In the Fall 2016 of The Next Page, Param Bedi, Bucknell’s Vice President for Library and Information Technology, discussed our enterprise information systems transformation. We are moving to numerous “best of breed” applications, all of which are Software as a Service (SaaS) offerings running in cloud environments. With so much of our data transitioning to these new systems, we’ve also had to rethink our data and business intelligence platforms. As we began to architect this solution, we asked ourselves if there were any advantages to running these platforms in the cloud as well. We eventually made the decision to build our data warehouse on Amazon Web Services (AWS). Since we began this effort, we have learned a lot about the platform including its many advantages and numerous services not generally available in on-premise data centers, as well as many things that need to be considered in a cloud deployment.
The advantages of public cloud platforms are probably too many to mention here, but let’s focus on some of the key ones. First of all, cloud deployments, by their very nature, result in simplified infrastructure management since an organization no longer needs to manage and maintain physical infrastructure. This also means that organizations are able to avoid many of the capital investments associated with infrastructure refreshes and expansion.
Cloud services also offer lots of scalability options. Most on-premise data centers can handle vertical scaling (increase of resources on a server) and horizontal scaling (addition of new servers to share the load), but AWS also offers Auto-Scaling. Consider a web application which requires additional processing power during one month of the year. Auto-Scaling will automatically create additional load-balanced servers, and when the load returns to normal, it can be automatically downsized. Of course, Auto-Scaling cannot be used on all AWS services, but when it can be leveraged, it’s a game-changer.
Another significant advantage of cloud services is that they often allow shifting of the focus from disaster recovery to high availability. Amazon currently has 16 regions across the world and each of those regions has 2-5 availability zones (essentially, separate physical data centers). Many AWS services have the option of built-in high availability across zones. For extra protection, this can be taken these even further, creating high availability solutions that span regions.
In the past decade or so, data centers have gone from big iron to decentralized servers to virtualization to public cloud. But AWS is disrupting that even further by pushing towards the elimination of servers altogether (at least from their customers’ standpoints). An example is their serverless computing solution, Lambda. In a traditional environment, applications and source code require some server to host and instantiate them. With Lambda, servers are no longer a requirement—developers simply write their code, deploy it, and pay for the compute time that is consumed.
And perhaps the most important advantage of cloud services is the agility and flexibility they provide. But, with this flexibility comes additional challenges, there are some key things that must be considered when deploying solutions in the cloud. Once a concept has been proven and begins moving it through its lifecycle, some governance must be put in place. In addition, universities are held to high standards when it comes to protecting personally identifiable information, so security controls and standards are critical.
At Bucknell, we’ve established a governance committee specifically for this purpose. The goal of this committee is to create a balance between the agility and flexibility of the cloud, while preventing a free-for-all environment, which would inevitably lead to poorly managed systems, lack of standards, and porous security. Through this group, we’ll establish standards, policies, and procedures; we’ll implement best practices around security, cost management, and architectural integrity; and we’ll act as a broker for others on campus, promoting consistency in the management of security, compliance, and costs.
To summarize, there are a great many advantages of the public cloud, in general, and AWS specifically. But organizations must be sure to balance the flexibility of these environments with a governed approach. At Bucknell, we’re still learning and still working all this out, but we fully expect to obtain the full value of the cloud while also ensuring that our solutions are secure and properly managed.