"How can there be an extra 30% overhead in applications like Apache Spark that other optimization solutions can't touch?" That's the question that many Pepperdata prospects and customers ask us. They're surprised—if not downright mind-boggled—to discover that Pepperdata autonomous cost optimization eliminates up to 30% (or more) wasted capacity inside Spark applications. And that's on top of and in addition to conventional resource optimization techniques such as Cluster Autoscaling, Instance Rightsizing, Spark Dynamic Allocation, and manual tuning of applications.
A Long History of Customer Success
At Pepperdata we take this skepticism in stride, since we've been encountering it—and overcoming it—for the last decade, from leading enterprises and innovative startups alike.
In those last ten years, Pepperdata Capacity Optimizer has been operating and hardened in some of the world’s largest and most demanding compute environments running data-intensive workloads. By keeping these complex and highly sophisticated data clusters perfectly in balance, day in and day out for years, Pepperdata has pioneered and refined a unique and industry-leading set of features and expertise that reduces costs for data-intensive workloads an average of 30-47% with no manual tuning or application code changes.
By autonomously optimizing cluster CPU and memory in real time, Capacity Optimizer empowers customers to extract the most from their investment in large-scale compute environments, on prem or in the cloud.
For over a decade, Pepperdata's entire technical team has been singularly focused on providing the most robust solution for compute resource optimization. That's a level of focus, dedication, and diligence that benefits every customer.
And the results of these efforts are immediately apparent within minutes of installing Capacity Optimizer. Pepperdata's dashboards clearly visualize the immediate ability of Capacity Optimizer to recover 30% or more of wasted compute resources.
Here's one example screenshot from a relatively small cluster that generates over $30,000 in annual savings. This chart shows this cluster's total dollar savings per month, which varies from month to month based on usage of that cluster. The “Cost Savings” heading at the top summarizes the savings: $31,174.70 over the course of a year.
Figure 1: Cost savings achieved by a single cluster running Pepperdata
For many of our large-scale enterprise customers deploying Pepperdata across thousands of nodes, the savings multiply exponentially. For such customers, including those in the table below, Pepperdata Capacity Optimizer is like receiving a check for $5,000, $10,000 or more in the mail every day.
Figure 2: Daily and annual savings achieved by a sampling of Pepperdata customers
It's gratifying for us at Pepperdata to see the amazement, surprise, and even incredulity in our customers' eyes when they first take a peek at our dashboards and realize their savings potential. We are thrilled when customers save three to four times (or more!) the initial sum invested in autonomous cost optimization. And our ROI guarantee makes everything about working with Pepperdata so simple for our customers: you don't pay if you don't save.
Figure 3: Pepperdata customers achieve a typical ROI between 300% and 600%. Customers who spend $1 now on Pepperdata recoup $3-6 later via a reduced cloud bill.
Transforming Skeptics Into Happy Customers
Pepperdata has been transforming skeptics into happy customers—even evangelists—for over a decade. We are trusted by industry-leading enterprises and innovative startups alike and run every day in 20,000 nodes across the planet.
One such satisfied customer is Extole, who came to Pepperdata after spending several years optimizing their cloud environment for cost and efficiency. Because of the years of effort that Extole had already invested in tuning their environment, Pepperdata's initial pitch naturally met some of that familiar "sounds too good to be true" skepticism. However, within days of trialing Pepperdata, Extole achieved an additional 30% savings on top of their other optimization efforts to date.
In the words of Ben Smith, VP of Technical Operations at Extole, "There’s no way there’s a 30% overhead that everyone is paying… and yet you see it in the numbers that it's true."
Transforming Happy Customers Into Evangelists
A US-based division of a leading global financial services enterprise deployed Pepperdata Capacity Optimizer to achieve a 42% cost savings by reducing the number of instance hours for their deployment of Apache Spark on Amazon EMR.
Pleased with their savings, the US division highly recommended Pepperdata to their colleagues in another country. Based on this strong guidance, and success stories from other Pepperdata customers, the overseas business unit completed a Proof of Value (POV) and chose to deploy Capacity Optimizer on 1,600 dev clusters and over 1,000 transient production clusters in their Amazon EMR environment, which lowered their costs by approximately $100K per month. These results were realized immediately and automatically, without any code changes to Spark applications or any manual application tuning.
One happy customer shared their experience with Pepperdata, and that led to another happy customer. As more customers share their stories and experiences, the concept of eliminating unused capacity in Spark doesn't have to sound "too good to be true"—it can be a reality for companies seeking to optimize the price/performance of their Spark applications.
See (or Try) It for Yourself
Pepperdata Capacity Optimizer autonomous cost optimization delivers 30% or more greater cost savings for data-intensive workloads such as Apache Spark with no application changes. Pepperdata pays for itself by immediately decreasing waste, increasing Spark utilization, and freeing developers from constant manual tuning to focus on revenue-impacting innovation.
You, too, can eliminate the "too good to be true" skepticism by trying Pepperdata Capacity Optimizer for yourself! Request a free demo to see how it works or experience the savings directly with a free trial.
