Future of CPDLC

Future of CPDLC

57TH ANNUAL CONFERENCE, Accra, Ghana, 19-23 March 2018

Agenda Item: B.5.1 – WP No. 85

Future of CPDLC

Presented by TOC


When properly implemented CPDLC can prove a benefit to controllers in many ways including clarifying communications, reducing radio congestion and potentially freeing up more time to dedicate to high priority tasks. There are some areas that pose a concern, however, such as lack of frequencies and several networks with different protocols (multi stack). As CPDLC has developed there have been several different systems, including disparate systems in oceanic and continental airspace, and there is a lack of a clear plan toward harmonization.


1.1.  The early ideas for CPDLC (Controller Pilot Data Link Communication) resulted from ICAO meetings in 1983. The groundwork was put in place for its development in part due to the increasing frequency saturation on the bands used for ACARS and VHF (Very High Frequency) voice communications. Additional concerns that were considered included controller pilot misunderstandings, blocked radio transmissions, and an ability to communicate with aircraft that are off VHF radio frequency.

1.2.  Differences emerged early on with Boeing-aircraft utilizing FANS-1 (Future Air Navigation System) technology while Airbus-aircraft used FANS-A. Potential challenges were compounded by ANSPs employing different systems of infrastructure. The systems continued evolving and to the now familiar FANS1/A+ and ATN (Aeronautical Telecommunications Network). As new technology emerges, such as FANS-2, FANS-B, and ATN-2, they continue to be developed independently causing difference in their development. This may result in air or ground providers needing to utilize multiple different types of air and ground infrastructure or at least the knowledge that they exist. There continues to be a lack of harmonization which poses challenges on many fronts including air and ground, and network infrastructure.

1.3.  Furthermore, there are matters to contend with such as latency issues, provider and user aborts, and the quality of networks CPDLC relies on. The development of different systems and aircraft equipment has led to a variety of differences in functionality and procedures worldwide. The future holds much promise for CPDLC but there will be some hurdles to overcome to get there.

1.4.  There is a tandem information segment for the IFATCA website on an “Introduction to Data Link”. This material will cover differences in the systems and their implementation in further detail.


2.1.  Two Pertinent definitions from the GOLD Manual:

Controller-pilot data link communications (CPDLC): A means of communication between controller and pilot, using data link for ATC communications.

CPDLC message: Information exchanged between an airborne application and its ground counterpart. A CPDLC message consists of a single message element or a combination of message elements conveyed in a single transmission by the initiator.


2.2.  From early on there have been multiple systems that emerged to provide CPDLC services including FANS-1 and FANS-A; emerged for Boeing and Airbus systems respectively. These systems have some different functionality from ATN, which is the ICAO standard. Some ANSPs employ one, some both and others currently use none (Informational material “Introduction to Data Link” will feature further information on differentiation between types of data link and how they are executed).

2.3.  While technology is improving constantly there are still several issues that must continue to be worked on to achieve optimal results such as latency issues and provider aborts. The proper functioning of all elements of that system will allow for building upon an existing strong foundation.

Latency Issues

2.4.  Latency issues pose a major challenge to overcome, especially in aircraft utilizing FANS. While ATN may present a solution it also has correlated to a rise in aborts. Part of the issue for FANS 1/A aircraft is the while the ground may generate a timer the aircraft does not.

2.4.1  Since the early roll out of data link it was advised that there should be a dedicated frequency for data and a separate one for voice. In the EASA (European Aviation Safety Agency) 2014 report, “Technical issues in the implementation of Regulation” regarding technical issues stated that with a moderate level of traffic the band might be acceptable through the year of 2015 but not through 2020. Unfortunately, data link was rolled out utilizing a shared frequency, instead of a dedicated one, which is susceptible to overloading.

2.4.2  CPDLC in Europe via OSI (Open Systems Interconnection), the current methodology, would benefit from a Common Signalling Channel (CSC) which would make it ICAO compliant. To combine the CSC and data-communication on the same channel/frequency is not without issues, as can be seen by some of the challenges in Europe.

2.4.3  Delayed transmissions are impactful on CPDLC transmissions and are defined as a delay that exceeds the timer maximums. According to the current Performance-Based Communication and Surveillance Manual, ICAO Document 9869, delays in transmission for RCP 240 (Required Communication Performance, RCP, denotes the amount of time a controller would have to handle a situation. In the case of RCP240 that would mean a time of 4 minutes round trip for the communications.) should not exceed 210 seconds for 95.0% of transmissions and not exceed 240 seconds on 99.9%.

2.4.4  As is show in the diagram of Actual Communication Technical Performance (ACTP) from the GOLD Manual there is a discrepancy in some aircraft types, and even in some airlines, in contending with functionality. This chart is depicting aircraft utilizing SATCOM functionality. To utilise data link benefits such as reduced lateral separation minima is a Required Communication Performance 240 (RCP 240) must be complied with. The blue line represents 95% function or 180 seconds, the red denotes 99.9% or 210 seconds. In some cases, where the poor performers can be traced back to a specific type of aircraft in a fleet, it is possible for fluctuations to arise from factors such as different routes with their associated geographic issues and even weather phenomena.

2.4.5  This problem is amplified in congested airspace due to the finite amount of bandwidth available within the frequency. Currently the bulk of usage still remains in voice communications. It is further exacerbated during parts of the year with increased air traffic volume putting stresses on the infrastructure. In the future with projected growth in traffic numbers the problem of congestion remains.

2.4.6 Costs to update data link in Europe alone, based on a EASA (European Aviation Safety Agency) report, are estimated to be 831M€ by 2020 and 979M€ by 2025. The biggest portion of this would be dedicated to airborne investment.

Provider Aborts

2.5 Another hurdle to overcome is the situation of provider aborts. It is important to note that the provider abort issues are in the ATN domain. A provider abort occurs when there is a continued loss of connectivity. Depicted at right is the events that transpire leading up to a provider abort. An ANSP would attempt to send a message and after repeated failure to be received by the aircraft would lead to the provider abort. This is in part due to the safety checks that the ATN aircraft run to ensure timely delivery of the messages. FANS 1/A aircraft may not experience this same phenomenon but lose many of the safeguards that the ATN system provides.

2.5.1  The loss of connectivity is possible at any juncture of CPDLC including ATN and VDL Mode 2. One of the major causal factors of provider aborts is avionics. The problem is spread across many different equipment manufacturers and unit types. There are also noted issues with some ground based systems and infrastructure.

2.5.2  For the below chart from Eurocontrol there is still an unacceptable percentage of provider aborts. Even with the previous drop there is still a climbing rate of aborts as the number of sessions goes up. The purple line details the percentage of provider aborts as contrasted to the blue number of sessions.

2.5.3  VDR (VHF Data Radio) Deafness is the factor that is linked to provider abort in approximately 70% of cases. This occurs when an aircraft is able to continue with downlink of messages but ceases to recognize uplinked messages.

2.5.4  A task force was formed in 2013 to work on a Provider Abort Action plan. By the end of 2014 the task force concluded that there were solutions to radio deafness issues, but they were not yet in widespread use. Another factor that is under investigation is the geographic location of the provider aborts and whether there is any correlation to occurrence.

2.5.5  While FANS aircraft don’t experience provider aborts by name they may also experience connectivity issues. When two criteria are met by receipt of specific messages, it is considered an equivalent event to provider abort. The transmissions that are received to equate to provider abort are the UM159 message, which equates to ‘invalid data’, ‘insufficient data’, or ‘commanded termination’, followed by a DM62 message, which is ‘invalid data’ or ‘insufficient data’.

2.6  In the July 2017 issue of Interpilot Magazine the article “Air Traffic Controllers Asking for CPDLC Log-on” lists some known challenges that controllers contend with regarding CPDLC. There are cases of some aircraft, particularly those using ATN, not being able to log on to the system. Some ANSPs, such as Maastricht UAC and SkyGuide, have had to introduce “whitelist” filtering due to a number of aircraft having “bad” avionics. The resultant of these issues is in some cases provider aborts. This filtering has been causing problems additionally with dual stack airframes unless the pilots are aware of how to specifically select FANS 1/A.

2.7  Some of the challenges that have been faced with CPDLC include the inability to do complex clearances, 4D trajectories or time-based clearances. There is a limit to the number of elements that can simultaneously be transmitted and replied to by a crew in a single message, or even two or three messages.


2.8  Further complications may come from having dual stack, multi stack and multi network

situations. Dual stack is when infrastructure is in place to interact with both ATN B1 and FANS 1/A aircraft. This is seen in several locations around the world including Maastricht (MUAC), Oslo FIR, Shannon UIR and London UIR.

2.8.1  Dual stack also applies when more than one communication protocol exists for a given function. Among ATN systems utilizing both OSI and IPS (Internet Protocol Suite) dual stack allows for the maximum amount of interoperability. The dual stack functionality may be found across a variety of different system manufacturers.

2.8.2  Several different aircraft types utilize dual stack technology including the B787, A350, A380, and some modified B747-800. Utilizing both technologies simultaneously may increase workload in the cockpit. This is compounded by both air and ground infrastructure needing updating if two types are meant to be run. Airlines must start the ball rolling on equipage, meaning adding VDL-2 ATN to FANS 1/A as ANSPs should not be utilizing FANS 1/A in high density continental airspace.

2.8.3 Complications can arise when aircraft are equipped with both FANS and ATN. Those aircraft may experience a loss of connectivity when changing between the two data link systems. Dual stack aircraft have different capabilities depending on which system they log on to. In some cases, such as within Maastricht UAC, unless pilots are made aware of how to select a specific system there could be connection issues.

2.9 CPDLC’s usefulness is dependent on provider quality, especially as the systems are reliant on, technology outside of the ANSPs control, such as satellite communications providers. These outside influences include satellites that transmit the data and the airframe based portion of CPDLC. Different CSPs (Communication Service Providers) are used in different areas, among them ARINC in the US and SITA in Europe.

2.9.1  One example to illustrate the importance of quality assurance is the September 2017 Iridium malfunctions over the Pacific. In the reported incidents, the aircraft had either switched to Inmarsat or become disconnected from Iridium on a preceding leg. The messages that had been sent via Iridium were later received errantly by the aircraft and this discrepancy in times was not detected. It was reported that one aircraft climbed as the old message commanded. This incident resulted in many NOTAMs and a temporary cessation in use of Iridium services until the problem was resolved. These situations should have never arisen as the messages should have had time stamps on them and should have been discarded as a result of the outdated time stamp. These occurrences highlight the hazards of utilizing non-ICAO compliant technologies with their lack of safeguards.

2.9.2  Another consideration is the cybersecurity of the system. The National Business Aviation Association (NBAA) encourages pilots to maintain vigilance and ensure clearance, language, and context make sense. NBAA also pushed for pilots to maintain situational awareness, especially in areas that are more vulnerable to attacks.

2.10 In order to ensure equipage, it may be necessary to have mandates or to furnish incentives or conversely disincentives.

2.10.1  Even during a transitional period all aircraft stand to reap the benefits of early technology adopters, including increased likelihood for request approval. The early adopters themselves see advantages in benefits and improvements of services for both pilots and controllers. Even in cases where companies wait until mandate requirement the aircraft receive potential advantages in controllers having more time to focus on higher priority tasks such as separation. Beyond the basic levels of separation controllers may also find more time to entertain requests that may not have been possible previously.

2.10.2  In some cases, the financial benefits or burden may be the motivating factor for equipage is the ability to utilize critical routes. For example, in the case of R-Lat (Reduced Latitudinal Separation) on the NAT (North Atlantic Tracks) aircraft are mandated to comply by 2017 with FANS 1/A (or equivalent) CPDLC and ADS-C equipage or will lose out on the change to utilized the tracks between FL 350-390. This could result in the non-conforming aircraft flying a circuitous routing requiring far more fuel and even in some cases stops. The result is that the best equipped aircraft receive the best service.


2.11 From the IFALPA vision statement:

Air transport is a global activity and it is essential that flight operations work within a common set of standards and procedures all over the world. It is essential that the resulting system is seamless, with the “right” systems (that is hardware, software and new or enhanced technologies) being used in the “right” way in order that a total air navigation system can deliver a logical, efficient and above all, safe, system.

2.11.1 IFALPA’s statement and goals seem to be in line with the goals of IFATCA proposals. They highlight the importance of seamless system, which sounds to mirror the harmonization hopes by IFATCA. The safety aspect they conclude with is omnipresent at the heart of IFATCA’s goals as well.

2.12 Looking ahead to the future new iterations of existing programs, such as future iterations of FANS and ATN, are in the works. How much they will differ from the current systems both in terms of user interaction and functionality is not yet known. It is important to note that the systems are still being manufactured from different companies thus showing at least for the next round of implementation may still pose a challenge.

2.12.1  FANS 2, is being developed by Boeing while FANS B, sometimes referred to as Link 2000+, is in the works with Airbus. FANS 2/B will utilize PM-CPDLC (Protected Mode CPDLC) under ATN protocol in combination with existing FANS 1/A. While the internal part of the system may differ the impact to the controller should be minimised.

2.12.2  B2 will pave the way for increased integration with SWIM and 4D technologies. CPDLC version 3 replaces versions 1 and 2. To remain compliant, ATN B2, both air and ground, will keep interoperability with ATN B1.

2.12.3  FAA Advisory Circular 20-140C provides a list of supported services included with CPDLC version 3, which is still a far way off:

  • ATC Communications Management (ACM)
  • Clearance Request and Delivery (CRD)
  • ATC Microphone Check (AMC)
  • Departure Clearance (DCL)
  • Data Link Taxi (D-TAXI)
  • Oceanic Clearance Delivery (OCL)
  • 4-Dimensional Trajectory Data Link (4DTRAD)
  • Information Exchange and Reporting (IER)
  • In-Trail Procedure (ITP)
  • Interval Management (IM)
  • Dynamic Required Navigation Performance (DRNP)

2.12.4 As technology keeps evolving at an ever more rapid pace the need for consistent upgrades in infrastructure and equipage will need to maintain that pace in order to maintain functionality.

2.13 ICAO 2016-2030 Global Air Navigation Plan lays out a near and long-term agenda for several issues including CPDLC.

2.13.1 The 2017-2019 Global Air Navigation Plan (GANP) outline of the goals is that safety improves and remains paramount while simultaneously increasing capacity, efficiency, and security. While accomplishing this program should remain economically viable and work to minimize adverse effects on the environment.

2.13.2  Future innovations are broken down into Aviation System Block Upgrades (ASBUs) and their approximate time frames of availability. CPDLC falls within Block 0, the first and foundational segment. ICAO Block 0 is the currently a work in progress with hopeful implementation by this year (2018).

2.13.3  In order for ANSPs to proceed with further blocks and future technologies a solid roll out of Block 0 must be in place. Block 0 utilises currently available technology and is ready for immediate roll out in places where it is not already in use.

2.13.4  Block 1, starting in 2019, brings with it the potential of advanced CPDLC over Baseline 2 utilizing OSI. Subsequently Block 2 brings Advanced CPDLC with utilization of IPS. Along with other technologies these culminate in Block 3, post 2030, and full 4D operations.

2.13.5  Some of the real-world deadlines that have recently transpired or are coming up imminently include:

  • December 2017 for utilization of NAT (North Atlantic Track) requisite CPDLC along with ADS-C for FL 350-390.
  • January 2020 NAT requirements in expanded altitudes inclusive of FL 290-390 and requiring ADS-C
  • February 2020 European mandate FL 285 and above within European airspace

2.13.6  In the longer term, advanced CPDLC and ADS-C over Baseline 2 provision via ATN OSI and then later via ATN IPS will allow for momentum toward fuller 4D applications, though the capacity for anything beyond the most basic 4D applications is limited. This transition over time and is not planned to finish until after 2030.

2.14 The SESAR master plan states that it sought for a proportion of radiotelephony messages to be converted to CPDLC which would allow for flexibility of the radar team in responding, this is especially advantageous in high density airspace or with complex clearances. Moving forward there is a goal to clearly define radar team responsibilities in utilizing CPDLC to avoid confusion. It is vital that throughout use of CPDLC that the controller and crew maintain situational awareness. The choice of when to utilize CPDLC versus radio remains with the controller.

2.15 The United States Federal Aviation Administration (FAA) has a staged roll out plan for Data Link and CPDLC services from now until 2030 and beyond. Currently CPDLC services are limited to tower use in DCL relay but within the next two to three years there is hope that area control facilities in the United States will begin utilizing CPDLC as well.

2.15.1  The FAA’s NextGen technology programs look forward to the increased equipage of CPDLC in conjunction with other technologies, including ADS-C, to allow for reduction in separation standards and more efficient flight paths. The deployment of the blocks, such as CPDLC are critical for roll out of later stages.

2.15.2  Both OSI and IPS are suited for handling dual stack operations. The differences from the OSI system and the incoming IPS system include IPS not subdividing the upper layers of interaction into sublayers like OSI does. This results with session, presentation and application in the same group as opposed to split up. ATN and IPS both offer potential for more bandwidth and data volume.

Current Policies

2.16  There are several different policies that deal with CPDLC and related technology either directly or indirectly. Several of them were created years ago and are outdated with current and future technologies.

2.17  IFATCA Policy on CPDLC, COM 4.8 “CPDLC – Datalink Communications” (TPM 2017 page 3 2 4 10) states:

All implementations of CPDLC must demonstrate full compliance with ICAO ATN SARPs. However, in Oceanic and Remote Regions, where it can be demonstrated that CPDLC implementation improves controller pilot communications, it is recognized that non ATN compliant technologies may be deployed during a transitional phase.

The ICAO ATN SARPs and their progressive development form the definitive basis for any future CPDLC implementation.

In high density ATN CPDLC airspace, FANS aircraft shall be handled via voice R/T for safety reasons.


2.17.1  As the amount of connectivity improves the coverage areas of the world also improve. This is in part due to the increased use of satellite communications in places where VHF was previously the method in use. While the planet is still not at 100 percent coverage, with gaps existing in some polar regions and while transitioning between services, there is certainly an increased amount of coverage.

2.17.2  Radio communications still takes precedence in high density airspace for time-critical scenarios. In high density airspace, the size of sectors relative to the few minutes of time an aircraft takes to transit the sector may result in some functionalities being limited. For example, if a sector takes an aircraft only 4 minutes to transit and an aircraft was in full compliance with regulations there may only be the possibility to issue one CPDLC message. Especially in the short term, CPDLC does not intend to and is not able to eliminate voice communication. FANS is now utilized within some airspace, though primarily remote and oceanic spaces.

2.17.3  IATA’s (International Air Transport Association) stance on CPDLC reaffirms existing IFATCA policy on use of CPDLC in congested airspace:

Support CPDLC as the primary means of communication in oceanic and remote airspace where the quality of voice communications is often poor. At the same time, CPDLC should be considered for implementation in appropriate en-route airspace in order to relieve congestion on voice channels.

2.17.4  As long as data link and other data share the limited amount of space via VDL (Voice Data Link) Mode 2 there is a potential issue.

2.17.5  The policy has further problems in that it is anchored with specifically FANS technology. FANS has been fully implemented in many oceanic FIRs and provided the safety related requirements for it are met it has proven to be a safe and efficient system. This policy may need review or removal in a subsequent work programme.

2.18 From COM 4.5 “RTF Frequency Usage” (TPM 2017 Page 3 2 4 7):

IFATCA recognises the need for and supports the reduction of voice communication workload of controllers. However, simply omitting items without alternative methods of accomplishing essential checks compromises safety.


2.18.1  While CPDLC represents a great step forward toward reduction of congestion on the radio, it is not suitable as a full replacement for existing voice communications. CPDLC functions excellently for routine operations and when there is not a safety critical imminent situation. In those events where time is of the essence, voice communications still represents the preferred choice, an example of such a scenario is time-critical separation.

2.18.2  Seeing as voice communications are intended to be used concurrently that would satisfy the need for an alternative means of communicating.

2.18.3  Due to the nature of CPDLC functionality the policy may be considered for policy review in a subsequent work programme.

2.19 Recommendations from AAS 1.5 “Air – Ground Datalink” (TPM 2017 page 3 2 1 8):

Information regarding the equipage/ non-equipage of Datalink is notified to controllers at the operational position in the appropriate manner.

The number of types of datalink equipage in a single airspace, should be kept to an absolute minimum.

ATC personnel should not be required to distinguish between different levels of datalink equipage in the same airspace.


2.19.1  Due to equipage allowing or disallowing use of some airspace and procedures it is vital for controllers to know the aircraft status for timely planning, though these actions tend to go on in the background outside of the controllers’ sphere of influence.

2.19.2  Mixed mode is a concern moving forward. As controllers are forced to work a variety of different technologies with different needs there can be a rise in complexity. In some cases, such as Maastricht UAC, where a gateway is present to handle multiple systems, runs according to the slowest technology present.

2.19.3  It is unlikely that there will be singular equipage, such as INMARSAT or Iridium within an airspace; to limit to a singular equipage type would reduce the number of aircraft able to utilize the technology in a given airspace.

2.19.4  This policy may need review and amendment in a subsequent programme.

2.20 From AAS1.6 “Datalink Applications – the Use of LACK (Logical Acknowledgement)” (TPM 2017 page 3 2 1 10):

When Air Traffic Services are provided via Aeronautical Telecommunications Network (ATN) Controller Pilot Data Link Communications (CPDLC), the use of LACKs (Logical Acknowledgments) shall be considered mandatory.
When Air Traffic Services are provided via any CPDLC other than ATN, a capability which meets the Operational Requirements for LACKs shall be considered mandatory.


2.20.1  This policy on LACK was put into IFATCA policy in 2001, yet there are still systems that are not in compliance ICAO standards and these policies. FANS 1/A still lacks LACK, though it seems it is not used in its initial form today as aircraft are equipped with FANS 1/A+ now. FANS 1/A+ with PM-CPDLC features the ability to comply.

2.20.2  For 2018 TOC work programme, review of the LACK policy is recommended.

2.20.3  ICAO standards from Annex 10, section 8:

8.2.1 In all communications the highest standard of discipline shall be observed at all times. Recommendation— Consequences of human performance, which could affect the accurate reception and comprehension of messages, should be taken into consideration when composing a message. Note— Guidance material on human performance can be found in the Human Factors Training Manual (Doc 9683) and Human Factors Guidelines for Air Traffic Management (ATM) Systems (Doc 9758).

8.2.2 Ground and airborne systems shall provide controllers and pilots with the capability to review and validate any operational messages they send.

8.2.3 Ground and airborne systems shall provide controllers and pilots with the capability to review, validate and when applicable, acknowledge any operational messages they receive.

8.2.4 The controller shall be provided with the capability to respond to messages, including emergencies, to issue clearances, instructions and advisories, and to request and provide information, as appropriate.

8.2.5 The pilot shall be provided with the capability to respond to messages, to request clearances and information, to report information, and to declare or cancel an emergency.

8.2.6 The pilot and the controller shall be provided with the capability to exchange messages which include standard message elements, free text message elements or a combination of both.

8.2.7 Unless specified by the appropriate ATS Authority, voice read-back of CPDLC messages shall not be required.


2.20.4  In order to move forward there is a need for harmonization; there is currently no global solution for this dilemma. The currently planned development to achieve this is the future of ATN/IPS, but this will take several years to start to come to fruition.

2.20.5 Mandates, such as timelines for compliance, alleviate potential controller confusion. But even with mandates older aircraft requiring retrofitting and general aviation aircraft may pose problems. The possibility of handling the aircraft via voice makes for a much smoother transition than forcing a dual stack equipage situation. In some cases, there may be exemptions to equipage for aircraft nearing the end of their lifespan.

2.20.6 Due to the nature of having technology based instead of outcome based policy there may be need for policy review in a future programme.


3.1 CPDLC provides many potential benefits to controllers including reduced radio congestion, less confusion in communications, and potentially more time for controllers to work on higher priority tasks such as separation.

3.2 In the end, after the resolution of technical issues the hope is a more safe and efficient airspace system utilising technology. Many of the positives that controllers would experience would be similar to pilots, including less confusion in communication as well as getting the data sent via CPDLC giving less opportunity to error when getting lengthy rerouting. Users could see benefits in utilising airspace where CPDLC is required and potentially in more opportune reroutes around weather as technology improves.

3.3 In order for CPDLC to move into the future reliability of the systems contracted out must be of sufficient levels for use, such as satellite providers.

3.4 Issues currently impacting the CPDLC system must have long term solutions developed and deployed. These issues include both on the ground with infrastructure and in the aircraft systems.

3.5 Harmonisation of CPDLC is a vital element moving forward. If technologies diverge further it will become increasingly difficult for both controllers and pilots to use.


4.1 It is recommended that:

IFATCA Policy is:

IFATCA supports efforts to define global safety and performance requirements for data link services in order to:

  • achieve harmonization;
  • support further implementation to improve safety and efficiency.

And is included in the IFATCA Technical and Professional Manual.


Last Update: October 1, 2020  

November 20, 2019   420   Jean-Francois Lepage    2018    

Comments are closed.

  • Search Knowledgebase