Cost Savings From Cloud Database Solutions

CTOs/CIOs, you are overspending by 80% on SQL Servers. I met with a Fortune 500 CIO last month who was frustrated with his SQL Server costs. It was a frustration we’ve heard over and over and over again. They were looking at hundreds of thousands of dollars in hardware upgrades. "Our databases are slow," Sysadmins told him. "We need more power." After reviewing his environment, we shocked him with our assessment: His SQL infrastructure wasn't underpowered... it was dramatically oversized. His Azure SQL instances were running 192 CPUs with Enterprise Edition licenses and premium storage. The real issue? Not hardware limitations, but: poorly written queries improper indexing default configurations .. that were driving excessive resource consumption. Within 30 days, we optimized his database environment without changing a line of application code. The results were impressive: His SQL Servers now run on 16 CPUs instead of 192. We switched from Enterprise Edition ($8,000/CPU) to Standard Edition ($2,000/CPU) after confirming the enterprise features weren't being utilized. And we moved from premium to standard storage after fixing the actual bottlenecks. Hundreds of thousands of $$$ were saved. Performance improved by 35% across all critical business applications. Backup times were cut in half. As I mentioned above, this isn't an isolated case. In 20+ years optimizing enterprise SQL environments, we still consistently find the same pattern: The natural reaction from tech people is to add more: Be it CPUs, RAM, better storage… whatever. That should be your last option. Your cloud providers and vendors won't tell you this because right-sizing doesn't help their bottom line. And it’s also harder and more time-consuming to execute. But it’s worth it. Before you approve that next SQL infrastructure upgrade, ask yourself: Do you really need more power, or do you need proper optimization?

From my years at AWS and now at MongoDB, I see the same story unfold. IT teams get excited by the possibilities cloud computing and managed services bring to them. They lift their existing environments and shift them to the cloud. This is a useful first step. It helps them build familiarity with a new operating and governance model, develop new skills, and start to reassess architectures and technologies. But it is just the first step. Next up is to start exploring cloud-native features and best practices that optimize system performance and cost. This is exactly the journey Indeed has been on. As one of the world’s leading job sites serving over 350 million unique monthly visitors and 3.5 million employers across 60 countries, it was running MongoDB at scale — 600 database clusters managing 500TB+ of data. The Indeed team lifted from its self-managed MongoDB environment to #MongoDBAtlas over the summer of 2023. A few months in and the engineering team saw its costs had grown significantly. At around the same time, the company’s management mandated every team cut costs by 19%. The Indeed cloud engineering team turned to MongoDB Atlas Technical Services to help. Over the next few months, the teams collaborated to optimize the busiest clusters, reconfigure underlying infrastructure resources, and ditch expensive 3rd party solutions in favor of Atlas’ native capabilities. Did they hit the 19% saving target? They did….and then some. They actually reduced expenses by 27%. The team at Indeed generously shared how they did it and lessons learned along the way at the MongoDB .local Austin event. Take a look at our blog post recapping their session, along with a link to the session recording: https://lnkd.in/gecEUSyh If you are facing your own cloud computing cost pressures—and let's face it—everyone is way more cost-conscious now than a few years ago, then we can help. Get in touch with me and I can connect you to the best and brightest minds in the business.

"Drinking our own champagne" just yielded $470K in savings. I was visiting a customer in London earlier in my career and I used the term "eating our own dogfood." The cultured Brit I was with politely educated me on the use of champagne over dogfood. It has been in my lexicon ever since. So I was immensely proud when recently one of our engineering teams used CloudZero to save $470,000 a year AND make our software more performant. My initial reaction was to find a way to reward the team by hiring two to three more engineers. How did they do it? Within AWS Lambda, a Python-based file processing library was using up too much memory, and not scaling efficiently. It’s a common story: Infrastructure that works at one level of scale struggles after significant growth. So, they migrated the Lambda function to a faster, more efficient file processing library, anticipating that the change would result in higher performance. It did, and it also resulted in significant cost savings. Comparing the Lambda costs from the month before the change to the month after, we realized a ~$34,000 per month reduction. Annualized, that’s over $400,000 per year! This isn’t the kind of optimization that trawling for commitment-based discounts would surface. This involves richly contextual data, well-informed engineers, and a culture of engineering efficiency — CloudZero’s bread and butter. That’s not all. The same team found another optimization in the same underlying process to reap an additional $70,000 in savings. Visit the link below to learn how we did it. https://lnkd.in/eWXGFKcE #CloudCost #CloudSavings #FinOps