International Civil Aviation Organization
INFORMATION PAPER / ACP-WGI09/IP-02
2008-10-09
Montreal, Canada 20-23 October 2008
Aviation Questionnaire for IETF IPv6 Mobility Activities
Prepared by:
Christian Bauer (DLR, )
Serkan Ayaz (DLR, )
Presented by:
Robert P. Dimond (Verizon)
SUMMARYThere is an effort in the Mobility Extensions for IPv6 (MEXT) working group within the Internet Engineering Task Force (IETF) to work on a solution on Route Optimization for Network Mobility specifically for the aeronautical industry that might become the cornerstone of the future mobility solution. During this effort, several questions emerged that should be answered by as many members of the aviation industry as possible in order to guarantee the development of a mobility protocol that is suitable for the future communications system. This information paper presents these questions with the intention of soliciting as much input as possible from knowledgeable members of the aviation community such as members of ICAO ACP WGI.
1Introduction
The adoption of IP for aeronautical communications is becoming a reality. ICAO ACP WG-I is currently in the progress of finalizing Document 9896 “Manual for the ATN using IPS Standards and Protocols”. This new ATN manual is based on IPv6 for future air-ground communications.
Several parties are working on IP(v6) for the future communications system, both Europe (e.g. within the NEWSKY project led by the German Aerospace Center DLR), and in the United States (e.g. in FAA and NASA projects). Currently, there is an effort in the Mobility Extensions for IPv6 (MEXT) working group within the Internet Engineering Task Force (IETF) to work on a solution on Route Optimization for Network Mobility specifically for the aeronautical industry that might become the cornerstone of the future mobility solution.
During this effort, several questions emerged that should be answered by as many members of the aviation industry as possible in order to guarantee the development of a mobility protocol that is suitable for the future communications system.
The questions are separated per audience: The first set is relevant to all parties, whereas later on questions are specifically directed at the Aeronautical Communication Service Providers (ACSP), Air Navigation Service Providers (ANSP), Airlines and Airframe manufactures and finally Airport operators.
2Questions Relevant to All Entities (ACSP, ANSP, Airline, Airport)
2.1General Questions
2.1.1Would it be acceptable in case of Home Agent failure that the Aircraft and the Correspondent Node on the ground can not communicate with each other anymore?
Explanation: As shown in the figure above, within Mobile IPv6 (MIPv6) an entity called Home Agent (HA) plays a pivotal role as the airborne router send all its traffic via the HA. In the Network Mobility (NEMO) basic extensions to MIPv6, even if aircraft and Correspondent Node are within the same network (i.e. within the ANSP operated air-ground network as shown in the figure) a HA failure would prevent communication between the end systems. Please note that in this figure we are assuming that the HA is operated by the ACSP.
2.1.2In case the previous question is answered with NO, would it be acceptable for you (most notably ANSPs and airlines) to deploy additional infrastructure that allows communication in case of Home Agent failure. This infrastructure has to be deployed in the same network as the Correspondent Node or at least close to the Correspondent Node. (Please note that “new infrastructure” does not necessarily imply new devices, but instead e.g. existing routers could receive a software upgrade)
2.2Technical Questions
2.2.1Are you expecting that the ATN will be directly addressable and routable from the Internet? (Please note that this question is not related to using the Internet as a carrier but about the ATN being reachable from the public Internet as perhaps only a stub network. In more technical words, will ATN prefixes/subnets be advertised to the Internet?) If this is not currently planned, but is considered a possibility within a decade or two, then the answer should be “YES”, as this affects several factors.
2.2.2As a general question, what type of IP addressing should be considered for the future ATN? Provider Aggregatable (PA) or Provider Independent (PI) addresses? PA addresses cannot be transferred between service providers whereas PI addresses are not associated with any provider and can therefore be preserved when an airline changes its communications service provider. PI addresses can also permit more effective multihoming with a single PI prefix reachable through different service providers, rather than all nodes having multiple PA prefixes (one per provider). Hence, PA addresses are more restrictive than PI addresses.
2.2.3How restrictive are your current firewall settings? E.g., are you blocking all ICMP messages?
3ACSP Specific Questions
3.1General Questions
3.1.1Who is usually operating the Gatelink system at the airports? Is it under administrative control of the airport company or of a communication service provider s.a. SITA or ARINC?
3.1.2What is the current plan and progress of bringing CertiPath to the ANSPs? Is there current work in progress such that the national ANSPs will accept certificates issued by another aeronautical entity?
3.1.3Would it be reasonable for you to deploy multiple, distributed HAs in your network?
With multiple Home Agents (HA), the network will deliver packets between aircraft and end nodes on the ground with smaller delay and in addition, it will provide resilience against the HA as a single point of failure. Only a small number of HAs would be needed (for instance, two per continent could suffice).
3.2Technical Questions
3.2.1Gatelink already makes use of X.509 certificates. Which version of X.509 is used?
- Also, which Certificate Authority (CA) signs the certificates?
- Who is operating the CA?
- Are the certificates used at the 802.11 layer, for End2End authentication or for both?
3.2.2Secure ACARS already uses PKI and certificates. Is it correct to say that itis only used for End2End (between aircraft and airline ground end-system) or is it also implemented for layer-2 authentication (aircraft radio – base station)?
3.2.3Are you doing hot or cold potato routing within your network?
Hot-potato routing is the practice of passing traffic off to another autonomous system (network operated by a different entity) as quickly as possible. Cold-potato routing is the opposite – a packet is kept within the originating autonomous system as long as possible when routing it to its destination.
3.2.4Are SITA and ARINC peering with each other? Does this mean that traffic within the ARINC network can be directly sent to the SITA network over a boundary router if the destination is from SITA’s address space? Is this applicable to both ATN/OSI and IP networks?
In addition, is it already the case or foreseeable that the network of an ACSP can also be used only for the purpose of transiting ATS or AOC traffic?
4ANSP Specific Questions
4.1General Questions
4.1.1What does the operational concept for UAVs look like? Is it valid to handle them the same way as a “normal” aircraft? What are the expectations of who will operate UAVs? Are there any projections for the number of UAVs that have to be supported in the future, e.g. in the 2020 timeframe? What distance will UAVs probably traverse? Are UAVs expected to contain an entire onboard network or simply a single IP-addressable host?
4.2Technical Questions
4.2.1How is the ANSP network connected to the data networks at the airport? How does the routing path look like?
4.2.2Would you consider deploying an entity called “Correspondent Router” within your network with which an aircraft can perform mobility related signalling?
The advantage is that packets do not have to be routed via the Home Agent (cf. Question 2.1) but can take a direct path to the ATS node within the ANSP network. In addition in case of Home Agent failure, communication between aircraft and ATS nodes on the ground can still take place.
5Airline / Airframe Manufacturer Specific Questions
5.1General Questions
5.1.1How many end-systems on an aircraft do you usually have at the moment and are you expecting for the future for each of the following domains? ATS, AOC, AAC, In-flight entertainment
5.1.2Would airlines consider operating Home Agents themselves or would you expect letting SITA or ARINC handle this aspect? Might it make sense to answer this question based on an airline being either small or large?
5.1.3With how many different end systems on the ground will an aircraft usually communicate with during a flight? Where are these on-ground end systems located?
5.2Technical Questions
5.2.1How many subnetworks will you expect to have on an aircraft in the future? Will a subnet serve several domains (e.g. ATS+AOC+AAC) or are you expecting to have separated subnets (e.g. 1 for ATS, 1 for AOC, etc.)?
5.2.2When an airline changes its ACSP, would it be acceptable to also change the IP addresses/prefixes/subnets of the aircraft or should those be preserved?
5.2.3Should a strict separation between cockpit and Airline Administrative Control/Passenger Communications be assumed or is it reasonable to assume that a single airborne router handles both traffic categories?
Using a single airborne router for all communications - cockpit and passengers - could be considered problematic due to the increased vulnerability, most notably from the Internet.
6Airport Specific Questions
6.1General Questions
Who is usually operating the Gatelink system at the airports? Is it under administrative control of the airport company or of a communication service provider s.a. SITA or ARINC?
6.2Technical Questions
How is the Gatelink system interconnected to the rest of the ATN? How does the routing path from the end system on board the aircraft to the ground based end-system look like?
7Summary
The authors request responses to the questions presented in this paper be addressed to teh following people:
Serkan Ayaz (DLR) -
Christian Bauer (DLR) -
Wesley Eddy (Verizon) -
Seite 1