Maurice Manning has recently joined Berkeley's Research IT team as a cyberinfrastructure engineer (CIE), supported by a grant from the National Science Foundation, and has already begun to make an impact supporting multiple research teams on campus.
Initial Engagements
Maurice has already collaborated with researchers in the Adesnik lab (Molecular and Cell Biology) on multiple challenges, including running IPython workflows in parallel on the Savio cluster, and defining custom kernels for JupyterHub. He will also work with the team to prototype how their research workflow might be designed to provide near-real-time, closed loop feedback using Savio to analyze data minutes after it is collected. To that end, he has developed a component that automates the movement of data from instruments in the Adesnik lab to Savio’s scratch storage, via Globus. This component has been tested successfully at multiple imaging labs on campus, including the Cancer Research Lab. You can read more about collaboration with the Adesnik lab team in a recent Research IT profile of doctoral student Evan Lyall, Systems Neuroscientists use BRC services to understand the neural basis of perception.
One goal of the CIE grant is to help researchers use workflow tools more effectively. Maurice has been working with both Kea Johnston (Near Eastern Studies) and Arianna Campiani (Anthropology/Archaeology) to enable their teams to use the Photoscan tool from within an IPython notebook. The most recent release of Photoscan, which creates 3D models from images, expands the command line capabilities of the software. Maurice is creating an IPython notebook which integrates with the Box SDK to move image data from storage on Box to the Savio space, run the compute intensive Photoscan tool on Savio nodes, then move resulting model data back to Box. The complete workflow will be contained in a well-documented notebook that facilitates modification by researchers without requiring strong HPC or command line skills.
Additionally Maurice has assisted the staff at Phoebe A. Hearst Museum of Anthropology with technology evaluation for the visualization wall for the museum’s new gallery. Quinn Dombrowski and Maurice have collaborated to assist Digital Humanists such as Adam Anderson, Kea Johnston, and Arianna Campiani to employ cyberinfrastructure tools such as Singularity and Jupyter notebooks to scale research. Other engagements have included working with other Research Data Management staff to advise researchers on solutions for data archiving, data migration, and structuring data to increase efficiency of processing.
How the CIE can support Berkeley researchers
The CIE role was established to augment an existing team of consultants with a resource that can work side-by-side with a research team to understand issues at a deeper level, identify and resolve specific roadblocks, and give longer term support than Research IT’s consulting service can provide. The CIE coordinates with Research IT leadership and consultants as well as faculty to identify research challenges that can benefit from this type of engagement.
The types of engagements targeted by this grant will use the campus cyberinfrastructure to accomplish research, increase scale, and accelerate productivity. Such activities would include (but are not limited to):
help researchers define science requirements and translate them into computation, network, storage, and workflow requirements;
enable researcher to utilize new forms of compute such as Spark and cloud-based technologies;
help to engineer the computational research solution at various points in the pipeline, including: choice of hardware/software at the data acquisition point; automated flow of data over the network; integration into compute resource (access and scheduler); and parallel optimization of the computational component;
work with research teams to implement data pipelines consisting of optimized data acquisition, intermediate storage, high-speed networking utilizing the Science DMZ, and medium- to large-scale computational resources, including the Savio HPC cluster provided by BRC; and,
utilize containerization and infrastructure virtualization technologies (e.g., Docker, Open vSwitch) to allow researchers to more transparently move their own, local, familiar development and computation environments onto more powerful campus cyberinfrastructure.
Components created during CIE engagements are intended to be reused by the larger Berkeley research community. Code and documentation supporting CIE-supported solutions are currently available in the CIE github repository; Research IT welcomes inquiries from researchers who seek assistance in using these components. As new solutions and components emerge, Research IT will provide details in news articles, trainings, town halls, and documentation.
Maurice joined the Research IT team after working as a systems architect designing cloud-based distributed systems for the Oceans Observatories Initiative at UCSD as well as the California PATH transportation corridor management project here at Berkeley. Prior to that he worked primarily in the biotech and life sciences domains. For more details on his experience, please see his BRC profile. Maurice and other members of the Research IT team are developing relationships with CIEs at other institutions, creating channels that will facilitate the exchange of information and potential future collaborations. There are currently three other CIEs in the UC system.
Engaging the CIE in your research project
Experience has shown that collaborations between campus researchers and IT support staff are most effective when researchers engage with Research IT early in a project’s evolution, when technology requirements and resources are being identified. At this stage, the CIE may be able to provide assistance with specifications and design to reduce the time to establishing an effective analysis pipeline.
Research IT is actively seeking engagements of the types outlined above. If your research team or researchers you know could benefit from this type of assistance, or if you simply want to know more about the campus cyberinfrastructure supported by the Research IT team, please e-mail us at research-it@berkeley.edu.