CalREN Mid-Term Status Report

I. Statement of Project

The foremost need is for reliable continuous telemetry of seismic waveforms and earthquake event detections from our remote broadband digital seismic dataloggers to our central site at UC Berkeley. UC Berkeley currently has installed 14 high-resolution digital seismic stations in northern and central California, and will have a total of 21 systems installed within the next year. These stations provide valuable on-scale waveforms used to determine earthquake location, magnitude, and duration and frequency of ground motion during an earthquake. Continuous high-speed digital connections to our instruments allow these data to be delivered in near real-time for analysis at UC Berkeley. The second component is a reliable, high-speed link with the USGS seismic network in Menlo Park. The USGS runs a dense short-period seismic network in central and northern California, with a limited number of high-resolution broadband digital stations within the San Francisco Bay area, which complement the Berkeley network. The real-time exchange of seismic waveforms and parametric information between the Berkeley and USGS networks will allow UC Berkeley and the USGS to form a stronger real-time earthquake notification program and to develop the capabilities for an earthquake alert or early warning system.

The benefits of this project extend beyond the Seismographic Station, as rapid access to reliable earthquake information is critical for emergency mitigation operations. The mobilization of local, state, and federal disaster response in the first few hours after a major earthquake can be greatly enhanced by dependable near real-time estimates of location, magnitude, mechanism, and extent of strong-ground shaking. We anticipate that participation in the CalREN project will significantly improve the earthquake response capabilities of the UC Berkeley Rapid Earthquake Data Integration (REDI) Project, which automatically distributes information to public safety agencies, utilities and transportation companies moments after an earthquake occurs. The use of high-speed digital telemetry will make it possible to provide earthquake information within seconds after a major event, and lay the groundwork for the development of a system which could warn the public of imminent shaking, before the seismic waves arrive. Furthermore, the robust and reliable digital telemetry ensures that a complete archive of seismic data will be available to scientists and engineers for future study.

II. Statement of Objectives

Test and evaluate frame relay services for seismic data acquisition from remote sites

Under this objective, the Seismographic Station will replace its existing analog telemetry system with 56 Kbit frame relay connections. We will assess the effective bandwidth and the transmission delays and errors of this telemetry system for an earthquake notification system. We currently have operating connections to 4 of the Berkeley Digital Seismic Network (BDSN) sites and anticipate the addition of 7 more sites before the end of 1995.

Test and evaluate frame relay services for acquisition of other geophysical data

The Seismographic Station will experiment using frame relay services to transmit data from other geophysical instruments, such as continuous GPS receivers and electric and magnetic field sensors, that are co-located with BDSN seismometers. Continuous telemetry of the GPS data will permit rapid estimates of ground deformation useful for emergency response, while the GPS and electro-magnetic data will be examined for signals that may precede earthquakes. We anticipate testing this component of the CalREN project in the next week or two.

Test and evaluate frame relay services for exchange of earthquake data with the USGS Menlo Park

The Seismographic Station and the USGS Menlo Park will utilize frame relay services to exchange earthquake data in near-real time. These shared data will allow the Seismographic Station and the USGS to form a stronger real-time earthquake response system for northern and central California. The T1 connections for this exchange have been installed, but have not been tested. We are currently waiting on the delivery of hardware for the USGS of the connection before activating this exchange.

Test and evaluate frame relay services for seismic data acquisition with multiple recipients

The Seismographic Station will utilize the frame relay capability of establishing multiple permanent virtual circuits (PVC) to establish a emergency operations facility which will receive and process an identical data stream from a subset of our seismic stations. This will allow us to determine the practical aspects of configuring and administering multiple PVCs and data connections at our seismic stations as well as providing redundancy for earthquake notification. It will provide us with valuable experience that may lead to a future state-wide exchange of seismic data with Caltech in southern California.

III. Accomplishments to date

Based on these initial installations, we have begun to evaluate frame relay services as a component of an earthquake notification system. We have found the established frame relay connections to be extremely robust and evaluation of data transmission delays indicates that the frame relay service has eliminated the large, sporadic transmission delays and errors experienced using the analog telemetry.

The Seismographic Station will be in a position to complement the acquisition of seismic data with other geophysical information within a week or two when the Calnet connection (providing inter-LATA frame relay services) to our station SAO is complete and/or the connection at BRIB is activated. The site at SAO contains both seismic instrumentation and sensors to measure the electric and magnetic fields while the seismic instrumentation at BRIB is supplemented by a continuous recording GPS receiver.

Progress toward the goal of establishing a near-real time exchange of earthquake data with the USGS Menlo Park is currently delayed pending the delivery of the appropriate hardware for the USGS side of the connection (Table 2). In the meantime, we are coordinating with USGS personnel to develop the software necessary to exchange and incorporate the new data into our automatic seismic processing systems.

We plan to begin testing multiple PVCs in early 1996. This aspect of frame relay is a powerful advance in telemetry capability from the view of redundant communications. We plan to establish a "backup" processing facility which will receive a subset of data from the BDSN, providing emergency earthquake notification capabilities. This facility will serve as a testbed for establishing a state-wide system of redundant communication between UC Berkeley and Caltech.

Notes: "Other equip." refers to additional geophysical instrumentation:
GPS - Global Positioning System receivers;
EM - Electric and magnetic field sensors.

In order to inform and educate the public about the application of high-speed telecommunication by the Seismographic Station and CalREN, the Seismographic Station has established a document describing the experiment and the results on its WWW server (http://seismo.berkeley.edu/seismo). In addition to the WWW document, Seismographic Station personnel will give presentations at scientific conferences describing the project goals and results. For example, Mr. Douglas Neuhauser is giving an invited poster describing the Seismographic Station CalREN project at the 1995 Annual Meeting of the American Geophysical Union on December 13. The results of the project will also be described in the Bulletin of the Seismographic Station and in scientific papers.

IV. Problems and difficulties

• A. Unscheduled delay in the delivery of the initial frame relay equipment.
• B. Ongoing problems associated with the installation and configuration of the frame relay circuits by Pacific Bell.

A. Initial equipment delays

B. Problems with Pacific Bell Frame Relay Installations

• 1. Our initial circuits were scheduled for installation during the week of September 5, 1995. On September 28 we received the first written information concerning these circuits, which included installation dates, circuit numbers, and DLCI addressing. However, there was no information indicating that the frame relay network was not activated. Only when we reported apparent problems with the first frame relay connections were we informed that the network was not "turned up". The Pacific Bell frame relay service center was extremely helpful in troubleshooting our initial connection and we established our first frame relay connection after 90 minutes of work with the Pacific Bell service center.

• 2. When we tested our second connection at Byerly Vault (BKS), the circuit did not work. We placed a call to the Pacific Bell service center and were informed that it "tested good to the network interface, so the CSU must be defective", and that any service call would most likely be charged to UC Berkeley. After several phone calls to our Pacific Bell representative, we established that there was not an active network interface installed at our site, only an standard ADN unpowered RJ45 jack. UC Berkeley staff reported that there was 60-cycle noise on the line, so we opened a service call and a Pacific Bell technician was dispatched. The technician found there was an open circuit between the central office and BKS, and reportedly fixed the problem. However, the technician did not contact the frame relay service group to test the line. We discovered the circuit was still not operational, and opened a second service call on the circuit. This time, the service representative repaired the problem, and at our insistence, contacted the frame relay group to test the circuit with our CSU/DSU attached.

• 3. When we tested our third circuit at Jasper Ridge (JRSC), we could not establish a frame relay connection and placed a service call to Pacific Bell. During testing of the circuit, we discovered that the DLCI addressing was incorrectly specified as 21 instead of 16. When Pacific Bell corrected the circuit, we managed to establish a frame relay connection, but our equipment reported that circuit was being dropped and reset every 90 seconds. After two additional service calls to Pacific Bell, we performed a hard reset on the frame relay equipment at Jasper Ridge, and the circuit remained stable. The FRAD apparently got "confused" by the dynamic DLCI re-addressing of the circuit.

• 4. When we attempted to install our equipment at the Russel Reservation site (BRIB) on October 30, we again found that we could not establish a frame relay connection. We placed a service call and discovered that we were given the wrong circuit ID by Pacific Bell, and that they were unable to trace the circuit. We finally determined the correct circuit ID through the Frame Relay Port Order Number, and were informed by Pacific Bell that the circuit installation has never been completed due to "facility problems, even though it has been schedule for completion in the first week of October. The installation of that circuit has been escalated, and we expect it to be completed this week.

• 5. When we attempted to test our circuit to Richmond Field Station (RFSB) at the end of October, we discovered that the circuit installation had not been completed due to "facility problems" even though the circuit was scheduled for installation on October 12.

• 6. The DLCI addressing provided by Pacific Ball was incorrect for our Lick Observatory (MHC) installation. We were informed that the DLCI addressing for the PVC was 20 on the Lick Observatory circuit, and 16 on the T1 line at UC Berkeley. This conflicted with a prior assignment of DLCI 16 on the T1 line, and we informed Pacific Bell of this conflict. We were then told that the PVC DLCI addressing was 16 on the Lick Observatory line, and 20 on the T1 line, which also conflicted with a previously assigned DLCI on the T1 line. The Pacific Bell Frame Relay Service Center traced the line, and found that the DLCI 20 address referred to a connection between two Pacific Bell switches, and that the DLCI on the T1 line was really 23. However, the Pacific Bell customer representative could not confirm the correct DLCI addressing information for several weeks.

• 7. Pacific Bell has been lax in keeping us informed of the installation status of our circuits. We have not been always been informed when installations are completed or delayed to "facility problems", and have not always been provided with the correct circuit information in a timely fashion.

Each problem by itself is not a major obstacle. However, when viewed together, they do not present a favorable image of Pacific Bell's ability to configure, install, and support a frame relay network. There appears to be little communication between the circuit installation or repair crews and the frame relay service center. The information provided to us by Pacific Bell customer representatives concerning the physical details of frame relay installations has often been incomplete or incorrect. Pacific Bell has not been consistent in providing us timely or accurate information concerning the installation or configuration status of our frame relay circuits. Pacific Bell was provided with a complete and detailed list of all of our station locations before the CalREN project was approved in order to verify that frame relay could be installed at the sites, but this has not prevented at three "facility problems" with our initial installations.

The problems that we have encountered with our circuit installations and configurations have thus far not been major obstacles for us, primarily because we started with the stations closest to UC Berkeley. However, as we expand our frame relay network to incorporate circuits at more remote sites, it becomes much more critical that we are provided with timely and accurate information concerning the state of the circuits. It is essential that we know whether a circuit has been properly installed, tested, and configured before we travel to distant stations to install and connect our frame relay equipment. We request that Pacific Bell assign the appropriate staff and resources to this project to ensure that these problems are addressed in a timely fashion.

V. Summary

Through the CalREN Foundation, Pacific Bell is providing a valuable public service by its support of the Seismographic Station. By ensuring a robust and reliable flow of seismic data to UC Berkeley, Pacific Bell is contributing directly to the public safety of the residents of northern California. Information from UC Berkeley is utilized by public and private agencies to respond to earthquake emergencies, aiding in the recovery of lifeline systems such as transportation, water, power, and communications. We look forward to continued collaboration with Pacific Bell in contributing to safety and welfare of Californians.