Head in the Cloud? It’s Not Such a Bad Thing, After All
Typically, when you think of Amazon, you think of the online retail giant where you can purchase virtually anything from toys to toilet paper, in just a few clicks.
What you may not know is that Amazon has a whole other arm, separate from its retail division, that provides on demand, cloud-based computing services used by big name companies such as Pinterest, GE, Slack, Airbnb, and MLB. It’s called Amazon Web Services (AWS).
Here on the Farm, over 600 faculty, students, staff, business units, schools, labs, and institutes have accounts with AWS (which indirectly benefits the entire Stanford community), thanks to an Enterprise Agreement that University IT (UIT) negotiated and maintains with Amazon.
The computing needs of these users range from data processing and warehousing to networking, analytics, archiving, and much more. AWS provides all of this, while helping them conduct and scale their research, create game-changing applications, and provide information and services more efficiently—and sometimes with a significant cost savings.
Hosting and network connection solutions
This was the case for Scott Stocker, Senior Director of Web Communications in the Office of University Communications.
Scott and his team were using AWS web hosting services for news.stanford.edu. With the help of UIT's Emerging Technologies group, they configured another AWS service called Direct Connect, which provided a lower bandwidth rate for on-campus traffic to the site, therefore reducing costs significantly.
“Since so much of our traffic comes from on-campus users, Direct Connect ended up reducing our monthly bill by about 40 percent,” Scott said.
Direct Connect has also proved to be a successful solution for scaling other highly trafficked sites like emergency.stanford.edu and stanford.edu. Consider a situation where a new Nobel Prize winner at Stanford has just been announced, or an earthquake has rattled campus.
“You don’t want to run out and buy more hardware to handle the spikes in traffic each time those specific events happen; rather, you want to scale up and down as needed,” said Bruce Vincent, Senior Technology Strategist and Director of Emerging Technologies for UIT. “AWS enables us to do that quickly and easily.”
Research and database solutions
Elsewhere on campus, teams are using AWS to support research and provide database solutions in a scalable way.
At the Graduate School of Business, Dave Love and his DevOps team partnered with Research Support Services (RSS) to create and evolve a “big compute” solution called CloudForest, which uses AWS. CloudForest can launch a grid of hundreds of machines used in research to perform computationally intense workloads.
“The primary driver for us is scalability, with an ancillary driver around quick provisioning time,” said Dave, Associate Director of DevOps & Cloud Architect. “CloudForest gives us both, since there’s no practical way we could spin up hundreds of virtual machines within Stanford’s own infrastructure, and certainly not as rapidly as it can be done with AWS.”
Dave’s team is also partnering with UIT to explore how AWS might help with storage solutions for disaster recovery and archiving.
These are just a few scenarios where AWS has proven effective for computing needs at Stanford. But, according to Bruce, there is no slowing down when it comes to cloud-based initiatives. His team in Emerging Technologies is continuously testing solutions and helping to implement them across the university.
“You’re going to be hearing a lot of cloud strategy initiative news over the next few months, and it’s all good stuff,” he said.
Bruce is available to answer any questions about cloud-based services and Amazon Web Services. If you are considering whether it might be a solution for you, or if you’d just like to learn more, contact him at firstname.lastname@example.org. You can also visit aws.stanford.edu to learn more about Stanford’s Enterprise Agreement with AWS.