Unlike Software as a Service, Infrastructure as a Service (IaaS) uses the cloud to host applications that we manage. Benefits of IaaS include metered use (paying only for what you use down to the minute) and the availability of instant scalability (more computing power is readily available to increase or decrease, as needed). Some IaaS offer low or zero cost options to try out and learn about their services. Often IaaS products and platforms require more technical knowledge to setup and use than SaaS products, but can be more flexible for specific uses that aren’t served by a vendor product.
The most common IaaS platforms being used on campus are: Amazon Web Services, Google Cloud Platform, and Microsoft Azure but there are other vendors that offer IaaS platforms. On the research side, you may gain more specialized service offerings from XSEDE or NERSC. If you are interested in cloud infrastructure (or high-performance computing) for research, Berkeley Research IT has excellent resources for consultation and assistance focused on research.
Do I need to use IaaS to solve my problem?
In considering using IaaS, one of the most important questions to answer is whether you really need Infrastructure as a Service. Many common problems are supported through SaaS products, many of which campus has already purchased or contracted to be used campus-wide. For example, using Amazon Web Services to run file sharing software might not make sense if you can use the campus supplied Box or Google Drive SaaS options and they fit your need. IaaS gives more flexibility but also incurs more complexity and requires more technical expertise. Now if you have a unique need or a special purpose you should think about the next two questions.
Who is going to configure and support the service ongoing?
Using these services, just like a server in our data center (or down the hall, under your desk) requires technical expertise to setup and maintain. Setting up an account and putting data in the cloud without thinking about long-term support may put your data at risk and may make your service less reliable. As with security, IaaS solutions offer more flexibility but puts more responsibility on the user of the service for maintenance and updates.
What kind of information or data do I need for the service to work?
Example: Amazon and Google both have IaaS services that allow for general storage of data. IaaS services requires the individual using their service to take on more responsibility for setting up their service securely and maintaining the security of their data. This is different than many SaaS services, where much of this work happens behind the scenes and is not controlled by the user. This distinction is important to keep in mind when considering using an IaaS service.
As with SaaS services, where your information is coming from and who will keep it up to date are important questions to answer. For example: a custom application running on AWS may need UIDs or user emails to work properly. Having accurate data is what makes these services useful. Keeping this information up-to-date can be time consuming if not properly planned out. If you are handling information that is sensitive - student personal information, research data, grades, social security numbers - you should contact the Information Security and Policy team before moving data to a new service. Some campus services have already been vetted with guidance on what types of data can be stored or processed in them. As an example, view the UC Berkeley Box and Google Data Use Agreement.