Around 700 participants, speakers and exhibitors from 20 countries - the second Automotive Ethernet Congress on February 3 and 4, 2016 in Munich has once again grown significantly compared to last year's already very successful debut. The first keynote speech showed how such concentrated Ethernet expertise can be used to the benefit of all participants: Kirsten Matheus, BMW, repeatedly involved the audience in her presentation by conducting short surveys.

The most important results of this almost representative assessment of the current situation: the vast majority of participants have long since left the pure evaluation phase for the use of Ethernet in cars behind them, with around 20% even supporting the entry into the standardization of a 10 GB/s variant of Automotive Ethernet by a show of hands. In contrast, far fewer hands were raised in favor of a 10 MB/s variant brought into play by Matheus as a direct CAN competitor. Part of the appeal of Ethernet is precisely the prospect of high data rates at low prices.

However, Matheus emphasized that reducing automotive Ethernet to just a fast data highway is not enough: "With its service-based architecture, Ethernet is the best way to get to grips with the increasing complexity of modern vehicle networks." The switch to Ethernet does not simply mean an increase in available bandwidth, but also forms the basis for a completely new vehicle architecture.

Aurora Sere-Schneider, Continental, even spoke of an "architecture revolution" triggered by Ethernet in her keynote speech. Vehicle networking and the Internet of Things are the major drivers of this development, right through to future disruptive cloud-based architectures, which have long been commonplace in the IT world. However, Sere-Schneider also pointed out a serious difference between IT and automotive applications: the latter often have to meet strict ASIL requirements - a simple 1:1 transfer of existing solutions is therefore out of the question.

Standards and toothbrushes

Despite a generally optimistic mood, several presentations warned against calling into question the successes already achieved in standardization through individual special solutions. Harald Zweck, Infineon, summed up this problem with a nice metaphor: "Standards are like toothbrushes. Everyone agrees that they are useful. But nobody wants to use someone else's."

But you have to differentiate here: At the PHY level, for example, standardization efforts at 100 MB/s and 1 GB/s have been very successful so far. The situation is somewhat different for Time Sensitive Networking (TSN) for time-critical applications, for example: "There is no one TSN standard," Daniel Zebralla, Continental, made clear. Rather, TSN is a large toolbox that can be used as required - depending on whether latency times, traffic shaping or the reservation of resources are involved. As the individual standard elements of TSN are in most cases independent of each other, they do not necessarily have to be further developed together. However, the term TSN remains vague - it is always necessary to specify which variants are meant in each individual case.

With reference to PHY standardization, Mehmet Tazebay, Broadcom, once again took up the 10 GB/s variant of Automotive Ethernet mentioned by Matheus: He was skeptical that transmission via an inexpensive unshielded twisted pair cable (UTP) would still be possible at this data rate. Instead, an intermediate stage, for example at 5 GB/s, is also conceivable for the time being - if that is what the vehicle manufacturers want: "First the OEMs have to agree on a specific value for the next development stage, then we will find a solution."

Deployment above the clouds

In addition to presentations on PHY & Cable, architecture, communication and test & validation, the topic of security also played an important role at the congress. Automotive Ethernet is a great help in "making security an integral part of the car", emphasized Clara Otero Pérez, NXP, in her keynote speech. In fact, Ethernet already comes with numerous mechanisms to prevent unauthorized access to the network. However, as Yong Kim, Broadcom, and Roman Pallierer, Elektrobit, explained in their presentations, it is crucial to activate these mechanisms in a targeted manner. By default, Ethernet is far less protected than it should be for use in cars.

Even if more and more standards from the IT world can be found in cars in the course of vehicle networking, the expectations of the respective customers still differ considerably: "IT buyers know that they have to spend extra money on security. Car buyers simply take this security for granted," said Martin Brunner, Secunet. The same applies in both cases: Security is not a state, but a process that encompasses the entire product life cycle. The speakers agreed that security should not be a differentiating factor. However, there is still no committee in sight that would define generally binding security standards.

The fact that Ethernet can be extremely secure is demonstrated by its use in the avionics sector, where it forms the basis for networking the fly-by-wire system as so-called Avionics Full Duplex Ethernet (AFDX). In the future, however, Automotive Ethernet could also be used above the clouds: Airbus is examining the use of this significantly cheaper Ethernet variant for cabling the passenger cabin. In his presentation, Sebastian Hahn, Airbus, reported on extensive tests with 100-base T1 systems and confirmed that the technology also has great potential for use in aircraft.

However, there have also been numerous concrete advances in the use of Automotive Ethernet in ground-based transportation: Representatives from Audi, Volvo, and PSA Peugeot Citroen, among others, explained how they plan to use Ethernet in their vehicles. And in the accompanying exhibition, 30 companies presented concrete products and solutions - from tiny semiconductor components to refrigerator-sized HiL test systems.

The next Automotive Ethernet Congress will take place again in Munich on February 8 and 9, 2017. Anyone interested can already register online to receive further information.