Friday, March 13, 2015

Connected Airplane - How did we get here, what's next, and what are the big challenges?

The Connected Airplane
Operational data link will retain ACARS for the foreseeable future.

ATN applications will break loose once ATN-IPS is widely available.

The majority of data link growth will be in airplane health monitoring, which can use COTS radios.

There is no end of COTS radios coming onto the marketplace.  Regional and global coverage by multiple Ku band and Ka band service providers, using geostationary fixed and low-earth-orbit moving constellations.

On-ground cellular radio networks are the greatest data link bargain in aviation history.

AeroMACS is struggling with Catch-22 or Field of Dreams challenges between equipping the airplanes and the airports.  The difference (with WiFi) is 59 MHz of sweet dedicated spectrum set aside for safety communications.  We just need the governments to decide to allocate the funds, find qualified suppliers and build out the network.  There is also a little issue of making sure it doesn't actually interfere with Globalstar uplinks (who are watching the noise floor very closely.)

Securing the airplane networks from each other and untrusted public networks must be resolved for each radio network, airplane network, and end-system application.  Open Internet standards will guide the way to apply security, but constrained by the unique demands to develop and certificate to aviation-specific standards.  Getting industrial agreements may become a bottom-up approach to this issue, rather than a strict top-down edict.  The challenges of network security are not unique to aviation, agreeing on a standard and deploying equipment are.

Passenger applications would be best applied surgically and specifically, rather than broadly and boldly.  Low cost, low bandwidth is more popular than high cost, low bandwidth.  Setting expectations will motivate the migration, and high-end services will benefit by the managed approach.

Broadcasting live content offers the greatest entertainment benefits for the least cost in spectrum, at least in a pathway to engage every passenger with a free service.

The prevalence of passenger cell phones capable of operating over a WiFi network will undermine the benefits of a mobile cellular base station.

ACARS is a mobile data link protocol developed in the late 1970s.  I sent my first ACARS message from a lab while working for Boeing Commercial Airplanes in 1989.

Data link was part of the enhanced automation enabling the 747-400 to be flown without a flight engineer.   The 747-400 was the first airplane designed with data link as a core feature.  Maintenance information, crew messaging, flight plans, controller-pilot data link control (CPDLC), automatic dependent surveillance - contract (ADS-C), cabin crew messaging, connecting gates, airplane health monitoring are just some of the data link applications operational by 1995.

The 747-400 relied on VHF ACARS (packet 2.4 kbps) and classic Inmarsat Satcom ACARS (packet up to 10.5 kbps).  Satcom also offered multi-channel voice for passengers, crew, and pilots, a 2.4 kbps dial-up service including support for an on-board fax machine, and another packet channel for any data link application other than ACARS.

ACARS never gained respect in the industry.  Efforts to replace ACARS by the Aeronautical Telecommunications Network (ATN) have been continuous since the late 1980s.  It is sensible to want to move away from a limited character set store-and-forward service towards a bit-oriented mobile network.  With the advent of ARINC 622, ACARS could carry bit-oriented ATN applications like CPDLC.

While the 747-400 is out of production, no airliner since has been developed without fully embracing ACARS, even becoming a built-in feature.

Data Link Today
That was then.  And that is still largely true today.  ACARS works, and is good enough.

Thankfully radio systems have evolved.

ACARS VHF data link mode 2 (VDLM2) stepped up the data rate to 31.5 kbps over the same 25 kHz channel that used to bear only 2.4 kbps.  VDLM2 was developed for ATN, so this time we had to envelope the character oriented ACARS messages.

Inmarsat now offers SBB providing a native IP network connection and data rates that can exceed 300 kbps on one radio channel.

Commercial-off-the-shelf (COTS) radio systems using Ku band or Ka band and very large antennas are capable of delivering over 100 Mbps to an airplane.

Iridium Short Burst Data (SBD) provides global ACARS connectivity.  Iridium Certus data rates over 1 Mbps should be evident in 2017.

By 2018, several LEO satellite constellations (SpaceX/Google, OneWeb, LeoSat) may be launching offering global broadband data rates.

Gogo, formerly Aircell, has been offering a line-of-site service since mid 90's, at first re-using cellular services and now with 4 MHz of dedicated spectrum.  This spectrum was originally used by AirFone in the 1970s.  AirFone eventually offered dial-up data link around 9.6 kbps, telephony, and a hosted portal for breaking news and later a suite of cached content and Tenzing proxy email.  Now Gogo is able to bring as much as 10 Mbps to an airplane.

Reports from Gogo and others profess imminent announcement of a US FCC auction for some part of the 500 MHz between 14-14.5 GHz, for line-of-sight application (LOS).   LOS Ku would operate on a non-interfering basis to both existing geostationary satellite and emerging low Earth orbit satellites.  Conceivably more than one low Earth orbit satellite constellation will emerge in the next five years.  The LOS Ku-band service will operate within a link budget bearing no resemblance to the existing ATG network, so it is not clear what spectral efficiency these frequencies will yield.

Connecting the airplane just while it is on the ground has been a goal also reaching back into the 1990's.

There are over 10,000 airplanes flying today that turn on a cellular data modem when they land.  Reports that cellular doesn't work are akin to claims denying climate change.  7 kbps was laudable with 1G dial-up.  2G GPRS stepped up to around 40 kbps, then 2.5 G EDGE to about 120 kbps, 3G UMTS, EVDO, 4G LTE, HSPA+, 5G, my oh my!  Cellular data rates are now exceeding 100 Mbps.  Small cell networks are just gaining traction to serve the higher density demands in public and urban locations.  Airplanes leverage the investments from cellular networks to serve their broad base of customers.  Costs are dropping to unheard of levels; less than $1 per GB will be here soon enough.

802.11 WiFi has not become a prevalent aviation service due to a number of challenges and barriers.  The technology is less of an issue.  Getting every airport to equip and managing the spectrum are big issues.  Until enough airports are equipped, airlines won't spend the money to put on the radios.  Until enough airlines are participating, airports won't equip.  ISM frequencies are rather congested at airports.

AeroMACS is an 802.16e WiMAX initiative using set-aside spectrum for air traffic control applications.  The protocol permits using the service for other applications as long as to not degrade the channel.  Over the next few years, AeroMACS installations by both governments and by airports will create a private network that can serve an airplane on the ground at data rates perhaps as high as 10 Mbps on a 5 MHz channel.  It may take 100 access points to cover a given airport with small cells and high frequency reuse (in the gate areas particularly).

Wired networks to the airplane have not really gained traction.  Ethernet over power-lines is a feasible technology with data rates approaching 100 Mbps.  The fastest, and most secure data link to the airplane is provided by JetwayNet, with dedicated data rates over 10 Gbps available.

Sneakernet remains a method of choice, even today.  High valued content, software data loading, limitations in avionics are factors that lead a person to walk out to the airplane with a flash drive or a laptop computer.

Data Link Tomorrow
There is no end to the variety of data link options standing ready to serve aviation, or forthcoming in the near future.

Aviation relies on ICAO to develop Standards and Recommended Practices (SARPS) and RTCA/EUROCAE to develop Minimum Operational Performance Standards (MOPS) to create interoperable air traffic control.  Understandably, the reliability, integrity, and performance of a data link application used for aircraft separation is a matter of public safety.  While of great import, messages from these applications are quite small in size and frequency.

Commercial-Off-The-Shelf (COTS) radio systems are not developed to meet SARPS and MOPS (as they don't exist), or may offer limited controls favoring aviation.  There is no Technical Standard Order (TSO) applicable to a COTS radio.

Required Communication Performance (RCP) is a means to quantify a COTS radio service for utility with various applications without being prescriptive to its design.  RCP lets any service provider build a body of evidence professing their reliability/availability, integrity, and performance.  Its the show-me approach to qualifying data link.  And the oversight never stops.

The spectrum underlying the radio system and the licenses associated to it are of great importance.  Only three bands are licensed for "Route" (ATC) data link - comms: HF, VHF, Inmarsat & Iridium L band.  AeroMACS C band has the designation, but I don't think they have applied for a license (FCC) yet.

Route frequencies are the first priority for Route applications.  Use of any other frequency bears the loss of aviation priority, which leads whether a COTS link can be trusted to be available.  There is no fundamental risk by virtue of using a COTS radio. Law enforcement, military, and public safety applications are frequently carried over COTS data link service providers.

COTS radio systems can service any non-Route application, and could supplement a Route application as long as a Route radio remains available, and the time to recover from any failure of the COTS radio does not create a safety hazard.  The greatest benefit is presenting parallel data links, offloading from the Route radio to the COTS radio whenever possible.

Every year I see conference proceedings expressing the benefits of a connected airplane.  I feel that we re-discover these benefits over and over again.   What is holding us back?

It costs a lot to change anything.  We have to study, analyze and build a business case.  Suppliers, providers, and options keep evolving. The future technologies seem to be so much better than what we have now.

ACARS is alive and well.  However, there are some chinks, particularly on the ground in Europe, where congestion has arisen once again (since long ago upgrading to VDLM2.)

ATN-IPS is emerging as a way to embrace an IP network that can include a radio (COTS or otherwise).  Finally, ATN applications penned-in (ha ha) by OSI, could be free to roam over any IP network.

Regional mandates are driving airlines to equip for using CPDLC and ADS-C over an ACARS network.  The challenge is getting airplanes equipped, air traffic control equipped and flight crews and air traffic controls trained.  There is no end of delays along this point.  It takes money to make the changes.  Clear benefits need to build a business motivation.  The market place finds it easier to do nothing.  Whatever we do today, something better will emerge tomorrow.  Aviation safety brings forth conservative mentality fearful of change making things worse not better.

There is a middle ground.  The administrative communications, and to some extent the airline operational communications.  This is where transaction processing, airplane health monitoring, meteorological applications, and the electronic flight bag (EFB) live.  These applications are served poorly by the safety radio channels, if comparing both cost and performance to COTS radio channels.   The volume of this data traffic is currently over half the total data link burden.  The flood gates are opening with the advent of faster and lower cost broadband channels.  But today, few systems other than EFB could take advantage of the connectivity. It will take avionics upgrades to drive larger participation across airplane systems.  Intermediate servers are being offered to gather the information across the various avionics systems and then utilize a COTS radio as a store-and-forward secure gateway.

Network Security
Network security is the latest challenge to a connected aircraft.  While security by obscurity has served aviation well, the threats are much more sophisticated, and the reach across the airplane more pervasive.

Threats to an airplane network can arrive through a radio connection or through the onboard network.  The Ethernet networks on an airplane are distributed and federated to segregate the various levels of systems from those a part of aircraft control to those serving passenger owned devices.

AEEC is developing Project Paper 848 to provide a framework to ensure a radio can securely connect across the various levels of Ethernet networks.  The framework also extends secure tunnels across the open, public Internet into authorized data link services.  JetwayNet OASES is a one example of Project Paper 848 functionality.

Passenger Services
Serving the passenger as if they were in their home or office is a fool's errand.

Until recently, I always compared in-flight data link to cellular, as they tracked together. Then cellular LTE emerged and we are chasing it now too.

LTE is a curious waveform, given its many-to-serve-one approach to channel makeup.  Numerous trials have shown LTE is suitable for mobile applications, and it appears to be a very simple technology to adapt for aviation utility.  LTE is a favored choice for all new line-of-sight services.

I cannot understand how the public will ever be satisfied if they believe they can receive an equivalent service to what they are accustomed.  How do we deal with the cost/price/profit issue?

Net Neutrality brings forth a truth-in-advertising mandate that will be eye-popping to read.  Marketing speak of 100 Mbps will be reduced to what an individual can expect, and that may well be under 100 kbps.  At 100 kbps, it might take a couple of minutes to load an average web page (2 MBytes plus 95 serial http requests) even with performance enhancing proxy (PEP).

Streaming is a service that the industry will forever be chasing.  While I would be happy with a 480p at 1 Mbps experience, you might need 720p at 2.5 Mbps, or 1080p at 8 Mbps, or even 4k with over 30 Mbps.  High efficiency codecs will moderate these numbers only slightly.  It's hard to offer a retail price point below $10 an hour if you are streaming just 1 Mbps (or about 1 GB for $20.)

Crafting a host of specific, low-data, applications or gateways should be a growing trend.  Selling products for a few dollars that do just one thing seems a better bargain than paying a lot for something that doesn't do everything I want.  Text messaging is a huge, huge opportunity with both popularity and low cost. I've been saying that since Symonty hung cell phones to the walls at Tenzing in 2003 for in-seat SMS messaging using our special international SIM card set, in partnership with Panasonic, ARINC and Virgin Atlantic Airways.

It's passenger services that's powering the deployment of broadband COTS radio systems.  While operational enhancements affecting everyone have made-do with ACARS and baby-steps,  In-Flight Entertainment (IFE) runs in many directions at once, with each airline free to leap towards discriminating features.  Communications are powering all the latest features, especially considering the trend to embrace passenger-owned devices, in-seat power, and convenient wireless access.

Television broadcasting requires a wide beam service, either re-using public Direct Broadcast Satellites (DBS) or leased transponders broadcasting a private set of channels.  The airplane has to have an appropriate receiver, which in the DBS case is dedicated and the the leased case usually shared with the data link networking service.  It is possible to broadcast in a spot beam scenario, but the information uses bandwidth within each spot beam undermining frequency reuse benefits.

Live television streaming can waste spectrum whenever content is repeated.  Headline news stories cycle continuously.  Sporting events have many breaks as well.  Aviation would be much better served by caching broadcast vignettes and composing linear feeds or on-demand programming, which then leads to customized advertising opportunities.

Content licensing is always nightmare to reckon with.  The problem is fragmenting as individual channels present themselves as independent properties.

My preference is to have the programming cater to the time I can watch in my seat, rather than a matter of chance what programming is playing, along with the waste of mass-market advertising.

AeroMobile pioneered the use of the cellular air-interface by flying a mobile cellular base station back in 2004.  While the utility and ease of cellular roaming is laudable, it seems a WiFi network has greatest utility in comparison.  Thankfully, cellular providers are favoring WiFi as an alternative to connect cellular services.  No objection to cellular air-interface as a convenient feature in any case.

Peter Lemme

Copyright 2015
All Rights Reserved.