An Open Standard Architecture Viewpoint on What’s Next for AdvancedTCA and MicroTCA
The xTCA architecture’s comfort zone in mission-critical markets is only growing stronger, making expansion and longevity part of its future. Here’s why.
In the past decade, the AdvancedTCA (ATCA) open standard architecture has seen multiples of billions of dollars in revenue. The powerful architecture was a behemoth in the Telecom market, with Terabit/sec bandwidth, 10GbE to 40GbE signaling, 99.9999% uptime with full redundancy, and a wealth of great products from dozens of vendors. Less well-known was that ATCA and its sister-specification, MicroTCA, had also expanded into other mission-critical markets. Communications systems are moving to the cloud with cheap servers and monochrome technologies like Open Compute. Where do xTCA technologies fit now?
The Killer Cloud
The last mile proximity feature of virtually zero downtime is not as critical these days, with network virtualization and “good enough” services. Servers all over the world are delivering as “the cloud,” thanks to the Internet, thus the tremendous premium of space in data centers has decreased considerably. Server farms can be in much cheaper areas with creative solutions for cooling the buildings housing them and the electronics inside.
There is little differentiation in maximizing Ethernet data transfer as quickly as possible, therefore the advantage of an open standard architecture has also decreased. Open standard architectures like AdvancedTCA® have had the advantage of dozens of member companies creating various configurations of boards and chassis platforms. Of late, however, the requirements of the Telco market have become near ubiquitous, and no longer need the open architecture’s potential versatility. Where is there a place for AdvancedTCA in the Telco market going forward? What about other markets?
One area where AdvancedTCA still shines, and will for some time to come, is in performance density. In a limited space, the speed and reliability are second to none. Today’s ATCA chassis platforms boast speeds of 40GbE for data (and even 100G capable, although 100G switches are not standardly available on the market yet) and cooling to 400W/slot or more. Nearly 5.6 kilowatts of power/heat is dissipated, and powerful, multicore processors pack more punch than ever. ATCA’s large board footprint of 8U high x 280mm deep provides plenty of space for fast (and hot) processors. A wealth of chassis configurations for 2, 5, 6, 14, and other slot sizes exist in various cooling configurations (front-to-rear, side-to-side, etc.), as well. Figure 1 shows a 15U ATCA chassis from Pixus Technologies with four individually hot swappable, intelligent 576 CFM fans for systems that require cooling for a whopping 400W per slot.
Lest we forget the advantages of open standards, in this industry there have been instances where a company or armed service uses a proprietary solution and gets burned. Suppliers can orphan proprietary solutions, announcing they are no longer offering a certain product, or they may sell a vital business unit to someone with conflicting interests. (Recall that IBM sold its BladeCenter unit to a Chinese interest.) With an open standard, you have multiple vendors to choose from and a rich differentiation of products to provide solutions for every need. This is one of the reasons why AdvancedTCA has seen success in other markets such as Defense (more on this later).
ATCA is an Open Specification (like an Open Standard), which is not the same thing as Open Source. With open source, there is little differentiation since vendors are given specific BOMs, Gerbers, and other resources to build a monochrome, albeit open source product. The governing body for the ATCA open specification is called “PICMG,” the leading standards development organization in the non-desktop, or embedded computer market. With open standards and therefore open specifications from PICMG, multiple vendors gather to define a set of interfaces that work together with design guidelines. The result is a rich array of product features. Although the Telco market has narrowed itself down to a smaller feature-set of requirements over time, beneficial ATCA options were available that could have been leveraged to greater benefit. For example, the carriers for Advanced Mezzanine Cards (Advanced MC®, or AMCs) provide a wealth of I/O and storage options. Carriers can hold several AMCs with storage modules and the Rear Transition Modules (RTMs) can hold storage as well (Figure 2).
The xTCA architectures have other beneficial features with respect to the shelf manager. The architecture leverages the PICMG Hardware Platform Management (HPM) specification, whose core elements are now utilized by competing standards organizations and Fortune 100 companies. Of course, the shelf manager can have a power supply take over if another unit fails, can control the speeds of the fans if one fan goes out, and so forth. There are also features for Field Replaceable Unit (FRU) identification. For example, when a board is plugged in, it can identify the serial number and manufacturer of the board. And it will know how much power needs to be allocated to that board and can sequentially power up boards as needed. The shelf managers know the speed of the fans in the system and whether they are starting to work harder for the same level of cooling, which can indicate some sort of impending issue, such as a dirty fan filter. A potential failure can be averted before it ever happens.
In some MicroTCA systems chassis locator features are available, such that a single chassis in a massive data center can tell an operator (via a remote signal) exactly where it is located, flashing an LED for easy identification. There are also JTAG Switch Modules, which allow a toggle to be plugged in (in just one slot) for software uploading or de-bugging to any board in the chassis.
Overall, ATCA will still have a place in the Communications market for applications that need powerful performance and reliability. Of course, much of the ATCA infrastructure will remain with the architecture for upgrades, and remain scalable; moving to even higher performance. As ATCA’s share in the communications market has subsided in recent years, its presence in the Mil/Aero and Research/Physics markets has continued to grow. The P-8 Poseidon, CANES, and several other projects in the military have moved to ATCA. Increasingly, the market is seeing more specialty cards in form factors such as A/D and D/A converters and multi-faceted modules that combine the performance of FPGAs, processors, and carrier in a single board.
Another area of growth for ATCA is the Research/Physics market. Multiple national labs have experiments or applications that require massive computing power and specialized computing. ATCA’s strong ecosystem and scalability has made it a natural choice. The High-energy Physics market has also utilized ATCA with 99.999% uptime, superior reliability, rear I/O capability, and superior performance. There is also an “ATCA for Physics” specification that focuses on precision timing, I/O, and other features critical to the segment. The Stanford Linear Accelerator (SLAC) has recently adopted a massive amount of ATCA systems, and where SLAC goes, other labs seem to follow.
MicroTCA is used heavily by one of the most advanced communications/test companies on the planet. AMCs are typically about 75mm x 280mm x 0.8” wide, which allows up to 12 AMCs to fit within a 1U enclosure (when chassis is loaded in the front or the rear). With superior SWaP advantages to 6U and 3U architectures, the MicroTCA form factor offers tremendous performance density. High-end communications systems will likely continue to use MicroTCA. The architecture has carved out a strong niche as the high-performance specialist, offering performance density in a small size that no other open standard architecture can achieve. Much of the High-energy Physics market is also using MicroTCA.4, a specialized version for these labs. Figure 3 shows a MicroTCA system in DESY’s accelerator in the European XFEL project. DESY (part of the Hemholtz Institute) has been a key leader in the MicroTCA.4 efforts, with development of solutions with the critical precision timing needed for the experiments.
The MicroTCA architecture can offer the tremendous reliability, shelf management, and six-nines uptime (99.9999%) of ATCA in one-fourth the size. It’s a great fit for applications that don’t need the large board space available in ATCA. MicroTCA also has the benefit of inherent provisions in the core spec for precision timing with three dedicated clock lines. With its SWaP advantages, the architecture has been used in several high-performance mil/aero applications. It’s been used in NASA satellite systems, naval towed arrays, ground RADAR systems, and more. The technology has also been used in some of the coolest high-tech applications on the planet (most of which we can’t talk about. Shhh!)
The Road Ahead
What’s next as ATCA and MicroTCA move forward? AdvancedTCA is a trusted open standard architecture that will be used in powerful communications applications (whether Telecom or Mil/Aero) as well as complex systems requiring its board size, proven reliability, and impressive performance. MicroTCA might be expected in very high-end communications/test systems or any application requiring MicroTCA’s performance density and specialty features. The latter applies across applications in Mil/Aero, Test, Research/Physics, and other areas. Overall, the future remains strong for the xTCA architecture.
Justin Moll has been the Vice President of Marketing for PICMG since January 2017. He is also Vice President of U.S. Market Development for Pixus Technologies and a dedicated consultant. Previously, Moll was Director of Marketing for VadaTech and for ELMA Bustronic. He is the Chair of the Higher Speed Ethernet Fabrics for MicroTCA.0 and AMC.2 committee in PICMG.