If data is the new currency, as many technology experts believe, then CALIT2 just made a large deposit. The High Performance Wireless Research and Education Network (HPWREN) has installed a new virtual machine server and computer data storage complex on the fourth floor of the CALIT2 Building as part of HPWREN’s expansion from San Diego, Imperial and Riverside counties into Orange County. The complex will become a repository for all of the county’s HPWREN sensor and image data − and a rich resource for UC Irvine researchers.
The data servers are integrated with the NSF-funded Pacific Research Platform, which enables distributed computing on HPWREN data with over 6,000 central processing unit-cores and 600 graphic processing units.
HPWREN is a high-bandwidth, wireless, internet-connected sensor and communication network that supports applications in research, education and public safety. The network includes hundreds of cameras, including fixed 360-degree-view and closeup point/tilt/zoom models, as well as weather sensors and seismometers. It covers hundreds of miles of rugged terrain, from hard-to-reach areas across the remote Southern California backcountry almost to the Arizona border.
“It is the largest-scale and most fine-grained high-speed wireless network for environmental sensing in the world,” explains Larry Smarr, founding director of CALIT2 and the Harry E. Gruber professor in computer science and engineering at UC San Diego’s Jacobs School of Engineering.
Originally funded in 2000 by the National Science Foundation, HPWREN is now supported by its user community and is managed by the Scripps Institution of Oceanography’s Institute of Geophysics and Planetary Physics, in partnership with CALIT2’s Qualcomm Institute and the San Diego Supercomputer Center – all located on the UC San Diego campus.
Over the past couple years, HPWREN added installations in the Santa Ana Mountains, along the eastern edge of Orange County, and in Laguna Hills. It has become a valuable resource for the firefighting community as cameras provide live feeds of smoke and fire progress, and weather sensors provide wind direction and air humidity levels. It became a crucial ally for firefighters during the Holy Jim Canyon fire two years ago, as they relied on publicly available HPWREN images to decode the fire’s origin, stay apprised of its progress and plan their containment strategy.
Tirtha Banerjee, UCI assistant professor of civil engineering, investigates ecosystem disturbance processes through characterization, modeling and data analysis. His group plans to access the HPWREN image data for characterizing the landscape and use the associated sensor data for calibrating fire models. He is principal investigator of a UC interdisciplinary, multi-campus team seeking to help mitigate wildfires in California by developing better tools for understanding the use of prescribed burns to reduce wildland fuel.
“The common element in both ecosystem disturbance science, wildfires and prescribed fires is characterization of fuels, and we anticipate the HPWREN data will be useful for this work,” says Banerjee. “These characterizations are important for understanding fire behavior, but cannot be captured well from satellites and can be expensive with manual data collection.”
In addition to wildfire research, there are potential collaborations for researchers interested in designing their own studies using HPWREN capabilities. The network offers multiple opportunities for scientists interested in environmental sensing, data science, computational photography, machine learning and information sharing.
“It’s a completely application-agnostic platform on which you can place sensors,” says Smarr. “You configure it as you need for your specific research goals. This has included everything from listening for wolf howls, to 3D sonic arrays for bats, to observing ground and plant moisture, to most recently wildfires.”
HPWREN’s servers are interconnected by the Pacific Research Platform, a seamless research infrastructure that allows scientists to securely share large amounts of scientific data through high-capacity optical networks provided by Corporation for Education Network Initiatives in California (CENIC). Together the HPWREN and PRP create a wireless/optical fiber cyberinfrastructure that links UC Sand Diego, San Diego State University and UCI HPWREN servers at 100 gigabytes per second, providing data redundancy, disaster recovery and constant availability.
“There are a number of multiple and parallel developments within HPWREN, such as emerging new technologies, expansion of the backbone infrastructure and additional camera technologies that are all coming together at this time in a perfect storm of opportunity,” said Greg Hidley, a CENIC research engineer, HPWREN technical developer and a PRP collaborator, at a recent CALIT2@UCI workshop held to inform UCI engineering and computer science researchers about HPWREN’s capabilities.
Not only will UCI researchers be able to use this connection, with eyes across the county, but they will easily be able to access massive amounts of data from the entire network using the new server and storage complex in the CALIT2 Building. The PRP uses a Ceph open-source software object-, block- and file-level storage platform that adds over 500 terabytes of available distributed storage. In addition, there is a new industrial machine-vision camera system, along with a next-generation image-fetching system; and a Nextcloud file-sharing and collaboration platform. All together these new resources will support the creation of a repository for Orange County HPWREN sensor data and camera images. With this abundance of data, researchers can apply computational photography and machine-learning techniques to investigations that help predict the future by looking at the past.
CALIT2 Irvine Director G.P. Li encourages UCI faculty to take advantage of the network and the large quantity of accessible data now available.
“We have multiple researchers at UCI investigating different aspects of wildfires in California, and this is an important resource for them,” says Li. “It could be useful for climate change research as well; imagine if we pointed the cameras at the ocean and observed changes on the coastline, sea level and tides.”
“G.P. had this vision all along,” says Smarr. “He saw that data growth was an exponentially increasing means of conducting research and education, so he asked me and the HPWREN leadership to expand HPWREN data generation to Orange County and to use the PRP to establish its storage complex at UCI.”