How Connectivity is Changing in the Face of SOSA

Connectivity is Key to the New Standard

As the SOSA Technical Standard progresses, the recent herculean efforts undertaken by the technical working groups have elevated the standard even further. The additions to Snapshot 3 further define system realizations in accordance with the MOSA directive from the DoD and other government agencies.

Although several factors in systems design, network speeds and power supply requirements have been examined and defined as the SOSA initiative nears its first release, one of the areas most impacted is connectivity. With SOSA Release 1.0 slated for mid-2021, it’s important to take a look at the evolution of the connectivity through Snapshots 1, 2 and 3, paying special attention to the accommodations made for speed increases and data throughput to meet the growing bandwidth needs of applications.

Network Needs in Systems Aligned to SOSA

Ethernet has evolved in its capability to allow for high speed, low latency transfer of data, and can now move data quickly with priority, replacing interfaces like PCI Express for data plane communication. It has become the primary network technology specified for use in SOSA systems. It supports many of the goals in SOSA including configurability, interoperability and is widely used. (Figure 1)

Figure 1: Example of a 3U Backplane Profile with Network Slot Profiles

In fact, SOSA relies extensively on Ethernet networks to pass information between SOSA modules, and currently supports high speed 1/10/40 Gb Ethernet (with 100 Gb in development) as well as low latency and deterministic transfer of data. High speed switches are used to implement Ethernet plug-in card switches.

SOSA systems are comprised of SOSA modules that must communicate. This means that SOSA module interconnects provide facilities to pass information between SOSA modules and physical network interconnects are defined to pass the information between SOSA plug-in cards.

Because the SOSA standard defines plug-in card hardware by OpenVPX VITA 65.0 and VITA 65.1, OpenVPX backplanes implement slots to accept plug-in cards and allow for interconnection between SOSA hardware entities to implement the necessary networks required in SOSA systems.

Network connections are implemented over the in the interconnected pipes found in the backplane, and the pipes are logically organized into various planes in OpenVPX. These connections may be physically implemented in copper in the backplane or externally via fiber optics.

System Requirements Defined by Industry Parameters

Although initial targets of the snapshots incorporated high degrees of network speed, as we prepare for the first release of SOSA, it’s evident that the need to meet these increasing throughout needs has resulted in shifts to the connectivity requirements in SOSA systems. (Figure 2).

Figure 2: Increasing networks speeds have influenced performance attributes in SOSA aligned systems.

The need for this increased performance has influenced several connectivity design factors

• A new VPX connector (TE’s RT3) was developed to support speeds of 25+ Gb/s
• Module profiles were added to SOSA and OpenVPX to describe the required protocols
• VITA 67.3 I/O connectors were updated to support RF contacts and fiber optic MT ferrules
• Shift were made in slot profiles for payload & I/O-intensive SBCs as well as for switches and timing slots

‍

Ensuring Connectivity as the SOSA Ecosystem Builds

Given the nature of SOSA, designed to foster interoperability across manufacturers and product technologies, ensuring connectivity within a system is critical.

One notable industry event that clearly demonstrated the strength of the SOSA ecosystem as well as helped validate the interoperability of the SOSA aligned components was the first Tri-Service Open Architecture Interoperability Demonstration (TSOA-ID) event for the media, acquisition community and industry influencers, hosted in January 2020 by the Georgia Tech Research Institute in Atlanta, Georgia.

Elma’s SOSA and CMOSS Aligned 12-Slot 3U Development Platform served as the heart of joint efforts by five SOSA Consortium member companies to build and test a fully functioning system aligned to SOSA. These industry partners, that participated in the demo using Elma’s development platforms included Behlman Power, Concurrent Technologies, Crossfield Technology, Curtiss Wright, Interface Concept and Spectranetix (a Pacific Defense company).

Connectivity needs in systems aligned to SOSA have kept pace with changing market demands, even during the initial stages of the standard’s development. Ensuring reliable throughput for high network speeds creates a path for continued strengthening of SOSA across all branches of the DoD.

Learn more in this informative webinar “How the SOSA Technical Standard Implements Highly Standardized, Configurable and Interoperable Network Communications Protocols”, hosted by Military & Aerospace Electronics, and experts from Elma Electronic, Interface Concept and Pentek.

Upgrades and Enhancements for Legacy VME and CPCI Ethernet Switches

 In the past few years, several end-of-life (EOL) announcements in the embedded computing market have both caused angst and opportunity. Making the shift away from a tried-and-true solution always brings with it the need to review not only the mechanical elements of an embedded system, but the integration and networking elements as well. And when that review is forced upon a designer, as in the case of an EOL announcement, it may mean forced choices of not-as-optimum alternatives. Or it could be something different altogether.

EOL is not the only option

One product segment particularly affected is high-performance embedded Ethernet switches in VME, CompactPCI, and VPX form factors. Component availability triggered board-level EOL notices, with some suppliers forced to reduce or eliminate product offerings in some of these form factors.

But others looked at this shift as a way to strengthen the Ethernet switch and add new capabilities by implementing a technology refresh to an existing product line, instead of just shutting production down or scaling back the availability of its switches. The ComEth product line by Interface Concept (sold in North America by Elma Electronic Inc.) is an example of this how lack of component availability does not need to result in an EOL announcement, but rather, can foster new opportunities to help ensure a strong embedded ecosystem.

How Ethernet speeds have evolved

Created in the early 1970s, Ethernet has seen many improvements and enhancements over the past several decades, with the speed of data flow increasing again and again. Today, Gigabit is considered the low-end, with many embedded systems requiring Ethernet connectivity speeds of 10, 25, 40, and even 100 Gigabit. It can even be faster than that, but 10 and 25 Gigabit SERDES lanes, which can be combined in groups of four to form fat pipes supporting 40 GigKr4 and 100 GigKr4, are commonplace in today’s embedded systems market.

In addition, these embedded systems often utilize Ethernet switches when more than a single point-to-point connection is required. The ComEth family of Ethernet switches is a good example of a modern switch that provides the high performance switching and routing features most desired by system integrators. This includes a feature rich switch management application called Switchware that lets the user configure an extensive list of parameters and configuration settings using either a convenient browser interface or Command Line Interface (CLI). Once a switch configuration is created, the configuration file is saved, exported and then loaded onto other switches, facilitating easy configuration across each switch.

System enhancements for VME & CompactPCI

Interfaces such as to VME and CPCI backplanes remain at 1000BaseT speeds due to the limitations of the backplane connectors, although other interfaces, like front panel ports, have been updated to support faster uplink speeds. For example, some ports that supported 1000BaseT in the past now offer a wide range of speeds up to 10 Gig-BaseT capability. Such ports often support 10 and 100 Megabit, Gigabit, 2.5, 5 and 10 Gigabit speeds.

This facilitates overall system enhancements, where other equipment in the system may also need to accommodate bandwidth increases. It is interesting that sometimes Gigabit is not fast enough and 10 Gigabit is more than enough, so these intermediate speeds of 2.5 and 5 Gigabit may help solve the problem of matching the switch’s speed to the needed bandwidth. There is also more use of SFP, SFP+, and QSFP ports, allowing users to select — and even mix — a variety of media interfaces.

Some systems require both copper and fiber of different speeds and a modern commercial-off-the-shelf (COTS) Ethernet switch allows the system integrator great flexibility to be compatible with higher-level system requirements. In some cases, even the need for quite old 100 Megabit fiber interfaces is still supported with these new designs for applications that must have it.

Backplane designs for VPX

In the VPX market, backplanes often utilize SERDES interfaces, which had been Gigabit or 10 Gigabit speeds for the past few years. With enhancements to VPX backplane designs and connector technology, new generation VPX form factor Ethernet switches now support speeds of Gigabit, 10 Gigabit and 25 Gigabit SERDES on a single lane and, when four lanes are combined into a fat-pipe,current products are supporting 40 GigKR4 and even 100 GigKr4 speeds, with more speed improvement on product roadmaps.

Enabling innovation instead of EOL shutdowns

Component availability is an issue we will be facing for a while, but by using this challenge as an innovation opportunity, embedded system designers can move towards new models of development to capitalize on computing technologies, like enhanced speeds in Ethernet switching, and improve system performance.

Browse the entire line of available Ethernet switches.

Factors of Increased Heat Generation in OpenVPX Systems

 It’s no secret that higher performance means higher thermal management
 requirements. Denser electronics packed into smaller spaces oftentimes leaves designers with the challenge of finding more creative ways to dissipate the increased amount of heat for conduction-type cooling methods. OpenVPX enables extraordinary leaps in aggregate system bandwidth and processing speeds that mandates new methods to meet the resulting thermal challenges.

OpenVPX has introduced optical and RF signals to the backplane, removing these otherwise discrete connectors from the front of the cards. While the new backplane connections eliminate what would otherwise be a jumble of cables, the aggregate high-speed signals that now traverse the backplane rapidly heat up the system, exacerbating the already difficult-to-manage temperature increases.

Some of the most complex cards are being used in applications such as signal intelligence for communications and to record signals on the battlefield — including enemy communications — taking in audio inputs and triangulating the source of enemy fire.

Many high-performance applications require processor and FPGA (Field-Programmable Gate Array) system bandwidth that drive up the thermal load on the inside the chassis, necessitating the need for new thermal management strategies. One example is a recent aerospace application that required many RF inputs — 36 payload slots each with 16 RF signals and many large radar arrays that require vast amounts of RF I/O signals.

Tight Spaces Mean More Heat

Embedded sub-systems must sometimes be packaged to fit existing tight spaces in aircraft, ground vehicles, submarines, spacecraft and other rugged, compact environments, and has led to the need for optimized SWaP-C (size, weight and power-cooling*). While OpenVPX offers significant improvements in field-deployed system signal integrity, speed and capability, it has created new challenges in these space-constrained installations.

As higher performance systems are implemented, the choice between 3U VPX and 6U VPX becomes a matter of what functionality can be packaged on the smaller card vs. the larger. And as processors and FPGAs enable more capability, the 3U VPX form factor is favored for its reduced size and weight. This pushes the existing convection and conduction cooling techniques defined by the standard to their limits.

That concentration of power in a smaller board has heavily impacted chassis and backplane designs and complicated thermal management in systems using a 3U card, making heat dissipation a larger issue. However, new cooling options under the VITA 48 umbrella are working to accommodate the increased heat in these high performance systems.

Beyond Traditional Convection and Conduction

Most current applications find conduction cooling, as defined by VITA 48.2, and its respected cohort convection cooling sufficient. But the added complexity and heat generation of new boards and connectors quickly push current system cooling methods beyond these defined limits.

As VPX has grown in popularity, the VITA standards committees have defined additional cooling methods under VITA 48 to ensure future thermal needs are adequately handled. Current iterations are:

• 48.4, liquid flow-through, probably the most efficient up to 450 watts per card
• 48.5, air flow-through, which has the advantage of being able to meter the air to specific cards
• 48.7, air flow-by
• 48.8, air flow-through cooling without sealing for small form factor 3U and 6U VPX modules (ANSI ratified October 2017)

The environment in which the boards will be developed and tested is typically different than the final deployed unit, so a lab chassis, for example, can usually rely on just fan cooling, whereas a deployed unit might need conduction cooling. The proper cooling method for a deployed system should be based on the most practical design and take into account the housing, the card heat sink and the chassis itself.

See blog post on VITA 48.4 and alternate cooling methods

OpenVPX into the Future

OpenVPX has allowed new definitions for VPX backplanes and systems, giving system architects and end users a far wider range of choices in critical high-speed applications, paving the way for more open architecture and multi-vendor interoperability in the future. It fosters technology growth over time, without requiring changes to system architecture. It uses adaptations within the standards themselves to enable new capabilities and build HPEC hardware.

System density is only increasing, and end users are still searching for ways to fit smaller boxes into more compact spaces, so they can put even more electronics into their applications. Which, or course, means more heat.Get more details and information on elma electronic