MicroTCA.4: The next inflection point in open standards platforms

The perfect fit: MTCA.4 emerges as a bridge in the core/edge dilemma, and so much more.

8Andy Grove spoke of Strategic Inflection Points more than 10 years ago as significant changes that affect how businesses make decisions. Some may not be paying much attention to the new MicroTCA.4 specification, but the groundbreaking standard, in conjunction with new technical advances, makes this event one that should have everyone taking note. This truly is disruptive technology.

The new .4 standard was spearheaded by the physics community, but is extremely important for , military, commercial, and industrial applications. It introduces critical new features to the original MicroTCA.0 specification to make it truly five-nines available, serviceable, high performance, high bandwidth, and extremely flexible for numerous I/O configurations. Why is this important? There is a critical dilemma right now with standards-based platforms. There is a significant performance gap between and MicroTCA, and until now, there were few solutions.

has a large footprint and offers extreme computing power, making it ideal for core network applications, but is not very flexible or scalable. MicroTCA.0 is a small footprint, ideal for edge-based applications, but does not scale up well. This leaves many aggregation layer applications caught in the middle, demanding a mid-size compromise with plenty of performance, I/O flexibility, and high-speed bandwidth, in a cost-effective, smaller footprint with finer granularity and modularity that delivers five-nines availability and ruggedness.Concurrent with the ratification of the new .4 specification, extreme technical advancements have produced increased processing and networking throughput in much smaller packages. Think of Moore’s Law. Performance levels that were once only possible on ATCA cards just a few years ago are now possible on AdvancedMC cards. This includes powerful and high compute-density processing such as , hyperthreaded general purpose processors, and multicore packet processors. In 2002 when ATCA was being ratified, a typical Pentium 4 processor was running with 42 million transistors. Today, a quad-core i7 Xeon processor with 750 million transistors can run on a single AMC module.

Additionally, high bandwidth, managed 10 GbE/40 GbE fabrics are now available to support current demands on network traffic and the forecasted explosion over the next several years. 40 GbE will become more prevalent in the aggregation layer for mobile infrastructure, aggregation routers for IP routing and switching, IMS services, and service delivery solutions.

These technical advancements and the MTCA.4 standard converge together to offer the ideal solution to bridge the gap between ATCA and MTCA (Figure 1). This is a boon to a multitude of , telecommunications, and military applications. These systems bring the best of both worlds; in a small form factor they prove vastly more serviceable and scalable, with high levels of performance, network bandwidth, and availability.

Figure 1: MTCA.4 serves as a next-generation platform for bridging the “core/edge” dilemma.

ATCA and MTCA.4 are very complementary. Because of ATCA’s high level of density and compute power, ATCA solutions are the best fit for core network solutions with high-density computing and very large pipes of bandwidth. MTCA.4-based applications can sit in front of ATCA systems, performing specialized tasks. It supports all the same AMC modules available in the ecosystem today, and the platform management architecture is compatible.

What is MTCA.4?

The MTCA.4 specification was officially adopted in October 2011. The initiative was launched in 2009 when companies realized the immense advantage of enhancing the MicroTCA.0 specification by adding such capabilities as hot-swappable Rear Transition Modules (RTMs) for AMCs and precision timing. The addition of RTMs for all AdvancedMC slots and () slots increases a MicroTCA system’s serviceability and achieves five-nines availability. The RTM provides a new level of flexibility that adds more functionality in an AMC-RTM mated pair. It supports backwards compatibility so all AMCs currently in the ecosystem can be installed in MTCA.4 platforms. It has retention screws on its faceplate for higher levels of ruggedness. It also includes precision timing, which is critical for clocking, synchronization, and interlock signals. It is also important to note that the MTCA specification can support not only 10 GbE, but also 40 GbE on the backplane to all AMC slots.

Much more than physics

The physics community was instrumental in driving this new standard, and the majority of the enhancements they made will greatly benefit numerous industries. It can be used for evolving next-generation network applications in the telecom industry such as 4G, IMS, and SIP-based services, media gateways, security and packet inspection systems, and signaling gateways. The Aerospace and Defense industry can utilize it for sensor acquisition, lawful interception, communications, and control systems. The Enterprise market will benefit by using it for fault-tolerant and high-speed transaction systems. The Industrial Controls Industry will find it ideal for applications such as services.

Rethinking the RTM

The MTCA.4 specification defines the RTM to be roughly the same size as the master AMC card. As shown in Figure 2, the active and hot-swappable RTM plugs directly into the AMC above the connection to the midplane (also called Zone 3). Directly from the AMC’s RPIO connector, it draws power, -based management signals, hot-swap signals, and any I/O signals that are required. Since it is not connected to the midplane, the RPIO connector can be proprietary and send whatever signals are relevant to the RTM. Typical I/O signals can include PCI-Express, USB 2.0, Ethernet, and SATA for storage.

Figure 2: The MTCA.4 RTM connects directly to the AMC above the midplane, rendering the RPIO connector proprietary.

Extending the AMC front panel

The MTCA.4 specification adopted the extended flanges and retention screws for both the top and the bottom of an AMC and RTM faceplate that were incorporated in the “Rugged” MicroTCA.1 specification. This helps to ensure tight and secure retention to the chassis and greatly increases the shock and vibration specifications.

Shown in Figure 3 is an example of a MTCA.4 platform. It features redundant MCHs and MCH RTMs, 12 payload AMC slots and RTMs, redundant fan trays, redundant power supplies, and redundant AC or DC power inputs.

Figure 3: The MTCA.4 platform supports Multiple Controller Hubs, a dozen AMC slots and RTMs, AC or DC power inputs, and more.

MicroTCA.4 on the job

Detailed below are two example applications where equipment manufacturers can take advantage of MTCA.4: 1) Evolved Packet Core (EPC); 2) Military sensor/.

LTE EPC solutions

MTCA.4 systems are well suited for EPC functions such as the Mobile Management Entity (MME), Serving Gateway (SGW), and PDN Gateway (PGW). ATCA has already found success in large EPC gateways for large metropolitan deployments where a million subscribers can be served. As service providers start to deploy in mid-size cities, towns, rural regions, and niche applications such as public safety communications, a more cost-effective approach is MTCA.4-based systems, where hundreds of thousands of subscribers can be served (Figure 4).

Figure 4: MTCA.4 systems provide a cost-effective method for translating LTE services to mid-size and smaller deployments.

In Figure 5, an example system is configured to support all three functions of an EPC. The processors are arranged for both control plane and data plane processing. In addition, other services, such as a Policy and Charging Rules Function (PCRF), can be integrated in a redundant fashion.

Figure 5: As an LTE EPC solution, MicroTCA.4 maintains functionality for control and data plane processing, among other services.

Military sensor/data acquisition applications

As the military deploys more unmanned vehicles, the need for COTS systems that can perform sensor acquisition and data processing becomes more prevalent. These systems can use AMCs with and functionality, Graphics Processing Units (GPUs) with OpenGL for writing graphic-intensive computing applications, and General Purpose Processors (GPPs). Solid State Storage (SSS) can be located in multiple locations: 1) In the RTM behind the processors; 2) On adjacent AMC storage modules; 3) On Mini-SATA drives on the processors themselves.

Figure 6: MTCA.4 can employ AMCs, GPUs, and GPPs in sensor acquisition and data processing.

As the world moves quickly to all IP-based solutions, and with network data traffic growing at explosive rates, there will be more and more need for high-performance aggregation layer applications. While ATCA’s size has made it an excellent solution for the core of the network, it is not ideal for the aggregation layer. But fear not, cost-effective open standards-based platforms are now available with the new MTCA.4 specification. Its introduction could not have come at a better time. With exponential advances in smaller package computing technologies thanks to Moore’s Law, the AMC cards that were originally designed to add functionality to ATCA cards are now powerful enough to handle applications themselves. Add to this the availability of 10 GbE and 40 GbE high-bandwidth networking available in the MTCA.4 specification, and the new standard is designed to support the explosion of network traffic and IP-based services for years to come. Chalk up one more inflection point, Andy!

Tony Romero is Senior Product Manager at PT. Tony has worked extensively in the system architecture and product development of both and MicroTCA platforms for over eleven years. Prior to working at PT, Tony worked for Intel Corporation, Ziatech Corporation, and Dell Computer Corporation.