Subscribe to our blog
Thanks for subscribing to the blog.
Slower processing speeds and higher costs happen. And they’re understandable in certain circumstances; but four hours for batch processing in a cloud native application is unacceptable. And four hours is a conservative estimate for the amount of time Cheyenne, an Enterprise Architect at a major sales and marketing company, had spent waiting on her process to finish. She’d even had time to slip out and grab an ice cream while she sat on the sidelines, waiting for her batch to finish processing.
In her role, she depends upon her applications to reliably and quickly process customer data. In order to prepare the data for processing, each CSV tab within a batch must have its own file. There are tens to thousands of tabs in each job. Batch processing is completed using SMB/CIFS protocols within Azure NetApp Files, and each batch takes about four hours to finish. That process uses 60 compute nodes within Azure NetApp Files.
This is part of a series of articles about Azure files
Cost-Effective Batch Processing: Can It be Done?
Imagine Cheyenne’s distress when she received even more CSV files while her application was still breaking down the previous batch. The more quickly she and her team go through customers’ uploads, the more files they can handle, and, ultimately, the more customers they can satisfy. Because her team needed an efficient method to get through workloads faster, Cheyenne needed a faster way to process files in her cloud-native batch application. And on top of that, Cheyenne’s company desperately needed a cost-effective solution to manage the thousands of files they sift through each day.
Azure NetApp Files Speeds Up Processing by 6x
Fortunately, right at the peak of her despair, Cheyenne found Azure NetApp Files (ANF). With ANF, she cut the duration of that agonizing batch process from four hours to just 40 minutes because of its ultra low latency and extreme performance, which is underpinned by a user-controlled system for rapidly changing performance and service level. That’s a 6x reduction in compute costs and processing time.
That figure may seem extraordinary: Why did Cheyenne and her team endure slow compute for so long? What does it all add up to? The decrease to 40 minutes led to an uptick in productivity. Lower costs, more time, and improved customer service. The team’s output has picked up now, as has their team spirit because, let’s face it, accomplishing your goals feels good. It drives you.
And all of this magic? It took place within the Azure portal. Since NetApp is built into Microsoft Azure, the processes around accessing data, provisioning new volumes, adjusting throughput, and even billing have dovetailed into a single, intuitive space.
Life is good.