Proof-of-Concept HPC in the Cloud Solution Speeds MicroSeismic’s Data Analysis Process for Faster Client Insights
MicroSeismic, Inc. (MSI) provides oil field services to energy companies seeking ideal locations for hydraulic fracturing. Using its specially-developed proprietary solution, MSI helps its clients “see” underground to accelerate their operational decision-making. The process reduces the budget and time required for the exploration of fuel deposits and offers other benefits too. MSI’s technology can pre-determine the volume of resources in an underground deposit. Therefore clients gain the ability to evaluate long-term production worth before proceeding with oil extraction. Especially in times of economic uncertainly, the foresight can help MSI’s customers prioritize their business investments.
They can also determine the ideal spacing between wells for the maximum extraction of subsurface fuels. MSI’s technology can reduce disturbance to the surrounding environment by minimizing the number of fracturing sites tapped.
Accessing the unconventional fuel deposits deep below the ground involves injecting water, at extremely high pressures, into chosen well sites. The technique breaks the surrounding rock to release the subsurface fuels. MicroSeismic supports its clients during this phase of extraction by monitoring the hydraulic fracturing operations for effectiveness and efficiency. To date, the MSI team’s track record of success includes more than 45,000 fracture stages around the globe.
How does it work?
While the extraction concept is straightforward, the process of accomplishing it is not. MSI’s approach involves a complex array of sensors and extremely powerful high performance computing (HPC) systems. Thousands of sensors, called geophones, are coupled with a vast antenna array above the fracturing site. The process is akin to techniques used for earthquake monitoring but on an even grander scale.
The sensors capture terabytes of data while the geophones “listen” for the sound of rock cracking under pressure. Analysis of that information offers a 3D subterranean map of subsurface activity, even miles underground.
During MSI’s analysis process, the enormous trove of data divides into smaller batches. Each block of data represents about five minutes of information capture, making it easier for HPC systems to ingest and analyze. The nearby data center that hosts MSI’s HPC infrastructure tackled nearly six million data analysis jobs in 2019. The effort necessitated roughly 1.5 million hours of combined high performance compute time.
Supplementing on-premise HPC solutions
Microseismic is currently evaluating ways to augment their existing HPC solutions to accelerate time-to-insight. Doing so can help MSI’s teammates focus more of their time on customer services and less on data processing. Plus, the flexibility to scale HPC cycles up or down, as workloads dictate, helps MSI manage its capital expenditures.
With their goals defined, MicroSeismic sought a consulting firm to help them in their search for a supplemental HPC solution. That scouting effort ultimately led them to Six Nines IT, a Premier Consulting Partner in the Amazon Web Services (AWS) Partner Network.
Six Nines worked closely with MSI to address the scope of their HPC needs and develop an innovative proof-of-concept (PoC) solution. After evaluating several AWS solutions, including CPU and GPU optimized instances, the team recommended complementing MSI’s nearby data center with AWS EC2 C5d cloud-based HPC instances. Powering these instances are underlying 2nd Generation Intel Xeon Scalable processors with Intel AVX-512, Intel DL Boost technology, and AI capability built-in.
The underlying Intel ingredients, optimized for HPC and AI, provide the AWS C5d instances significant flexibility for scale. An HPC master node can spin up other cloud-based compute instances dynamically, effectively adding HPC nodes as workload demands dictate. The reverse is true too. When a data analysis job no longer needs extra horsepower, added instances will shut down. For this reason, the solution offers the extensibility and performance levels required to interpret MSI’s customer data quickly, while reducing the cost-per-terabyte of data processing.
HPC in the Cloud enables new opportunities
If the PoC is adopted, Six Nines will work to optimize MSI’s proprietary software for the new HPC environment. By making the application “cloud-native,” MSI and Six Nines can ensure the code realizes its full performance potential on AWS instances with Intel hardware. MSI’s longer-term goal is a combined HPC solution capable of processing the massive volumes of subsurface data in real-time. The PoC system can help MSI’s geophysicists to spend less time processing of data, and free them to focus more on data interpretation.
According to Peter Duncan, President, and CEO of MicroSeismic, “Data sets that previously required an overnight analysis could reveal meaningful insights in as little as thirty minutes.” Preliminary benchmarking by AWS of the C5d instances offer MicroSeismic up to a 10x workload performance improvement in comparison to their existing on-premises HPC solution. Added Duncan, “Based on MicroSeismic’s proof-of-concept analysis, Amazon EC2 C5d instances and underlying Intel technologies offer the potential to deliver faster results and greater value to our clients at a lower cost.”
A PoC, like that proposed by Six Nines, demonstrates the exceptional value derivable from HPC systems, whether on-premise or as a cloud-based service. By putting each workload in the location where it runs best, companies like MSI have a flexible foundation for faster client results.
Read the full MicroSeismic case study.
# # #
This article was produced as part of Intel’s editorial program, with the goal of highlighting cutting-edge science, research and innovation driven by the HPC and AI communities through advanced technology. The publisher of the content has final editing rights and determines what articles are published.
