University of Michigan: Revolutionizing Research Through Virtual Infrastructure
How the University of Michigan’s ARC-TS gives researchers secure, on-demand computing for sensitive data
Summary
University of Michigan researchers need to analyze large, sensitive datasets under regulations such as HIPAA and CUI. Building separate IT infrastructure for each research group was slow and costly. ARC-TS built a private research cloud on VergeOS that subdivides one pool of resources into secure enclaves with their own enforced policies, set up on a moment’s notice. Researchers reach their data securely from anywhere, and ARC-TS no longer goes through data center recertification for HIPAA and CUI.
Secure research computing on demand, without a new build for every group.
Research at the University of Michigan
Research is integral to the University of Michigan, a world-renowned institution with a rich history of interdisciplinary research and annual research expenditures above $1 billion. Its achievements span science, engineering, medicine, the social sciences, management, education and the humanities.
Research institutions like U-M need a custom approach to storage, networking and analysis to meet expanding and ever-changing needs. Connecting people, labs, data and computers, addressing security and regulatory concerns, and doing it without breaking the bank has historically been a tough problem.
About ARC-TS
To accelerate the availability of state-of-the-art cyberinfrastructure services and give the academic community a competitive advantage, U-M established Advanced Research Computing, Technology Services (ARC-TS). ARC-TS supports 2,600 independent groups and provides researchers across campus with the computational resources they need to produce groundbreaking scientific discoveries.
ARC-TS designs, obtains and maintains cutting-edge resources for data-intensive and computationally demanding research. Each group it serves has its own computing needs, and a large majority have specific data security requirements under the Health Insurance Portability and Accountability Act (HIPAA) and the Children’s Online Privacy Protection Act (COPPA).
To let researchers analyze sensitive datasets securely and stay within regulatory guidelines and agreements, VergeIO teamed with ARC-TS to provide high-performance, secure and flexible computing environments. These environments are secure enclaves that can be set up on a moment’s notice and take far less labor and cost than a traditional IT infrastructure.
Infrastructure That Keeps Up with Researchers
Researchers at academic institutions face a growing challenge. Their work calls for the analysis of ever-larger and more complex datasets, many with regulatory restrictions. Establishing a traditional IT infrastructure for each research group is time consuming and expensive. To stay at the forefront of academic and scientific research, U-M researchers need flexible compute and storage environments that deliver consistency, performance and reliability.
Health care research alone involves large amounts of sensitive information and diverse data types. That research is imperative to the advancement of patient care, and it requires the analysis of millions of clinical observations, including genomic, hospital, outpatient, pharmaceutical, laboratory and cost data. Work at that scale needs a secure, high-performance computing framework with multi-tiered storage.
U-M researcher Dr. Bryant Oliphant, who focuses on orthopedic trauma health services, spent much of his time figuring out how to get the datasets he needed within the limits of his computing power.
“Traditionally I’ve had to basically totally modify the database to get datasets, and [then I have to] append and merge them accordingly just to get them underneath the RAM limit for my computer. It just doesn’t work.”
Dr. Bryant Oliphant, U-M researcher, orthopedic trauma health services
The Virtualized Data Center
VergeOS is the University of Michigan’s private cloud environment for research. It brings processors, memory, storage and networking together and delivers scalable storage with quick, easy deployment of applications as needed. ARC-TS subdivides that collection into smaller units and allocates them to research projects as needed, where virtual machines and containers use them.
The infrastructure reconfigures automatically as resources are added, which gives ARC-TS flexibility without changing hardware or redoing capacity planning.
Global governance keeps each environment secure and supports greater collaboration among researchers. Each data center operates with its own set of automatically enforced policies and requirements. Stringent security protocols, including two-factor authentication, next-generation firewalls and system intrusion detection, let researchers reach their data centers safely and easily over the internet.
Five Research Services on One Platform
The ARC-TS research cloud on VergeOS delivers a suite of key services:
- Secure Enclave Service. A restricted-data desktop and server solution.
- On Demand Services. A container provisioning and management solution.
- Research Virtual Desktop Infrastructure (VDI). Capabilities similar to the Secure Enclave Service, suited to unrestricted data.
- Database Hosting. Research-oriented offerings including traditional RDBMS, NoSQL and time-series databases.
- Data Pipelining Tools. A wide variety of data ingestion, management and storage tools.
The Secure Enclave Service is the first generally accessible enclave service on campus for unrestricted, HIPAA and Controlled Unclassified Information (CUI) data. Researchers can reach their data easily and the security and compliance controls stay in place.
The Hardware
The ARC-TS environment runs on 60 nodes: 40 compute nodes and 20 storage nodes.
| Component | Compute nodes (40× H2400i-E5) | Storage nodes (20× S2400i-E5-HDD) |
|---|---|---|
| Processors | 2× Intel E5-2680v4 | 2× Intel E5-2620v4 |
| Memory | 512 GB DDR4 2400 MHz | 128 GB DDR4 2133 MHz |
| Networking | 2× 40GbE adapters | 4× 10GbE adapters |
| Flash | 2× 800 GB NVMe SSD (DC P3700) | 2× 800 GB SSD (DC S3610) |
| Disk | None | 12× 6 TB 7200 RPM SATA |
Secure, Efficient Research Technology
VergeOS lets researchers across many disciplines deploy and manage resources in a secure, scalable environment, with out-of-the-box capabilities:
- Efficiency. Built on cutting-edge Intel components, including NVMe storage.
- Flexibility. Robust, feature-rich software-defined networking.
- Cost-effectiveness. Data deduplication for more efficient use of storage.
At U-M, IT administrators can quickly spin up complete virtual enclaves, and researchers reach restricted data securely through a virtual private network. Tasks such as system auditing and updates are automated, which makes the system easier to administer. ARC-TS no longer has to go through data center recertification and compliance for HIPAA and CUI, saving time and resources. Consultants and support staff also share a common platform for support.
ARC-TS can stand up environments quickly and support its researchers better, and researchers reach their work securely and on demand.
“I used to be locked to my desk. And if I had to do something big, it was like, do I have to drive in on the weekend… or if you’re away on a trip, you can’t. This is huge. This is a gamechanger.”
Dr. Bryant Oliphant, U-M researcher, orthopedic trauma health services
Dr. Oliphant’s views are his alone and do not represent the University.
VergeOS gives universities and institutions a more achievable, easier-to-manage solution, with the scalability and flexibility to meet the research community’s ongoing needs. It also supports U-M’s Data Science Initiative, a major university investment that expands opportunities for students and faculty in fields such as transportation, health sciences, learning analytics and social science.
| Area | Traditional approach | With VergeOS |
|---|---|---|
| New research environment | A separate IT build per group | Secure enclave set up on a moment’s notice |
| Compliance | Data center recertification for HIPAA and CUI | Recertification no longer required |
| Researcher access | Tied to a desk and its RAM limit | Secure access over VPN from anywhere |
| Administration | Manual auditing and updates | Automated auditing and updates |
| Growth | Hardware changes and capacity planning | Automatic reconfiguration as nodes are added |
The Future of Research
The VergeIO research cloud gives U-M researchers scalable, secure data centers. It protects data integrity, maximizes collaboration and supports the production of groundbreaking scientific discoveries. The partnership between ARC-TS and VergeIO opens access to data-intensive research by removing cost barriers and cutting the time needed to build a traditional architecture.
ARC-TS administrators can keep meeting the ongoing needs of the research community, which matters enormously at an institution of this size. What once seemed insurmountable is now attainable: a simple, manageable system for research faculty and students.
The approach has the potential to change the way research data is stored and managed far beyond the University of Michigan. VergeIO continues to look for ways to create greater collaboration and data sharing among researchers across universities, organizations and enterprises.
Build a Research Cloud on VergeOS
The University of Michigan gives 2,600 research groups secure, on-demand computing on one VergeOS platform. Talk with a VergeIO engineer about your research environment.