 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
Executive Summary
Nearly 100 IGS colleagues gathered November 2 through 5, 1998 to attend a workshop dedicated to the infrastructure of our collaborative international organization. The IGS has been an operational service for nearly five years, providing GPS data and products from a globally distributed network of high precision GPS receivers to our internal users, especially the Analysis Centers, as well as to numerous external users. Since the inception of the IGS, we have not sponsored a workshop of this type, focusing on the network issues and how future planning of the IGS affect and are critically dependent on the foundation of the network. This workshop provided a great opportunity for people working within the various components of the IGS to meet and discuss current configurations, problems and resolutions. A good deal of time was devoted to understanding the many future requirements and how to begin incorporating these into our existing infrastructure. One of the key shifts in the IGS network is the realization of emerging application networks. For example, a subset of the IGS network is used for ITRF, a different subset has the characteristics to support precise time transfer, and yet another subset may meet the requirements defined for support to the low Earth orbiting (LEO) missions. This is a theme that was raised in nearly each position paper.
The final day of the workshop, November 6, was devoted to a summary session for the program committee to pull together the recommendations of the workshop. These recommendations were reviewed by the IGS Governing Board at its December meeting and approval was sought for the appropriate recommendations. The proceedings of the workshop are presented in this document and are also available on the CBIS web site. The final result of the workshop is intended to become a "Network Operations Plan" for the IGS Network System.
In brief, the workshop was considered to be productive and engaging, both in the workshop meetings as well as during the discussions, breaks, social receptions, and dinners. Carey Noll put a tremendous amount of effort into the local organization of the workshop and it was a total success. We are all indebted to her for her dedication and positive attitude in everything she does. Thanks are also due to her staff for making the whole event possible.
A more detailed day by day summary follows.
Day 1 - Monday, November 2
The first day was devoted to bringing the attendees together to focus on the status of the IGS, current performance issues, and future requirements. The welcome was given by Dr. Clark Wilson from NASA headquarters, the funding sponsor of the workshop, and followed by Dr. Vince Salmonson, Director of the Earth Sciences Directorate at GSFC, which includes the local organizing institution of the workshop, the CDDIS.
Prof. Gerhard Beutler provided an excellent keynote summary of the 'State of the IGS', stressing the important contribution of the improving and expanding IGS network, as well as the wealth of information in past IGS data. He pointed out that the IGS Network is the combining element of all space geodetic networks. He also showed the importance of complete and correct station information for daily analyses as well as for historical data in the very likely event of future reprocessing of the data sets.
The workshop was fortunate to have many representatives present from the Analysis Centers (ACs), and it was exceptional that the Analysis Center Coordinator, Jan Kouba could personally attend and participate in the whole workshop. The link between the network and the ACs must be strengthened and this workshop was a good attempt in promoting stronger connections. The second key talk, 'Current Network Performance', was given by Dr. James Zumberge, with co-authors Kouba, Tim Springer, and Werner Gurtner, was therefore extremely relevant. Some of the key points and recommendations that stem form this position paper were to highlight previous AC discussions and conclusions in regard to discrepancy resolution, improved and consistent naming conventions, and the need for clear guidelines at all levels of IGS operations. The need for instructions on how to make and document hardware changes to ensure the continuity of station solutions by the ACs was emphasized. The usefulness of network monitoring tools, e.g., the IGS network analysis routines, was acknowledged and more detailed metrics on data noise was suggested both here and again in subsequent discussions as a future improvement.
After a break, presentations on the future of GPS and as well future requirements of the IGS followed. We were fortunate to have David Minkel, Deputy Director of the NGS, to speak to us on 'GPS Modernization', the future GPS space systems. Minkel is very involved in civilian side of planning the dual-use enhancements to the future GPS space systems. He noted that the third frequency is converging on the region of 1181 MHz and that Block II satellites are lasting longer than expected, which in fact may delay the implementation of the new satellites with the third frequencies by a few years. He also stated that the U.S. military is now considering a new and improved signal structure and possibly leaving the original (less precise/robust P/Y code) to civilians.
The Future Requirements session was comprised of an excellent panel speaking for the future needs, requirements, and optimization of the IGS. Overviews were given by the IGS projects for tropospheric, ionospheric, timing, and Low Earth Orbiter (LEO) applications. Additional presentations included an overview of the IGEX (International GLONASS EXperiment) and its current status by Jim Slater. Seth Stein, the newly appointed Scientific Director of UNAVCO, gave an update on plate tectonics and noted that the IGS is responsible for providing the global framework that so much of the science depends on. A subsequent presentation by Frank Webb, Chair of the SCIGN GPS Board underscored the rapid densification of regional arrays and began to address how the IGS can develop an interface with these regional networks. Real-time application perspectives and communication issues were pointed out by Mike Whitehead of Satloc Corporation.
Day 2 - Tuesday, November 3
On Tuesday morning, the Physical Site Specifications and Communication and Data Links sessions were scheduled. In the first session, chaired by Werner Gurtner, presentations were made on monumentation, site selection, multipath detection, status/updates of Hatanaka compression, and two antenna-related presentations. It was noted physical monumentation and its' description in the site log could be improved. A number of global IGS stations are roof top installations, and even for properly (geodetically) monumented points, the monument-related site log entries are often rather sparse and inadequate. The site specification session ended with a consensus that a unique and consistent file naming convention must be developed and adhered to for site logs, station data files (RINEX), and station solutions files (SINEX). All of these conventions should be based on a unique (and officially adopted) four-character station name identifier. There was some discussion of the proposal forwarded by the ACs and it was recommended that the data centers and interested network people respond to this scheme to resolve any remaining issues, by proposing alternative approaches. Werner Gurtner agreed to handle the discussion on these issues by email and compile the suggestions with the goal to reach some resolution in the very near future.
The next session was devoted to 'Benchmarking' the IGS Network. It has been noted that the IGS communication paths are not as efficient as they could be. It was pointed out that ACs must shop frequently at many data centers (DCs) to retrieve data fast enough for generation of the rapid products. To this end the Central Bureau and the Infrastructure Committee, with technical assistance by UNAVCO, have been working on a questionnaire to map the communications and data flow of the IGS network. By understanding the characteristics of the current network configuration, we can work to develop more efficient paths and methods to support the ACs. The consensus seems to be developing that data 'pushes', from the lowest levels all the way up to DCs and ACs may be a more efficient approach to data distribution.
In the afternoon the newly designed and expanded IGS CBIS web site was demonstrated. It was agreed that this site is a very useful improvement over the original site. A few minor improvements were suggested and everyone is encouraged to explore the site and provide comments and suggestions.
A long poster session was held in the afternoon with a large block of time set aside for people to discuss posters on an individual basis. A reception with great food and drink was hosted by the people at Ashtech; many thanks to Robert Snow.
Day 3 - Wednesday, November 4
Data center issues began the day and a number of issues were raised, such as SIO's concentrated effort to clean up past records, site information, etc. and the Seamless Archive approach. However, there was considerable discussion on the list of recommendations from the ACs on how to deal with inconsistent, non-compliant, or 'bad', poor quality data. The ACs want to exclude such data from their analyses, as it corrupts the solution and is difficult for them to deal with. The DCs, with Carey Noll and Jeff Dean in the lead, agreed to investigate the situation and develop an approach that would help to ease the ACs¹ processing. The ACs have proposed separate data directories; however, the DCs did not agree with this concept and would like to devise a different solution.
The Network Monitoring session was chaired by Ruth Neilan and Angie Moore; most of the time was devoted to discussion. The CB is committed to resolving the data discrepancies and consistency problems and plans to improve quality monitoring in conjunction with regional network managers. It was noted that each IGS agency must be responsible for the validity of their information, and that the CB could monitor and respond, but that each agency has the ultimate control. Updating of site logs and where the logs should reside in order to minimize future problems was also discussed. Consensus was that appropriate software tools (such as an automatic site log generation and editing program) should be developed and used at the lowest level possible, closely monitored and assisted by the Network Coordinator at the CB. The challenge to eliminate the data discrepancies and incomplete information by the end of 1998 was issued by Ruth Neilan. (Errors and inconsistencies files were posted at the workshop and individual stations were challenged to come forward with corrections as soon as possible).
In the late afternoon, the Network Upgrade Panel addressed network enhancement issues and GPS instrumentation. Yehuda Bock presented a compelling case demonstrating that even mundane changes, such as antenna reinstallation/antenna replacements (of the same type) can create significant (mm) changes. It was argued that a significant subset of the IGS stations (e.g., the 47 ITRF stations) should receive special care. It was emphasized repeatedly that the IGS needs clear guidelines on how to manage past, present, and future hardware changes so that solution step functions are mitigated.
The last session was devoted to the GLONASS and GPS manufacturers (3S Navigation, Trimble, Allen Osborne, Javad, Ashtech, and Leica). All representatives assured the audience that Y2K and Wk 1024 problems will be tested and by early next year firmware upgrades (if necessary) will be distributed to users (free of charge). The newest receiver and antenna demonstrated to the IGS was from Javad; this antenna seems to include a few innovations, such as a new choke ring design with separately tuned for L1 and L2 frequencies (somehow electronically), a newly design, forty channel GPS/GLONASS receiver with tracking aiding (a high elevation SV aids the low elevation SV tracking, purportedly increasing the low elevation SNR by 10 Db!).
Jan Kouba has pointed out that while he was listening to the representatives from the receiver manufacturers, showing how every one is nowadays supplying six pairs of observables (code and phase on C/A, P1 and P2) it occurred to him that the IGS currently has a problem. Namely that the TurboRogues (the vast majority of receivers in the IGS network) under AS give C/A and (C/A + P2-P1) instead of P2, and that, we cannot match this pair with any of the above three pairs. T he implications of this fact are inconsistencies/incompatibility (at 1-2 ns level) in the satellite clock solutions and L1/L2 satellite calibration biases. Clearly this needs to be thought through and the IGS should make some decision for current and future hardware updates. This concern was discussed with some of the hardware representatives and clarification/confirmation on what specific receivers exactly output should be forthcoming, and hopefully soon.
On Wednesday night the group was graciously hosted by Leica for a guided, one-hour historical candlelight tour of Annapolis, led by several local women in authentic 17th century dresses, and with the full moon rising. This activity was followed by a banquet dinner at the historic Brice House of Annapolis. Many thanks are due to the folks at Leica, Tom Stansell and James Stowell, for sponsoring such an enjoyable social event.
Day 4 - Thursday, November 5
On Thursday, the New Communication Technology session included discussions of the next generation Internet and the STARTAP network and were presented by Steve Goldstein, Director at the National Science Foundation of the Networking and Communications Division. Currently some fast connections devoted to research and education are already globally available. It was suggested to talk to your local Internet service provider to explore the possibility of getting connected. Steve's slides are included in the proceedings. In the second presentation, a private communication consultant, Greg Peisinger, concentrated mainly on getting data from stations via various communication techniques, as well as the real time applications.
The workshop ended by asking attendees to join in working group discussions: Physical Sites, Data Centers, Network Upgrades (including communications and future requirements), and Network Monitoring. After the lunch break, the leaders of each of the above four groups presented summary recommendations and conclusions for the forum discussion.
The Physical Site groups recommendations included the outstanding issue of the unique four-character ID, and as noted above, it was taken upon by the group leader, Werner Gurtner. Also recommended was rigorous standardization of site log formats and time stamping of site logs, with incorporation into the SINEX files.
The Data Center group presented a list which included the need for developing DC guidelines, immediate site log updates (hours not days), planning for handling and use of hourly data, a commitment to look into the flagging of data considered non-compliant or 'bad' by ACs, and to work with the CB to develop the station functional characteristic matrix (identifying what IGS stations may meet requirements for various applications).
Network Monitoring also presented a set of recommendations, including team and community building, noting that most problems stem from lack of communication thus focusing on the need for improving communications amongst DC/AC and IGS station managers. 'Accreditation' or re-registering of IGS stations should be considered. The group also supported the 'challenge' by the CB to eliminate station discrepancies.
The Network Upgrade group suggested that a selected subset of important stations be systematically updated with 'homogenous' hardware and that these sites have two 'hot' installations which are established and analyzed. This created a great deal of discussion that was impossible to resolve in the short time remaining; it was recommended that the Infrastructure Committee should perhaps prepare a proposed plan on how to approach this issue.
Day 5 - Friday, November 6
Friday morning was a summary session for the program committee and the session chairs. Ruth Neilan encouraged Jan Kouba to begin by presenting the 'AC shopping list' (i.e., the recommendations of the Zumberge, et. al. position paper presented on Monday). This list was not completely included in the context of the established panels and discussion groups and it is important for the AC considerations and suggestions to be addressed.
Each panel moderator again presented the final recommendations of each group and these are included in this proceedings document.
At this time, it is most appropriate to thank the program committee, session chairs, and panel members for a job well done. Many thanks to all the contributors and participants. A very special and heartfelt thanks goes to Carey Noll and her staff for their outstanding local organization and attention to detail. The workshop was a wonderful experience for all who attended and has produced many sound recommendations for the inevitable improvement of the future IGS network.
(Prepared and summarized by Neilan, Noll, Kouba, and Moore)