December 19, 2023
By Holly Smith.
In late October, the Research Computing Center (RCC) hosted Cloud Days and officially unveiled Skyway. Skyway is an integrated platform developed by the RCC that enables end users to seamlessly burst compute jobs from Midway to commercial Clouds: Google Cloud–GCP, Amazon–AWS, and, coming soon, Microsoft Azure. Skyway is an open-source package that enables end users to submit workloads to the Cloud the same way they do on-premises, access high-capacity data storage when using Cloud resources, and use software packages and environments on the Cloud that mirror those on the premises. “Skyway makes it easy to take advantage of Cloud computing from the existing HPC environment. This is particularly beneficial for on-demand, time-sensitive workloads because researchers will have access to the necessary resources right away,” said RCC Senior Computational Scientist Trung Nguyen. Additionally, faculty and researchers benefit from the University negotiated discounts and are provided with an estimated cost of the Cloud usage a priori. Cloud Days attendees participated in hands-on demonstrations of Skyway.
Cloud Days also included specialized workshops facilitated by each of the Cloud vendors. During the Google’s workshop, three topics were covered. Firstly, what users can do right now securely and safely with Google’s generative artificial intelligence (AI) technology was presented. This portion went over important security aspects, such as data sovereignty and security of IP and responsible generative AI. The next section featured Google AI demonstrations and use cases. Finally, the workshop concluded with a discussion on the Google Vertex GenAI platform.
During the AWS workshop, AWS first introduced some basic AI and machine learning (ML) concepts and algorithms, and how Amazon SageMaker and AI services make it easier for researchers to apply AI/ML in their research. AWS then used SageMaker Studio to walk attendees through some AI/M examples, including basic regression, classification, and clustering. AWS rounded out the workshop by showing attendees how easy it is to get started with GenerativeAI on AWS.
During the final engaging and informative Azure workshop, Microsoft delved into the world of Azure OpenAI and its applications in research. The session was divided into three key sections. The first focused on understanding Azure OpenAI. Microsoft kicked off the workshop by introducing attendees to Azure OpenAI, its capabilities, and how it can be leveraged for research purposes. Everyone explored its features, benefits, and real-world applications. The next section was centered around compliance and responsible AI. In this portion, Microsoft discussed the importance of compliance in AI and how Azure OpenAI ensures the responsible use of AI. Topics such as data privacy, ethical considerations, and regulatory standards were all examined. The third and final section was on researcher enablement and skilling. It was dedicated to empowering researchers with the skills needed to effectively use Azure and Azure OpenAI for research. Microsoft shared resources, discussed best practices, and offered insights into how to maximize the potential of Azure.
“I am pleased to see all the hard work of the RCC team and successful collaboration with the various Cloud vendors come to fruition with this official launch of the Skyway platform. Adding a hybrid, on-premises and Cloud solution that is easy to use is instrumental in expanding RCC research computing support services to the University’s research community,” RCC Director, H. Birali Runesha said.
The RCC is grateful to the whole Skyway team as well as our Cloud partners Google, AWS, and Microsoft for making this event possible. Learn more about Skyway or request an account on the Skyway website.