module
An intelligent I/O module can employ two AdvancedMC modules: One has 4x 1 GbE ports to the front and software configurable interfaces to the fabric side (PCIe, 4x 1 GbE, or XAUI), and the other has 2x 10 GbE to the front and PCIe to the fabric side.
Using two AdvancedMC slots adds functionality:
n Host processor modules as main controllers or for load balancing.
n Storage modules as mass storage devices for processor AdvancedMC and Network Processing Units for UDP/NAT or load balancing processing in combination with a storage AdvancedMC module.
n Switch blade functionality can be increased by using NPU-based AdvancedMCs, supporting 10 GbE for NAT functionality directly on the switch blade.
n In combination with a general-purpose CPU-based AdvancedMC, all the intelligence for traffic management will reside on the switch with hot-swap capabilities.
n Integrated design increases the added value of the hub to encompass accompanying user applications, freeing up slots available for other valuable payload boards.
AdvancedMCs have proven they can cowboy up when it comes to handling high-performance, high availability processing chores. As versatile and flexible COTS building blocks for AdvancedTCA and MicroTCA-based systems, AdvancedMC modules can get the job done for a variety of necessary next-generation networking functions. Designers see AdvancedMCs as a cost-effective solution that can speed time to market for multiple applications on the same platform. And because AdvancedMCs have been so successful at these chores, a strong ecosystem to support them is thriving and innovating as well.
Sven Freudenfeld is responsible for North American Business Development for the Kontron AG line of AdvancedTCA, AdvancedMC, MicroTCA, and preintegrated OM solutions. Sven possesses more than 15 years of experience with voice, data, and wireless communications, having worked extensively with Nortel Networks in systems engineering, Sanmina-SCI in test engineering, and Deutsche Telekom in network engineering. Sven holds an electrical engineering degree from Germany. He is VP of and Chair of the CP-TA marketing workgroup focusing on the interoperability of COTS standard building blocks.