Datalink: Sorting Fact from Fiction

Kim O'Neil
Advanced Aviation Technology Ltd.
March 2001

Abstract

The ATM industry appears to be facing a dilemma: "official" support from some National and International Administrations appears to be driving States towards VDL mode 2 with a hoped for upgrade to VDLmode3. Yet this (effectively political) support for VDL mode 2 is without either a feasible, convincing, technical or economic argument. Comparison of VDL modes 2, 3 and 4 clearly indicate that VDL mode 4 holds the greatest promise with the widest range of applications - and has the actual benefit of being a demonstrable and working data-link.

This paper discusses some of the issues surrounding the introduction of datalink for Controller Pilot Datalink Communications (CPDLC). A copy of the accompanying PowerPoint presentation can be found here.

1. Introduction

Real confusion surrounds the practical issue of implementing data-link services.
First issue: - who is responsible for selecting from the datalink options?
Is this a decision made by ICAO or the Civil Aviation Administrations, Telecom providers, airlines or is it someone else? What roles do bodies such as ECAC or Eurocontrol play in this decision? Is there a "decision process" at all? Or is it based on some kind of arbitrary consensus? Can an airline, State or region "go it alone"? On what basis is the decision made - is it Safety, Cost, benefits, near term or long term functionality, commercial interest, spectrum efficiency - all of these or some other parameter? Many of these question don't have easy answers.

However, it doesn't take very long to realize that there is much at stake in the choice of datalink, that it is not just some kind of altruistic decision-making process. It is, in fact, a highly charged, commercial and heavily politicized process. The choice of data-link will have "knock-on" effects in many other areas. Those committed to existing technologies have a lot to lose and are not keen to open the door to new players.

For example, although VDL2 and VDL 3 share a common modulation scheme (D8PSK) there is no "upgrade path" from VDL2 to VDL3. From an operational perspective, they are, in fact, competing and incompatible technologies. This fact may surprise many. In particular, VDL2 does not fit into FAA's "Nexcom" programme (transition to digital voice).

2. Safety

In any complex debate, it is always wise to clear the table. So, is safety an issue? The answer is, of course, that safety is always an issue. There are few areas that are not affected at some time by questions of safety - if only because public concern is high and the airline business can be radically affected if safety is brought into question. That said, the issue of safety is strongly tied to applications: what the technology is used for determines how 'safety' can be affected.

Data-links transport data from one point to another point (or many other points). Key safety issues might typically be: the integrity of the link, its reliability, transport delays and security issues (such as authentication). Which of these properties are significant really depends on the application. The application also determines the upper and lower bounds of these parameters. In practice, most applications are purely commercial in nature - and these issues often result in a trade of cost against performance of the link. For example, security can be a major consideration in many data-link applications (and encompasses many parameters). However, security protocols tend to significantly reduce the efficiency of a data-link and add to its cost.

However, aviation is a highly regulated industry. All aspects of the aviation business are subject to regulatory oversight and certification: equipment, people and applications. Where safety is an issue, we have put in place regulatory structures to ensure that operational applications are safe (and that the chosen technology is sufficient to the purpose). In short: if it isn't safe, it won't fly.

Equipment may be installed in aircraft and 'certified' - but this does not mean that such equipment can be used for any purpose. An operational application must be certified as a system. It is at this point that issues such as message delay times become significant. For example, if the performance of the data-link cannot meet the needs of time critical communications, then it won't be certified for that purpose. Operational procedures must also be put in place and the approval of these procedures for a given purpose determines whether and/or how a technology really gets to be used in anger.

Thus the safety question is reduced to its essentials: can it be certified for the purpose for which it was intended? For if it isn't certified, it certainly won't fly - and many airlines know from bitter experience that the road to regulatory approval is often long and hard.

• Success depends upon developing careful strategies to avoid costly failures.

3. Key Decision Factors

The key decision factors on the choice of data-link depend on who you are and what your interests are. ATC providers, Airlines, Telecom providers and manufacturers have very different outlooks and objectives. Few of which are compatible. Everyone wants to sell their services (including airlines who are also owners of at least one service provider). Everyone has their own priorities. Given that there are many factors related to the choice of datalink, it is not surprising that industry consensus is much harder to achieve than might first appear.

3.1 Committment

Yet it is clear that some decisions have apparently been made: several Civil Aviation Administrations have indicated that they will not be installing VDL mode 2. Instead, they prefer to encourage communications service providers to so. Contracts have already been signed between some Civil Aviation Administrations and communication service providers - giving mutual access to their respective networks. This is apparently a "short term" measure to provide air/ground communication services until VDL mode 3 is in place (which is a major assumption in itself……). Of course, CAAs will also encourage airlines to equip - resulting in a disproportionate burden for airlines. Conversely, other administrations have made quite different commitments to VDL mode 4 (e.g. Russia, Sweden etc.).

3.2 Vested Interests

The dominant communication providers have their own interests to protect. They supply services via ACARS, HF and satellite communications. They would like to increase their capacity as the business grows - and business is growing. They also need to protect ACARS until it is replaced by a high-speed link. In fact, it is relatively difficult to increase current business based on ACARS due to the saturation of the ACARS network (which will generate significant customer dissatisfaction as time goes by…..). Naturally enough, the current communication service providers would also like to maintain their effective monopoly of services (which is essentially based on ownership of frequencies and their proprietary networks). New frequencies will allow their business to develop and expand. The transition from ACARS to VDL also requires new frequencies. In fact, this is their true bottom line: more frequencies.

3.3 Costs

There seems little prospect that costs and services provided by the current service providers will match the economies made possible by the Internet (although the Internet backbone will become a major carrier of operational communications). Yet significant reductions in operating costs should be a key decision factor.

3.4 Airline Operational Communications (AOC)

However, ATS providers and communication service providers need the cooperation of the airlines. Airlines must be convinced of the Cost/Benefits of datalink before equipping their aircraft with any technology. This has, for example, lead to the formation of the C/AFT team - a Boeing sponsored Airline grouping examining the cost/benefits of ATC/AOC applications. The general conclusions of this group have been that AOC pays now, ATC pays later………. Although there has been a recent surge in belief that CPDLC offers substantial capacity 'nearer term' benefits (although when exactly……….). One significant criticism of C/AFT is that it made a premature committment to a particular datalink without analysing the consequences........

Other Key Factors are:

3.5 Functionality

As always, some want to solve today's problems whilst others have a more strategic view. The question is: "how far into the future do I need to look"? and "how much of what is promised is real"? Conversely, it would not be wise to spend a lot of money on a quick fix only to discover that the chosen solution was a "dead end" - requiring even more resources to upgrade or replace later (with much disruption etc.). Hence operators need to consider current and future requirements over a representative period of operation, say 10 to 15 years. Over such a time period, the datalink chosen must be highly adaptable over a wide range of application.

3.6 Spectrum Efficiency and Spectrum Planning

This is a key and unavoidable issue, as the aeronautical spectrum is crowded and under pressure from external sources. Whatever solution is chosen, spectrum will be required. The datalink service must be able to co-exist with existing users. In the VHF band, this means the datalink must be able to fit into 25kHz frequency allocations without interference to other users (who are mostly AM double side band voice). It must facilitate frequency planning and provide high capacity - to obtain the highest return per frequency allocation. Strongly associated with spectrum efficiency is the International process of agreeing frequency allocations. This is a slow process. Existing services cannot be simply displaced to accommodate new services. To offer a datalink solution, spectrum must be made available across the whole of the operational volume of the data-link service. This requires the cooperation of the National spectrum managers.

3.7 Transition

Without an effective transition strategy, any datalink, however good, will be doomed. The datalink must also be able to co-exist for a very long period with existing systems. There will be NO quick or sudden transition. Even with a very co-ordinated transition (which is not very likely either!) it will take a long period to put in place the ground infrastructure required (the dates for which, keep going back….) - and equip sufficient aircraft to develop viable operational applications.

Any data-link, which requires exceptional frequency protection measures such as guard bands, is fundamentally unsuited for CPDLC operations. Further frequencies will, of course, be required if CPDLC services expand. Unfortunately, as currently specified, each VDL mode 2 frequency assignment requires two guard band frequencies, meaning that five frequencies are required to provide a single VDL mode 2 channel! The relatively slow rate of introduction of CPDLC services and benefits, also means that voice back up will also be required for many years. CPDLC services may help reduce voice traffic, but will not help to release VHF voice frequencies.

• Consequently, VDL mode 2 will greatly increase pressure on the VHF spectrum.

3.8 Cost/Benefits

Cost/benefits are a most sensitive issue for airlines as they bear most of the cost in direct and indirect charges: equipment, installation, certification, maintenance and operating fees. ATS providers bear a much smaller financial burden. Communication providers have some incentive to invest or risk losing their commercial position.

Assessing the true cost/benefits means looking at an investment across it entire operational life - which means looking a long way into the future. The problem here, is the difficulty of separating fact from fiction. Many things are promised - but history teaches us that relatively few things live up to the claims made for them……..

To make sense of what is possible, we first need to compare what is on offer. From this we can see what each link can and cannot do - we can then begin to assess each link against our future needs - calculate the costs and guestimate the benefits.

4. Comparison of Datalink Options

A comparison of the different data-link options is given below:

Figure 1: A review of ICAO Mobile VHF Digital Communications Systems

Notes:

1) VDL Mode 2 does not support air-air communications. Ground stations required for all VDL mode3 communications including air-air (as VDL mode 3 groundstations provide timebase). VDL Mode 4 provides air-air communications without need for ground infrastructure.

2) VDL Mode 4 incorporates the use of two global signalling channels for a full surveillance infrastructure plus local channels as required and it can also re-use the timeslots of distant users under high traffic loads (benefiting from 'capture effect' of FM). This gives redundancy and graceful degradation. VDL Mode 4 is also robust to the failure of the standard network time source (i.e. GLONASS or GPS). If primary timing is lost, timing can be obtained from an internal oscillator, another user, from a clock in the ground network or from another satellite (e.g. television satellites) connected to the ground network.

3) A potential weakness of VDL Mode 3 is the shared voice and data spectrum. A stuck transmitter could prevent both voice and data communications. This could be overcome with suitable implementation. Doubts have also been raised over robustness of the proposed modulation scheme (D8PSK) for both VDL Mode 2 & 3.

4) Channel time is wasted during simultaneous transmissions and in the gaps between random messages.

5) VDL Mode 2 and 3 communications are lost if the ground station fails. Note that both air-air VDL mode 3 and air-ground VDL mode 2 use a 'contention access' protocol with similar behaviour to the Ethernet LAN protocol. This has relatively low capacity (only 20-30%) of channel bandwidth can be used and the link fails catastrophically if overloaded, although VDL Mode 2 includes some measures to try and prevent catastrophic failure. At high capacity loading the access delay becomes very long and unpredictable. A CSMA protocol is therefore fundamentally unsuited to the delivery of time-critical data, e.g. differential corrections, CPDLC, TIS-B or ADS-B data, except in very low traffic densities.

6) Both VDL Modes 2 and 3 have lower capacity than VDL Mode 4. VDL Mode 2 capacity is lower because is based on a 'contention access' protocol, which is inefficient under heavy traffic loads. VDL Mode 3 capacity is lower because it does not re-use spectrum as efficiently. D8PSK offers a CCI value of approx. ~26-27 dB. VDL Mode 4 uses GFSK with a CCI value of approx. ~10 dB, and it automatically re-uses timeslots. Over a wide geographic area, VDL Mode 4 is a significantly more efficient user of the spectrum, particularly for ADS-B, TIS-B, CPDLC and DGNSS applications.

7) There are concerns about the flexibility of voice/data channel allocations in VDL Mode3. VDL Mode 3 may be unable to re-distribute channels between voice and data to adapt to requirement changes as data becomes more frequently used. This could cause particular problems during transition as data link traffic grows. A particular problem for VDL mode 3 is how to transition from analogue to digital voice given the spectrum constraints (both analog and digital channels will be required during the transition period - which could be many years).

5. Conclusions

It is not the task of ICAO to select or mandate a particular standard but rather to agree on common standards. The 96 articles of the Chicago Convention establish the privileges and restrictions of all Contracting States, and provide for the adoption of International Standards and Recommended Practices (SARPs) regulating air navigation, recommend the installation of navigation facilities by Contracting States and suggest the facilitation of air transport by the reduction of customs and immigration formalities. The Convention accepts the principle that every State has complete and exclusive sovereignty over the airspace above its territory and provides that no scheduled international air service may operate over or into the territory of a Contracting State without its previous consent.

Consequently, it is up to the individual member states to decide on their preferred implementation (hopefully, in agreement with the majority of other member States). It is not for one or two States to decide on what others shall or shall not do.

VDL Mode 3 hopes to integrate voice and data communications. However, this could reduce robustness, since failure of the VDL Mode 3 unit would result in the loss of both voice and data communications. Also, integrating voice and data may not bring cost savings - given that VHF R/T transceivers are inexpensive and significant redundancy would have to be included. VDL Mode 3 also has significant transit delay, transition and flexibility problems, which need much further investigation.

Only VDL Mode 4 appears to meet all the broad requirements defined (so far) in various operational forums. VDL Mode 4 also has several advantages over the other data links listed:

• VDL Mode 4 can provide a CDTI and ASAS including air-air data link in remote areas. VDL Mode 3 cannot because it requires master ground stations to operate.
• VDL Mode 4 supports all types of data link applications, including time-critical ones. VDL Mode 2 does not support air-air communications and cannot support time-critical applications except in very low-density areas.
• VDL Modes 2 and 3 have much lower capacity and require more bandwidth than VDL Mode 4.
• VDL Mode 2 has poor degradation characteristics, suffering very rapid failure on overload. VDL Mode 4 exhibits graceful degradation through functions such as slaving on time from other stations, re-use of the timeslots of distant users, etc.
• VDL Mode 4 has an incremental growth capacity through the local addition of 25 kHz channels for gradual increase of capacity, new services and functions. This is made available through a Directory of Services, which also enables autotune functions. All VDL Mode 4 units can exploit this additional capacity.
• VDL Mode 4 offers very high spectrum efficiency and an ADS-B capability with long range, redundancy and robustness.
• Only VDL mode 4 offers the possibility of seamless transition.

Advanced Aviation Technology Ltd.
The Old Post Office,
The Street, Compton,
Surrey GU3 1ED. ENGLAND.
Tel. +44 1483 811 311.
Email: kim.oneil@aatl.net