During last weekend’s frenzied holiday shopping, online retailers pushed the limits of their databases. The strain looks likely to continue during the next four weeks, as retailers continue to promote discounts and incentives in virtually every combination possible.
One way online retailers can prepare for those stressful traffic loads, and handle seasonal jumps in activity, is to optimize their databases. I spoke with Paul Campaniello, vice president of global marketing at ScaleBase, to discuss how the implementation of practices, including distributed databases, can help prevent ecommerce disaster.
SPfR: The holiday season is in full swing. Cybersales were up this year. Have performance expectations changed in recent years?
Campaniello: Performance is critical. So much of what we do is making sure [performance goals are met]. Today, there are many next-gen apps out there — not the traditional Oracle, SAP, and PeopleSoft — but apps like Facebook and Twitter, which are Web based. Our solution is built on MySQL DB, just like Facebook.
SPfR: What do solution providers need to be mindful of when working with retail customers?
Campaniello: What happens when everyone, for example, goes to Zappos and clicks on the same pair of $400 Jimmy Choo shoes? You click into or out of the cart, and you “get the hourglass.” If you “get the hourglass,” that’s a database issue. When you try and optimize a transaction, what you’re really trying to manage is the scale and flow, power and speed of the communication between the application and the database.
SPfR: Some might wonder, why not over-provision? Better safe than sorry, no?
Campaniello: We just did a Webinar and talked about SPRAIN pain: Scalability, Performance, Relaxed consistency, Agility, Intricacy, and Necessity. The idea is that when you stretch a database beyond its normal capacity, SPRAIN will stretch the database. You can solve the problem by throwing hardware at it. But if you don’t look at ways to scale your database, you’re not fixing the underlying problem. You want to find a way to scale your database.