Vehicle design moves toward a networked future with 1394 automotive standard

OPINION 

By Max Bassler, Chairman, 1394 Trade Association

The increasing application of electronics in the automobile has resulted in dozens or even hundreds of processing islands throughout the auto. The requirement to share data among the nodes and the efficiency of a shared-media interconnect make an automotive network necessary in next-generation designs. Current and future designs include the need to move video around the auto for safety, navigation, and entertainment applications.

 The 1394 Automotive (1394 Auto) standard offers the best combination of topology, performance, cost, and data security for these designs.  And all of the connector and harness systems selected for 1394 Auto are shipping today worldwide.

 The Genesis of the 1394 Auto Standard

 The 1394 Auto standard is derived from, and is compatible with, the widely deployed entertainment and computing interconnect that is alternately referred to as IEEE 1394, FireWire (originally an Apple term but now broadly used), or iLink (Sony). A decade ago, 1394 technology was already shipping in applications such as digital cameras. The designers of the 1394 architecture optimized the interconnect from day one as a peer-to-peer multimedia-capable bus. Jointly, the 1394 Trade Association (1394TA) and the former IDB Forum realized that FireWire capabilities would match future automotive networking needs and together defined IDB-1394 (Intelligent transportation systems Data Bus using 1394 technology).

The original IDB-1394 standard specified the use of optical media in the vehicle because at the time auto designers felt noise would be an issue for copper media. Subsequently, silicon developments and improved shielding technologies have evolved that allow the use of copper cabling in the auto. The 1394TA recently augmented the original work with a copper specification and auto designers now have a choice of media. The 1394TA will shepherd the auto spec now under the brand 1394 Automotive.

The move toward a networked automobile began more than a decade ago. As electronics moved into systems ranging from mission-critical braking and engine control to entertainment and convenience features such as navigation and power mirrors, the electronic islands had to be tied to the driver’s seat with what became bulky wiring harnesses. The harnesses added significant weight and cost, and were unreliable and inefficient to install and service.

 Today, mission-critical systems are tied together with CAN and LIN networks optimized for the reliability and real-time response needed in the power train. The bulk of the electronic devices (or ‘nodes’) in a vehicle now provide telematic or infotainment functions including DVD video and GPS navigation, or driver-assistance functions such as adaptive cruise control or lane-maintenance warnings. These nodes need more bandwidth than LIN or CAN provide, and they are well suited for the multimedia technologies developed in the computer and entertainment market segments, and the car makers don’t want mission critical buses to be affected by passenger compartment systems.

Consider the type of systems that car makers plan to integrate into new designs. Today, some vehicles include simple low-resolution rear-facing cameras that assist a driver in backing into a parking space. Soon, multiple high-resolution, forward-, side- and rear-facing cameras will handle the parking assist functions, allow electronic systems to monitor how well the driver is following the travel lane, and offer adaptive cruise control that matches the speed of traffic. Such features will begin as options in passenger automobiles but may quickly be mandated in commercial vehicles.

At the same time, infotainment capabilities are expanding. Bandwidth-intensive navigation systems that started as driver-guidance aids now allow the driver or passengers to search for restaurants, tourist attractions, or entertainment venues. The systems are beginning to support video and will be integrated with satellite video, interactive gaming, and DVD systems. In high-end vehicles, the automakers plan to offer a high-resolution screen at every passenger seat, and allow each passenger to individually choose the screen content. Back-seat passengers might separately play a game or watch a movie while the front-seat passenger searches for a dining spot and the lane-control camera keeps the driver alert.

 Point-to-point options

Automakers face the choice of using relatively inexpensive and inflexible point-to-point links or more robust networks that offer flexibility. With video support as a baseline requirement, automakers might interconnect video sources and displays in a next-generation vehicle.  The simplest video implementations today use a point-to-point connection from source to display.  Issues of control and synchronization to other auto functions are real important.

For example, single-camera parking-assist systems often rely on a LVDS link between the camera and the display. Such a point-to-point design is a reasonable choice in a low-end vehicle. But automakers also benefit from adopting a single technology for use across their product line, and LVDS technology will struggle to serve in a high-end multi-source multi-display design.  There is no accepted LVDS-industry standard, and different vendors apply different coding schemes.  Some LVDS chipsets implement low bit rate backward channels by time division multiplexing. Virtually any LVDS device on the market uses DC balancing.

Some automakers and third-party suppliers are considering using HDMI, developed for the living room, in cars. HDMI links high-definition video sources such as cable and satellite set-top boxes to HDTV sets.  It carries uncompressed HD-quality streams point-to-point. . HDMI proponents cite the content security features of HDMI as an advantage; and automakers understand the need for a secure interconnect for cars because they can’t take the chance that their designs might allow pirates to capture pristine digital copies of protected music, movies, and other content.

But HDMI is not ideal for an automotive role. Even home consumers have been plagued with expensive HDMI cable or cheaper cables that often don’t work. Also, HDMI works best over distances shorter than two meters, and to use HDMI in a multi-source, multi-display system, designers need to develop a switching center to route the video streams. There are HDMI switches for home theater, but they carry a premium price.

The HDMI interface carries uncompressed video that requires 15-30 times the bandwidth of a compressed stream. However, many of the sources an automobile might use such as DVD are already compressed, and codecs for digital cameras are affordable. By sending it to the receiver in a compressed format and decoding after it is received, there is no need for HDMI.  More important, interconnects such as 1394 Auto with DTCP (Digital Transmission Content Protection) are recognized as secure by content owners.

 There are now network technologies and compressed video for high-end designs. A few automakers have been using MOST (Media Oriented Systems Transport) networks, a standard defined by a group of German automakers and third-party suppliers led by BMW. MOST is purely an automotive standard with no synergistic technology in the computer or entertainment market segments, and with a bandwidth of only 22.5 Mbps, MOST technology doesn’t offer the performance required in next-generation designs.

The standards movement

There is a short list of network-technology candidates for next-generation designs, including 1394 Auto, an upgraded version of MOST, or – and this is a long shot — an automotive flavor of Ethernet.  Automakers who in the past have been quick to develop proprietary technologies now are looking for backbone systems that are both flexible and ‘future-proof,’ able to last and improve through extended product design cycles.

Comparing 1394 Auto, MOST, and Ethernet

Engineering teams considering 1394 Auto, MOST, and Ethernet should consider these variables:

            – aggregate maximum bandwidth

            – media access control

            – isochronous capabilities

- reverse/legacy compatibility

- media choice

- topology

- silicon costs

            – IP license fees

            – silicon industry support

            – synergy with other control technologies

            – synergy with broad consumer technologies

            – flexible topology

            – auto protocols and software

Aggregate bandwidth ultimately limits the A/V features that an automaker can integrate in a vehicle. Looking forward, the automakers plan to integrate more and more video. Consider a likely scenario presented at an automotive forum by Fujitsu in 2007. The example included four driver assistance cameras (rear, front, and each side), a navigation system, a DVD video source, and a satellite TV source. The example pegs the driver assistance camera specs at 640x480x30 (640×480-pixel resolution and a 30 frame per second update rate) with a 16-bit color depth. Each camera generates a 147-Mbps video stream for an aggregate 588-Mbps stream. The DVD and TV systems operating at 720x480x30 and 16-bit color generate 166-Mbps streams. The navigation system operating at 800x480x60 and 16-bit color provides a 415-Mbps stream. All together, just the video traffic can exceed 1-Gbps.

The 1394 Auto standard supports the same range of data rates available in the mainstream 1394 family. Today the IEEE-1394b standard defines data rates as fast as 800 Mbps (called S800 or FireWire 800). In fact, virtually all FireWire products (ICs and end products) support the prior 400-Mbps rates (S400 or FireWire 400) rates and most new products add S800 support. The 1394TA has also defined 1.6- and 3.2-Gbps versions of the specification.

The first-generation MOST technology shipping today operates at 22.5-Mbps maximum rates and is referred to as MOST25. The MOST group has also defined a second-generation technology called MOST50 that would double the data rate. It appears however that the MOST industry will not produce MOST50 products and instead will focus on a 150-Mbps version called MOST150, a spec completed in the spring of 2008.

Clearly a single MOST network will not support all of the A/V traffic that automakers envision and the next-generation MOST technology will fall short.  1394 Auto will support most of the automakers’ intentions today and faster technology is already here.

Ethernet brings the same advantages and disadvantages to the auto that it does to home video-centric networks. While inexpensive, Ethernet remains a best-effort network that is not designed to carry streaming video. There is realistically no way that 1-Gbps Ethernet could serve in the Fujitsu application detailed previously. And low-cost 10-Gbps Ethernet still appears years away. Ethernet as a ‘best-effort’ system also lacks the quality of service guaranteed by 1394.

Backwards compatibility

Auto networks must also be backwards compatibility to prior-generation products. With a backwards-compatible implementation, the automaker can mix existing, stocked subsystems – say a digital audio players that doesn’t require the bandwidth of a video device – with the newest, fastest subsystems. Both 1394 Auto and Ethernet can support all traffic from prior-generation standards on the latest network implementation. Ethernet handles slower traffic at the network switch on a port-by-port basis. The 1394 Auto standard allows any two nodes to communicate at the highest speed that both devices support. A legacy video camera can be connected to an S800 1394 Auto network and it will work fine. Likewise, a new S800 peripheral can be connected to a legacy 100 Megabit/second 1394 Auto bus, and the S800 peripheral will operate fine at S100 speeds.

MOST, however, does not offer backwards compatibility. The MOST150 standard will be fundamentally incompatible with MOST25 technology. That means engineering teams with MOST25 experience start over on the technology learning curve. Moreover there will be no economy of scale with older MOST products that can’t be carried forward to new designs; all new harnessing also will be required.

Media Types

There has also been lengthy discussion of the type of media over which MOST and 1394 Auto can operate. As the idea of an automotive network matriculated, prevailing wisdom dictated that optical fiber would be the best media choice. Fiber would help reduce the weight of a harness relative to copper and would be immune to the noisy EMI and RF environment in a car. MOST was defined from the start based on optical media.  The first 1394-centric automotive spec also specified fiber; IDB-1394 has been fiber based since initial publication.

Today, automotive engineers are reconsidering the fiber versus copper choices, since advances in signaling and silicon technology have solved the EMI and EMC challenges of copper.

While Ethernet certainly operates over copper and optical media in the data center, it’s impossible to judge how Ethernet might be implemented in a car. Ethernet proponents are just starting to work on automotive-centric standards.

Topology

It’s also tough to speculate on an Ethernet topology for the car. To take advantage of the economy of scale from the IT world, an automotive Ethernet network would almost have to use the same Ethernet switch ICs used in IT products, and require lots of support.  Also, the ICs would have to be manufactured to automotive temperature grade.

MOST provides little choice in topology. MOST defines a physical ring topology similar to the original Token Ring LAN defined in the early 1980′s. Unfortunately, ring-based networks have proven less cost effective and even less reliable, unless the network uses redundant dual rings, than other topologies.

Conversely, 1394 Auto is the most flexible of any network or bus-based interconnect. The standard allows for bus, star, ring, daisy chain, tree, and other topologies, and the topologies can be mixed and matched using 1394 Auto in a ring for fail-safe operation that will survive any single cable or device fault.

Cost, maturity, and reliability

Cost matters in any design, and in vehicles is of utmost importance. Many factors such as the number of silicon players, maturity in a technology base, harness flexibility and licensing fees come into play in understanding applied cost.

Without question, 1394 Auto benefits from synergistic deployment in the computer and entertainment space that continues to drive down silicon costs. But it’s not just an advantage in silicon. There are 1394-based industrial control cameras that can quickly be adapted for use in autos. Even peripherals such as rugged 1394 disk drives might find a home in the auto space. And all of the connector and harness systems selected for 1394 Auto already are shipping worldwide.  In contrast, MOST does not have a synergistic market segment that can provide similar leverage.

Automakers also favor mature technologies. Ethernet and 1394 Auto both bring advantages in terms of technical maturity and broad use.  1394 Auto specifically offers a decade of usage in video-centric applications. The 800-Mbps flavor for 1394 technology is already widely available. The original MOST technology is mature. But MOST150 products aren’t on the market yet.

Reliability is another mandate in the auto. With no auto-centric version of Ethernet specified, it’s impossible to speculate how that technology might adapt to the car. There are industrial versions of Ethernet, but no apparent candidates that could quickly make the auto transition.

The original flavor of MOST has definitely proven reliable and MOST150 will likely be reliable too.  But 1394 technology is ultimately the most reliable: it has been deployed in far more difficult environments than vehicles, and continues to be applied.  For example, the F-35 Joint Strike Fighter relies on 1394b technology to link more than 70 nodes in systems that provide real-time mission information, weapon systems control, and engine and flight controls.  That’s a tough and unforgiving proving ground, and 1394 is continuing to be applied in this critical environment,

The 1394 Auto standard offers another advantage:  flexibility. A 1394 Auto network lets engineers provide 1394 connectors at various locations around the car. Auto dealers can take advantage of those connectors to easily install complex options such as navigation, entertainment, and safety systems.

1394 Automotive is the Superior Choice

When you consider Ethernet, MOST and 1394 Auto from the system level, it’s clear that 1394 answers the needs of the automakers and provides superior performance. The automakers came to the 1394 Auto group with three major concerns for future automotive networks.

First, they wanted a mature proven technology and 1394 Auto clearly wins in terms of maturity. Second the automakers wanted a network technology with data security and content protection that would ensure that copyrighted works such as music and video would be secure while moving around the network. The 1394 Auto technology is long proven in terms of content protection in consumer electronics and brings the same advantage to the automobile. The DTLA (Digital Transmission Licensing Administrator) has endorsed 1394 for content protection. Third, the automakers were looking for a technology with a cost-effective and viable forward-looking roadmap – starting at 400-Mbps speeds. The S400 and S800 versions of 1394 Auto are shipping now with faster, backwards-compatible versions on the way.

The reality of the auto network choice comes down to the willingness of an automaker to trade off fidelity, flexibility and cost. Point-to-point links can provide a cost advantage, but lack flexibility. With the limited maximum transfer rate of MOST, the automaker must either use more expensive compression technology (read codec ICs) so that more-highly-compressed video can traverse the slower network, or be willing to sacrifice fidelity. The 1394 Auto network allows the automaker to choose proven low-cost video compression technology and still offer fidelity superior to what MOST can offer. And 1394 Auto offers maximum flexibility in terms of topology, media, and ease of use.