Google Cloud helps to drive tech innovation at Bloorview Research Institute

Established in 2004, the internationally recognized Bloorview Research Institute (BRI) brings together a multi-disciplinary team of scientists who work collaboratively with clinical staff, clients, and families to generate clinically linked and applied pediatric rehabilitation research.

BRI is guided by three themes: exploring human dimensions of care and disability, developing and improving interventions, and understanding developmental diversity.

BRI is housed within Holland Bloorview Kids Rehabilitation Hospital, Canada’s largest pediatric rehabilitation and continuing care teaching hospital. The hospital is affiliated with the University of Toronto and provides care to over 8,000 children with complex medical needs and disabilities every year. Watch the Holland Bloorview video to learn more about their mission. 

Business challenge 

BRI is composed of nearly 250 personnel and supports over 40 researchers and investigators in 12 different labs, with specialties such as autism, critical disability and rehabilitation studies lab, orthotics, prosthetics, engineering, and inclusive communities. The data collected and encoded in the labs range from brain signals, blood oxygen levels, muscle vibration signals, prosthetic limb movements to video and MRI scan images. 

“As we continue to expand our research enterprise as Canada’s largest hospital-based research institute in childhood disability, our research technical requirements are becoming more complex,” says Pazikah Kozak, Chief Information Officer at Holland Bloorview. “SADA has partnered with the Bloorview Research Institute to implement a robust storage solution on Google Cloud to ensure our researchers can access their data more quickly, securely, and with greater flexibility.”

The advancement of data-intensive compute research practices combined with aging infrastructure began to present challenges for BRI. Data was stored in an on-premises data center comprising servers, some of which were approaching ten years old. “With an older infrastructure, you have to worry about how much longer you can operate on existing servers before they are obsolete, and what investments in new equipment will look like,” says Ka Lun Tam, Research Engineer and Special Projects Manager at the research institute. 

Additionally, research only moves as fast as the data can be crunched. “In terms of performing analysis, we typically had desktop computers and workstations dedicated to running various data analysis scripts,” says Tam. “On a good day, it will run for an hour. Or it may take a week due to the volume of data. Using the equipment we had, the lag time between acquiring data and having results was becoming a pain point.”

At times, BRI would purchase additional supercomputers to run one-off analysis for a particular lab. Once the immediate business need was accomplished, the equipment would often sit idle for an extended period of time. Eventually, many labs started offloading work to third-party cloud computing platforms to improve efficiency.

“Individual labs were developing unique, independent relationships with a variety of cloud providers,” says Tam. “It was a bit of a hodgepodge. So we were looking to come up with one institute-wide vendor that could manage all our cloud, computing, and storage needs as well as open the door to enhanced AI and Machine Learning capabilities.”

Solution

The engineering team at BRI reviewed the portfolios of leading cloud service providers and selected Google Cloud to run a proof of concept (POC) before making a final decision. Google Cloud encouraged BRI to engage SADA, a multiple-time Google Cloud Sales Partner of the Year, to help guide the trial process.

SADA gave a designated team of BRI researchers a Google Cloud sandbox to play in. It was really helpful for us to explore all the different functionalities of Google Cloud by using real-life data. We also appreciated SADA’s weekly support calls so we could ask them questions while we were investigating.”

Mani Kang, | Director of Research Operations and Business Development at Bloorview Research Institute

The POC focused on compute functionality, but the team also discussed best practices for storage and organizational structure within Google Cloud. BRI needed to limit and grant access only to the researchers working on specific studies to ensure their standard operating procedures and regulatory compliance requirements for data privacy. “SADA helped us create all the file buckets that we needed and showed us how to automate the process so every lab, each dictating their own unique structure, would have a configuration that’s repeatable,” says Tam. 

With the POC proving a success, BRI decided to migrate their existing infrastructure and go all in with Google Cloud for their research operations moving forward. BRI is the first hospital organization in the Toronto, Canada area to go all in on Google Cloud. Next, SADA worked with BRI’s information management and technology department to secure connections from their onsite data center to Google Cloud. The transfers of data were staged and completed on a per-lab basis.  

Impact

As a result of working with Google Cloud and SADA, Bloorview Research Institute was able to establish an institute-wide platform for cloud, compute, and storage systems. In the process, SADA guided them through a proof of concept that led to the migration of 15 TB of data from on-premises to Google Cloud. Additionally, SADA helped Bloorview Research Institute:

• Lower overall costs through efficiencies from the cloud environment
• Leverage Google Cloud cloud-native tools
• Improve ability to run compute-intensive research and analysis

It was surprisingly fast and simple to move to Google Cloud. With SADA’s support using Google Cloud’s Storage Transfer Service, the migration of files was straightforward. I was expecting us to be transferring our first terabyte every night for at least two or three weeks. As it turned out, we copied two terabytes on that first night.

Ka Lun Tam | Research Engineer and Special Projects Manager at Bloorview Research Institute

Nearly all the data from each of the labs have been migrated to Google Cloud. SADA is continuing to advise BRI on leveraging Google Cloud to increase performance and improve elasticity to expand their research and analytics capabilities.

“In addition to significant capex cost savings, our staff utilization has improved,” says Tam. “We no longer have to spend time considering whether or not a server is going to die or if our backup drives are full. We’ll never run out of disk space on Google Cloud.”