Cloud Computing

Maximizing the effectiveness of shared resources on the cloud

Overview

The Cloud and Interactive Computing (CIC) group at the Texas Advanced Computing Center develops, deploys and administers production-grade cloud resources, applications, services and tools for the national research science community. CIC has four major areas of focus: programmable cloud infrastructure-as-a-service, managed cloud-native platforms and applications, consulting, and research and development efforts into new cloud and interactive computing technologies.

Infrastructure As A Service (IaaS, Systems)

CIC operates multiple national-scale production clouds funded by grants from agencies such as the National Science Foundation that provide self-service academic cloud computing capabilities.

Chameleon

Chameleon is a highly configurable cloud computing resource for large-scale cloud and computer science research. Chameleon includes both bare-metal and KVM provisioning with diverse and heterogeneous hardware types to enable experimentation with alternative technologies.

Jetstream

Jetstream is general purpose, programmable, cloud computing for US researchers and science gateway developers. Provides self-service provisioning through APIs, command-line tools and multiple user interfaces.

Platform As A Service (PaaS, Applications/Services/API)

On top of the lower level infrastructure-as-a-service, CIC develops, deploys and administers higher level cloud and interactive computing platforms, applications and tools.

Science-as-a-service API platform

Science-as-a-service API platform is a platform of RESTful API services for the open science community, TAPIS provides APIs for many of the common distributed computing tasks including, data and metadata marshalling and movement, code execution and job lifecycle management, identity and authorization.

TACC JupyterHub

TACC JupyterHub is a customized, self-service Jupyter notebooks and cloud computing tools and libraries for specific communities and domains of science. The TACC JupyterHub is integrated into other TACC platforms and systems such as TAPIS and the Corral high-performance storage system.

Abaco

Abaco (Actor Based Containers) is a functions-as-a-service platform based on Linux container technology and the actor model for concurrent computation. Abaco provides a web service layer over a container execution engine to enable reproducible, event-driven computing patterns.

Consulting

CIC engages groups both internal and external to TACC to help them design and build customized science clouds tailored to their research community and automate the management of the underlying resources and infrastructure. CIC specializes in the following areas:

  • OpenStack deployments for infrastructure-as-a-service offerings.
  • Distributed systems architecture and microservice/API design.
  • Systems administration automation with DevOps tools such as Puppet and Ansible.
  • JupyterHub and interactive computing environments.

Research & Development

CIC conducts research and development into multiple areas of cloud and interactive computing. Current areas of interest include:

  • Highly available, distributed cloud/object storage.
  • Automated, self-service provisioning and management of resilient stateful services such as databases and message queues.
  • Containers, unikernels and related technologies for high-throughput and reproducible scientific workloads.
  • Jupyter project/Jupyterlab notebook environments that bridge the gap between interactive computing and more traditional HPC.