People

Impact

Cornell’s HACC technologies (Isis2, SuperClouds, formally robust protocol synthesis) enable cloud-hosted machine learning solutions for data gathered in real-time, securely, fault-tolerantly, and with very strong consistency guarantees.

Cloud computing offers many advantages including:

• Ability to collect massive data sets and apply massive computing power on demand (as needed)
• New machine learning tools optimize decision making
• Potential to support huge numbers of user systems

Our technology adds:

• Scalable consistency, security and reliability
• Ability to update data even as machine learning algorithms process that data
• Formal reasoning about core components
• Tools and languages for quick development and deployment
• Elimination of vendor lock-in and stovepipes

We anticipate several tangible benefits to end users as a result of this project.

• Our approach extends today’s standard three-tier cloud infrastructure into a new world that mimics the breakdown of tiers and roles yet can achieve much stronger assurance.
• Our science basis yields deep insights applicable to other cloud computing challenges and problems similar to but not identical to the ones we focused upon.
• The outcome will be practical tools that let the cloud computing developer work in a familiar style, yet achieve applications that are automatically capable of providing high assurance, adapt and heal themselves as disruptions occur, and secure sensitive data, all while guaranteeing consistency of a formally rigorous kind and security for sensitive data.

Our goal is...

...a cloud computing system that can bring cloud-hosted machine-learning and automated reasoning to bear on critical military applications with high assurance, scalability, and cost savings.

From a project standpoint our main research goals are to:

Create a scientific basis for scalable, high performance replication with strong consistency properties and strong security, that explicitly models the phenomena most important in cloud deployments and use sound methods to overcome those problems, and publish the findings.

Create software solutions that can be demonstrated in real cloud computing settings and that demonstrate scalability and high performance while maintaining resiliency as disruptive events occur.

Work with early adopters from the Air Force, the networking industry and the cloud computing community to show precisely how our solutions can be used by application developers to solve their toughest problems.

Innovations

Isis2 – a new option for cloud computing, with:

• Scalability, consistency, security, fault-tolerance
• Ida: Machine learning platform that runs at cloud scale with strongly-consistent real-time updates
• DMake: quickly build large reliable cloud applications

Superclouds – virtualization and cloud storage, with:

• Computation migration, replication, and consistency
• VirtualWire: Virtual network migration
• Reliable cloud storage even in a contested environment
• Cloud-in-clouds straddles cloud provider boundaries