Providing a Better Online Shopping Experience with Faster Cloud Performance
Online shopping has increasingly taken over market share from brick-and-mortar stores. But the COVID-19 pandemic accelerated its popularity. With more customers browsing and buying, online retailers have had to boost their site’s performance. Any lag time in the checkout process can quickly become a missed sales opportunity as customers will move on to the next site to find the items they want to purchase.
Along with this need for faster performance has come a rise in cloud adoption. The cloud allows a retailer’s IT team to hand off the day-to-day system management responsibility to the cloud providers so they can focus their time on more valuable projects that affect the bottom line. On top of that, the cloud opens retailers up to exciting new technologies that they just can’t take advantage of through the use of on-premises data centers. These include the use of Augmented Reality (AR) applications that let customers “try-out” products at home such as IKEA’s app that lets you see how furniture looks in your home or Sephora’s app that allows users to try out different shades of makeup. These also include applications for better inventory and Point of Sale (POS) management that make the buying process seamless for customers.
Yet, for all its promise, retailers are still hesitant to move some of their workloads to the cloud. Mainly their mission-critical databases and applications. These workloads are too important to fail, which makes the process of refactoring them to be cloud-native unappealing. These workloads tend to be too big and complex, meaning they already require an extremely high level of performance. And during times of peak customer activity, such as the busy holiday or back-to-school shopping seasons, this already high level of performance will need an additional boost. However, since the cloud is a shared environment (unless you adopt a single tenant which is way too expensive for many customers), the cloud vendors put throttles on the amount of speed you can achieve. And these thresholds may not necessarily meet retailers’ high-performance demands. To add insult to injury, the pricing model for performance in the cloud is directly tied to the number of cloud resources being leveraged… which bloats cloud costs.
These workloads – such as Microsoft SQL Server, Oracle, and Exadata – are sometimes referred to as “untouchable” workloads. That’s because they seemingly can’t be successfully moved from on-premises datacenters to the cloud. This is especially problematic if there is a corporate mandate to go cloud-first.
This is something that a home goods and furniture e-commerce company recently dealt with. The company had a corporate objective – and a strict deadline -- to get out of all on-premises data centers. The team had moved its 40 Microsoft SQL Server hosts to the cloud but kept running into throughput limitations. They knew they didn’t have time to refactor to be cloud-native before the deadline was up. But they needed to get faster performance so their hosts would run.
The solution was to introduce a cloud platform to the company’s cloud infrastructure stack, acting as a virtualization layer between the workloads and the underlying cloud infrastructure. With a cloud platform, you can get a significant boost in the amount of performance that you get through the native cloud alone. Cloud platforms also offer enterprise data services – something that is standard in on-prem solutions but isn’t natively available on the cloud. These data services – such as zero-footprint clones, deduplication, and thin-provisioning – help keep cloud resources to a minimum. In turn, this helps keep cloud spending in check.
After introducing a cloud platform to their stack, the retailer was able to boost their Microsoft SQL Server workloads on the cloud to be 0.2ms faster than they had previously been on-prem. The platform also offered them 3.2 GB/s throughputs per SQL host. And on top of all that, with data reduction data services, the company was able to reduce its data usage by 3:1, helping them keep their cloud costs in check.
If your databases on the cloud are suffering from a lack of performance, stop by the Silk booth (booth 2071) at The Smart Retail Tech Expo. We’ll be happy to give you a demonstration of how the Silk Cloud Platform can give you up to 10x faster performance for your largest, most complex workloads in the cloud.