C/AFT ATS Datalink Focus Group Meeting

March 12, 1998

Everett, WA

Agenda

0900-0910 Welcome and Opening statements David Massy-Greene

0910-0940 ARINC Presentation Aloke Roy

0940-1010 SITA Presentation Vaughn Maiolla

1010-1030 Break

1030-1100 VDL Mode 2 Simulation Report Ken Clark

1100-1130 Mitre/CAASD Datalink Study Report Jim Simpkins

1130-1230 Lunch

1230-1500 General discussion

1500-1600 Develop plan for the April All Airline C/AFT meeting in Paris

Attendance

Summary

C/AFT has a core meeting in Washington DC on April 2nd, and an all-airlinemeeting in Paris on April 29th. David Massy-Greene will take a recommendationto the all-airline meeting for an interim step to ATN. This step will be ACARSover VDL Mode 2 using CLNP and something like TP4. C/AFT will also perform abenefit study to estimate the probability of long-term benefit of ATN.

Action Items

Group: Send corroborating studies as well as any benefit studies that areout there.

Group: Provide electronic copies of presentations to Kathleen

Aloke: Send assumptions that were used for growth predictions

Aloke: See if ARINC can do ACARS over VDL Mode 2 simulation in mixedenvironment

Nigel: Find out what Eurocontrol is doing in regards to frequencymanagement.

Boeing: Determine if we can undertake simulation of ACARS over VDL mode 2

Boeing: Begin development of long-term probabilistic benefit model

Charlotte: Determine if we can publish FAA CPDLC R.D on the C/AFT website.

Introduction David Massy-Greene

We are embarking on an important activity to try to unhook nexus that's beenoccurring in various CAA's around world to get progress in CNS/ATM. Which waydo we go in datalink – what flavor, and which one can we justify? C/AFTlooked at what it takes to put ATN on airplane, we didn't understand what waspushing ATN, in its pure form, to airplane at this stage. What are benefitsthat would pay for investment? Some applications may require level-B router onairplane (current estimates ~ $1/2 million per airplane). Started to look atwhat else was available to enable to go faster into datalink environment, butstill leave progression path to ATN. SITA and ARINC have both made separateproposals to go to VDL Mode 2, allowing bandwidth for AOC and ATC.

What would I like to do today? Develop consensus on path for C/AFT datalinkfocus group – how should we be proceeding in future.

A full cost benefit analysis requires costs and benefits. We don't knowbenefits, have some idea of costs. We can't determine benefits until we have anoperational concept. Free Flight TF3 provides ideas, but doesn't provideoperational concept. We plan on doing a probabilistic cost/benefit analysis, itwill take a while, and will be fairly undefined due to uncertainty inoperational concept/benefits. Would like benefits ultimately in dollars. Datalink is an enabler.

Aloke Roy ARINC proposal for transition to VDL/ATN

Link to Aloke's presentation.

Situation: VHF ACARS traffic is increasing. Multiple frequenciesare required to handle current traffic. Today there are 10 millionmessages/month requiring 4-5 frequencies. As user demand increases, morefrequencies will be required. Growth predictions indicate that up to 12frequencies may be required by 2004. This is not a viable solution. Frequencies are not easy to find – we're not going to get 12 frequencies.

Proposal: Implement a VDL Mode 2 Network which:

• Complies with ICAO VDL and ATN SARPs
• Uses Connection Less Network Protocol (CLNP) address for end-to-end datadelivery
• Supports direct delivery of user data to hosts using CLNP
• Supports optional X.25 Gateway for host connectivity
• Announces VDL service availability using ACARS squitter message
• Will allow simultaneous operation of interim VDL and full ATN basedavionics

ARINC's proposal has the following objectives:

• Provide additional data-link capacity by Year 2000
• Provide backward compatibility with existing applications and minimizeimpact on avionics
• Provide clear transition path to ATN
• Allow interim avionics and full ATN compatible avionics to operate oversame ground infrastructure thereby reducing industry costs
• Allow local routing and delivery of user data
• Provide the capability to add new applications without changing basiccommunications stack
• Gain operational experience with new architecture

ARINC Roadmap to ATN: Mode 1 trials were performed in 1997. GLASTARVDL Mode 2 demo will fly next month (FAA). European trials next year. Proposeindustry interoperability team, even though SARPS are well defined there are alot of options. Interoperability problems could arise, depending on optionschosen. Will have full VDL Mode 2 operational end of 1999 (150 - 200 groundstations per year). ACARS service will live its own life for at least 10 years. Would like to see all new airplanes equipped with VDL in 2000 (Airbus andBoeing). FAA automation in 2005 will allow airlines to take full advantage ofATN end-to-end.

Vaughn Maiolla SITA proposal for Migration to VDL Mode-2

Link to Vaughn's presentation

Situation: SITA agrees with ARINC's assessment of the currentsituation. SITA's philosophy turns around realizable benefits. To get fundingfor new investments, we need almost immediate payback. Most of the benefits ofCNS/ATM are in the future, and are dependent on new applications. No one haslooked at a proper Concept of Operations for CPDLC and ADS. SITA operates 3frequencies in Europe and that's not enough. These clogged arteries keep peoplefrom using datalink.

Proposal: VDL Mode 2 is here today and provides an intermediatestep supporting existing ACARS-based AOC applications during the upgrade ofaircraft fleet and airline ground system. X.25 over VDL is the preferredsolution because it is extremely simple and will eliminate many of today'sproblems due to poor ACARS protocol implementation. Other reasons are:

• ICAO VDL standard was developed to provide an ATN compatible VHF datalinkbut the VDL ISO8208/X.25 protocol can transport data in any format
• Allows same conversion between VDL or VHF ACARS air-ground connections anduser ground system interface
• Avionics can easily switch between VHF ACARS and VDL ground stationsduring the VDL deployment phase.

SITA's proposal has the following objectives:

• To provide performance improvements without an immediate replacement orupgrade of the existing ACARS datalink applications, on the aircraft or at theairline host
• To provide a migration path from ACARS protocol for air-ground datalinkcommunications to protocols specified in the ICAO VDL Mode 2 and AMSS Data-3standards
• To provide ICAO-compliant datalink services which can support any futureATS internetworking protocols.

SITA believes CLNP/TP4 add complexity and risk with no performance benefit,and that CLNP/ TP4 should be validated (on the ground) before going "live"with operational traffic. CLNP/TP4 validation is already a part of ATNvalidation and is task that is better left to ATNSI and EURATN.

VDL Mode 2 Simulation Report Ken Clark

Link to Ken's presentation

A CRDA between the FAA and ARINC was established in June 1996 to develop anintegrated end-to-end VDL/ATN simulation. The simulation explored theperformance and capacity characteristics of VDL Mode 2 system under various ATSand AOC traffic conditions. The intent of the simulation was to determine themaximum number of aircraft that could be serviced within each domain on a singlefrequency and meet the performance requirements.

The ATS requirements for message delivery times were based on operationalrequirements for the Aeronautical Data Link System prepared by the FAA Data LinkOperational Requirements Team (DLORT) in January 1995. The ATS message trafficmodel was based on the message types and message frequency as recorded duringthe controller pilot data link communications benefits studies for terminalairspace conducted by the FAA Technical Center.

ARINC developed an AOC message traffic profile to be used for the simulationincluding PDC and D-ATIS messages, as well as standard OOOI messages, and arebased on actual ACARS traffic measurements. AOC performance was required tomeet current AOC data link performance levels.

The study consisted of three steps, each representing various phases offlight. For each Step a message set was defined. All aircraft in the systemtransmitted the same message set. The number of aircraft were then increased todetermine the maximum capacity of the system, and still meet the performancestandards of DLORT. The model represented Y2000 level traffic.

Step 1. Airport domain. All aircraft located on the ground. Therewas a single ground station.

Step 2. Terminal domain. All aircraft airborne. The traffic profileused included terminal domain messages from the AOC traffic set and ATS messagesused in the FAA Technical Center study. There were 6 ground stations located inthe New York City area.

Results for Steps 1 and 2.

Performance Requirement

· 95% of all messages are required to be delivered within 3 secondswith 150 aircraft.

Result

• Shared ATS/AOC. Performance requirements met with 160 aircraft.
• Dedicated ATS Datalink. Performance requirements met with 250 aircraft.

Step 3. En-route domain. Northeastern United States.

Results for Step 3

Performance Requirement

· 95% of all messages are required to be delivered within 5 secondswith 400 aircraft.

Result

• Shared ATS/AOC. Performance requirements not met (simulation could handle150 aircraft)
• Dedicated ATS Datalink. Performance requirements met with 640 aircraft.

Conclusion

• Simulation shows VDL Mode 2 meets expected performance and capacityrequirements using a shared ATS/AOC data link system in terminal and airportdomains.
• VDL Mode 2 data link system dedicated for ATS applications can supportmore than 640 aircraft in the en-route domain, which is well beyond the DLORTrequirements.

Mitre/CAASD Datalink Study Report Jim Simpkins

Link to Jim's presentation

Jim Simpkins presented the results of a data link simulation study thatlooked at domestic en route operations for sectors that enforce restrictions dueto communication volume saturation for 1994 traffic levels. The premise was thatif the frequency problems could be reduced, controllers could accept morecommunications and reduce the miles-in-trail restrictions. Controllerteamwork/delegation of tasks, with the additional communication capacity, led tomore efficient and higher capacity use of the existing airspace. Results showedthat with data link there was a reduction of the 20 miles-in-trail restrictionto minimum (5 miles) and it resulted in 62% delay reduction for all 48 aircraft.In the arrival sector, average flight time was reduced by 4 minutes and 22miles. Controllers were able to achieve no holding with a 40% increase intraffic. Voice occupancy time was reduced between 78 to 84%. TOC (transfer ofcommunications) accounted for 30% of the messages. The follow-on study performedby CSSI did not have a manned simulation.

General Discussion

Aloke Roy: The SITA proposal is a subset of ARINC's, and SITA willhave to go to what ARINC proposes for ATN. SITA's proposal will solve today'sproblem, but it is not flexible for the future.

David Massy-Greene: What about message assurance? This is animportant issue for NATCA and IFALPA.

Vaughn Maiolla: Message assurance is not embedded in the SITAsolution, but it could be added at the application layer.

Aloke: ARINC's proposal has message assurance in the CLNP.

David: We must be careful in implementing a near-term solutiononly. Is it worth implementing a near-term solution that will take us to 2005but won't be fully implemented until 2003?

David: What thought has been given to frequency management, and howto accomplish?

Aloke: All I can say is that ARINC intends to field 2 frequenciesfrom the get-go. It's easier to put hardware in the field right away, ratherthan update later. There is capability within VDL that allows service providersto detect problems and change frequency. W can implement on-ground frequencyvs. en-route frequency. W can synchronize frequency changeover based on TOC,using input provided by FAA computers. VDL system is being designed to workwith external systems like that.

David: Is Eurocontrol dealing with frequency management?

Nigel Makins: I don't know but I'll find out.

David: Latency numbers in the VDL Mode 2 simulation report -- werethey transmission latencies or end-end?

Aloke: Latency from CMU through RF through ground station, throughroute exiting ARINC network. May need to add FAA / airlines network delay andapplication delay to get total time budget. According to Ron Jones published atATN Panel meeting total 3 ½ seconds for terminal, and around 6 seconds foren-route. This is how we came to latency requirement.

Al Burgemeister: Doesn't account for latencies within airplaneavionics. Jones people allocated 1 sec to avionics, but that is optimistic.

David: Vaughn , how difficult would it be for SITA to implement TP4and additional addressing capability.

Vaughn: It wouldn't be that difficult. We don't favor it becauseof risk at this point, we'd rather do it incrementally.

David: I think TP4 will become a requirement.

Vaughn: TP4 should be transparent to us. I'm not a fan of TP4myself. What does it buy you, look at CPDLC – just add pilotacknowledgment.

David: Adding this to our systems would be expensive.

Kent Loving: So would TP4.

Al: I understood that advantage of CNLP is to address and pointparticular data stream to gateway AFEPS or ATC/Airline host. X.25 doesn't haveequivalent addressing.

Vaughn: The issue you're raising would only be of concern in ATNenvironment where you're totally bit-oriented from end-end.

Aloke: Tentative yes, you can do networking with X.25 just likewith most other protocols. Advantage of CNLP is that it makes X.25 addressingplan transparent. In X.25 every terminal belongs to the same subnetworkingsystem. Each subnetwork will have to do its own address translations. X.25 cando it, but CNLP is already defined. Industry spent 10 years defining thisstandard, why throw it away?

David: With CNLP and TP4 it would be on path to ATN.

Kent: you mentioned datalink users forum. Wouldn't this be a goodplace to bring up the SITA vs. ARINC proposals?

Aloke: Basically yes, there is a split between two options, sowe'll do both on paper, and we'll see what gets implemented. That's as far asit is being planned on being taken so far.

David: ACARS over VDL M2 simulation is becoming very necessary. How long would it take Boeing to do ACARS over VDL Mode 2?

Kent: Don't know.

Kathleen Pirotte: Boeing will take action item to investigate this.

David: We need to have that answer.

Kent: We would need a lot of information from ARINC and SITA to doa good simulation. Can I count on you to give us information?

Vaughn: Yes, that should be OK.

Aloke: Yes, don't see a problem. Maybe we can give you Opnetmodel, but I'll have to check that out.

David: Question for Charlotte. Is there a way we can get aroundprioritization of messages? Charlotte Long: I don't know. It seemsreasonable that with correct latency prioritization should not be necessary.

Aloke: Priority becomes a concern when you're talking about tacticalvs. strategic messages. We won't start out using tactical messages.

Charlotte: We have tried to separate into tactical and strategicsets. Best we could do is Build 1 is strategic and later is tactical. We'rethinking about prioritization, but we don't have an answer yet.

David: Message integrity is not built into the system (Mode 2 orATN). We should retain CRC for message integrity (even if 622 convergencefunction is stripped out). If we take FANS-1 implementation, it isbit-oriented, even though it comes out in characters. C/AFT would recommendthat CRC remains in there.

Aloke: I agree.

Greg Saylor: If you need to go to increased number of frequencies,can you get them?

Aloke: No. if we don't go to VDL we will have problems. In someareas we have 5, but can't get them everywhere. It's going to be harder andharder.

Vaughn: Likewise for Europe. We operate 3 frequencies but theyaren't available across Europe. If demand drives us to 5 frequencies we'd be inreal trouble. We do have a VDL frequency for Europe, so we're OK in thatregard.

David: Where do we go from here?

C/AFT has a core meeting in Washington DC on April 2nd, and an all-airlinemeeting in Paris on April 29th. I'd like to take a recommendation to theall-airline meeting for an interim step to ATN. This step will be ACARS overVDL Mode 2 using CLNP and something like TP4. This is the way I would suggestwe go as an interim step.

Greg: Is there any cost in taking the intermediate step that wouldbe in excess of what we would need to take a longer term step.

David: This would require ARINC 758 CMU, possibly at differentcertification level than short-term implementation. We are still trying to findout what it will cost to get to level B CMU, even though we don't know what theactual requirement will be. I've spoken with both major CMU suppliers, who saythat they are planning to certify to level C, but building to level B. They'vedone all they think they need to do to go to level B certification if necessary. Software tracing has been done, software verification has not.

Aloke: Given what we know today, 90-95% wiill be reusable