Developments in Cloud Computing, Dynamic Data Routing, and Virtualization for an Efficient Seismic Network

The Southern California Seismic Network (SCSN), operated by Caltech and USGS, utilizes state of the art computing technologies to provide timely earthquake event notification , ShakeMap, and other data products. The SCSN contributes data to the ShakeAlert Earthquake Early Warning (EEW) system, which must process seismic data within seconds of being recorded, and SCSN maintains a robust and resilient network of more than 350 digital strong motion and broadband seismic stations to achieve this goal. Here we present recent and ongoing innovations in Cloud Computing, Dynamic Data Routing, and Virtualization that keep the network running securely and efficiently. The SCSN operates its own data processing and archival center at Caltech, but we have leveraged Amazon Web Services (AWS) capabilities to provide additional redundancy and allow critical off-site operations continuity. Current SCSN redundancy employs AWS Elastic Cloud Computing (EC2) instances which provide capabilities for data acquisition, processing, calculating origin time, location and magnitude, and distributing earthquake alerts directly from the cloud even if part or most of our terrestrial systems are incapacitated. Using AWS services, we also provide redundant capabilities for post-processing information, Shakemap and data archiving. Current developments to use Apache Kafka and native AWS services (like Kinesis, Lambda, Elastic Container Services, Simple Storage Services) to integrate Machine Learning (ML) for picking and associating events in real-time processing will be discussed. Dynamic data routing methods are being employed at SCSN to provide path redundancy capabilities for data acquisition from the field to both terrestrial and cloud servers using multiple pathways, making the overall system very resilient to partial telemetry failures. We have continued to improve data path diversity and use dynamic routing to increase fault tolerance within our network. Our expanded virtualization project at the Caltech/USGS data centers and the AWS cloud allows us to provide functionality by using Virtual Machines, and Docker containers. Virtualization allows us to make the most efficient use of high-performance hardware and cloud services, increasing the flexibility and scalability of our data processing systems.