Cloud has become the backbone of modern business, promising agility, scalability and resilience. But as adoption grows, many organisations are discovering that running workloads in the cloud is far from a free ride. Never mind cloud bill shock, the bigger issue is that without careful management most of that spend achieves very little.
“Oftentimes, cloud spend is just a number to people,” says Trent Cameron, CIO of TrueMark. “But real visibility is about understanding what’s behind that number, where it is coming from, which departments the spend is coming from and ultimately, what value it is driving for the business.”
For many executives, security used to dominate discussions about cloud adoption. Today, it’s a given. It’s built into the way cloud platforms operate. Cost, on the other hand, can be unpredictable, and can easily disrupt decision-making and long-term planning. For Cameron, this comes down to guardrails.
“When we don't have guardrails in place, we're not sure what people are doing,” he says. “That leads to the second thing, which is lack of education.” A lot of the time, speed comes before cost, and artificial intelligence (AI) is a great example of that. “Everyone wants AI in production tomorrow without following the chain back,” Cameron continues.
“It gets very expensive if everyone, instead of doing a search across an elastic database, is running queries through AI.” When every query costs 30 cents instead of fractions of a cent, that’s when costs begin to balloon. “It’s a lack of understanding of what things cost from both an engineering and management perspective,” he adds.
Undifferentiated heavy lifting
One of the strongest arguments for cloud is elasticity – paying for what you use, when you use it. Yet many organisations still run workloads continuously, as though they were operating in a traditional data centre. “The whole reason the cloud makes sense is not because it is cheaper as a whole. Anyone who has lifted and shifted will tell you it was more expensive,” he says. “Treating cloud as a data centre doesn’t take advantage of that elasticity.”
Cameron talks about development teams who, as an example, only work eight to ten hours a day. The problem is that their databases are running 24/7 which means the business is paying 24/7 and will continue to do so “because the tooling isn’t in place,” he adds. Developers aren’t necessarily concerned about cloud spend because it is something they’re not measured on, and the cycle continues. “It’s undifferentiated heavy lifting… the conversation never happens,” says Cameron. It's about creating visibility that connects technical resources to strategic business outcomes.
One example of how TrueMark has approached this problem is developing AutoState to close this gap. It’s an open-source tool that automates the starting and stopping of AWS resources using simple tags. Without reinventing the wheel, AutoState makes cost control practical, even in complex cloud environments. Cameron explains that while developers know exactly when they’ll be in the office, they often don’t know how to turn off AWS resources when they’re not needed. And even if they do know how, it’s usually not their responsibility or focus because they’re paid to develop, not manage cloud infrastructure.
“There’s a separation between responsibilities. In many regulated entities, even if a developer could stop and start their database, they may not have the permissions to do so,” he says. “But with automation and tagging, this makes it very simple.”
Automated innovation
With AutoState, a developer or DevOps engineer can manually, or through infrastructure-as-code, add standardised tags. “And that’s the beauty of having simple tools that do the right thing,” adds Cameron. “A developer doesn't care if this database is on when he isn’t working on it, and AWS doesn't charge you for it other than the storage. So really, it's a win-win to adopt these types of tools.”
Because AutoState is open source, it also removes the barriers that usually come with cost optimisation tools. Teams do not need licences, training or layers of approval – they can simply apply tags and let automation do the work. This, in turn, strengthens confidence at an executive level.
“Why would a board grant you more if you haven’t shown that you’re being effective with the tools you already have?” asks Cameron. Cost control is no longer just a line item; it is proof that cloud resources are being used responsibly and with intent. That innovation, efficiency and tooling have to be there.
“Do I want a developer sitting around trying to figure out how to start and stop an instance and yelling for permissions? No, absolutely not. Do I want that instance turned off? Yes. As we start to show that efficiency, we're really pushing that innovation,” ends Cameron.
Share