When many academic institutions needed to go remote, teachers recorded lectures with learning management tools for students to watch. Because the platforms were often cloud-based, students could stream the content when it was ideal. This worked well, but difficulties began to arise with cloud providers.
The problem wasn't getting the content into the platform, but rather its storage and distribution. Cloud providers charge for storage and, with platforms ballooning in size as groups of students stream lessons, for example, monthly cloud bills have grown out of control.
Due to the pandemic, many businesses and industries were also forced to move staff and platforms remote. To make the transition quickly, many organizations began to use cloud services. As life slowly becomes more stable, companies are looking at where the resources are, where people want to be and what is best for the company overall.
Remote workers and platforms, which were growing slowly pre-pandemic, have exploded and look to be the standard going forward as companies embrace the cost savings and flexibility for staff and the company. But now IT operations teams face a cloud hangover. Because they are the ones paying the bill, they must also be the ones to reevaluate the tools in place and look to manage cloud costs.
Reevaluate tools and services
First, examine the additional services your organization pays for in its main software as a service (SaaS) or other applications. For example, a local business that doesn't have a global reach probably doesn't need a service like AWS CloudFront that provides worldwide distribution.
Assume in that initial rush to the cloud, not all services or features were identified as necessary. It was quicker to simply check everything, but the time to reevaluate has arrived. With any cloud environment, it can be difficult to understand what is in use and admins might need technical and sales support from the cloud vendor.
One thing that occurs in any IT environment is the ominous warning of repercussions from admins disabling features or functions. Often, these alarms can scare even the most seasoned IT professionals to second guess themselves, so prepare to be on the phone with vendors to work through what each service does and the possible effects of its removal.
Additional services can be the easiest targets for mitigating a cloud hangover, as admins can turn them on or off and see the results quickly. The effect is often limited for tools and services, but not for data storage.
Revisit data storage tiering
Not all data is created equal. In data centers, admins account for this with data tiers. In the rush to move to the cloud, speed might have taken precedence to a data tiering and retention plan. Even if some level of data tiering planning occurred, it's important to revisit as data ages, more data is created and admins better understand data usage.
Most clouds provide data usage reports that give insight into usage and help admins plan for when data moves from the primary tier to the secondary tier, archival and removal. Data tiering isn't just a technical consideration. Admins must involve the data creators or owners to find out the life span of the data, criticality and retrieval needs as it must pair with usage reports.
This can be a challenge for the data owner -- if the data is in place and working, why change anything? Being transparent and sharing costs can help. IT isn't free, so why should IT admins pretend it is? Application owners must understand billing and cost savings and see the financial implications of where the data lives and the different cost levels for performance and capacity.
The other target to cure a cloud hangover is not the additional services to SaaS or other applications, but the main applications themselves. Often large-scale services are purchased as a blanket item to cover all personnel who might use it, not just those who definitely will.
This often occurs in large contracts for video conferencing services or virtual desktop infrastructure deployments from cloud vendors that were done quickly. These items can be some of the biggest cost-saving measures, but one of the worst things to pull away from end users. It doesn't matter if they used it, they are losing it; that is a problem. Some vendors might allow organizations to trim down subscriptions, while others might require companies to restart on a different subscription tier.
Usage reports must drive all cloud decisions. Using a tool or feature only once or twice does not count as active use. This is a tricky line to draw, but teams must draw it somewhere or end up paying for once-a-year usage -- and balloon the bill accordingly. This policy might create tension with staff, but usage reports help IT ops admins determine how to trim subscriptions.
The cloud hangover will be messy no matter the evidence or data. Pulling a service bothers users -- even if they never used it. Be open about what you're doing: Don't remove services in secret, as one or two vocal users can cause a massive headache if something disappears suddenly. The usage reports, combined with billing data, will provide the best argument backing -- and it could be an argument. Be prepared for opposition from users and middle management.