Alibaba’s Tmall Global now features goods from 14,500 overseas brands, 80% of them selling in China for the first time.
Using Internet-hosted technology can be risky, but it’s worth taking steps to ensure its reliability.
For retailers and others using “cloud” services to power their web applications and store tons of web data, Amazon Web Services, the technology services arm of Amazon.com Inc., has emerged in recent years as a prime cloud provider. So when one or more of its cloud-based data-serving web servers go down, as some did in late June as strong storms battered the eastern United States, it raises concerns about the viability of tying a good part of a merchant’s technology to a data center and web server farm operated by another company.
Indeed, the complexity alone of these outsourced data centers and server farms should give their users reason for concern. Amazon Web Services reported July 2 that the June 29 outage occurred in its Northern Virginia facilities when the storm caused a “large voltage spike” to hit the data transmission switching equipment in two data centers. As expected, the switches in one of the data centers transferred power without a problem to back-up power generators. But for some reason that Amazon did not publicly explain, the switches in the other data center didn’t successfully transfer power to its back-up generators—even though the data centers were using switching gear and power generators installed within the past two years and tested weekly, Amazon says.
This all goes to show that, even in a cloud-services set-up as sophisticated as Amazon’s, things can go wrong, leaving web sites without the power they need to function properly and serve customers. And with some web sites using the services of more than one cloud technology provider while connecting with consumers in more ways, including mobile commerce and social networks, their exposure to occasional outages is even greater.
But that’s not to say the cloud doesn’t offer a good option for supporting web site technology.
As retail web sites become more complex with interactive applications, and with more and more data to store, cloud-based computing and storage systems offer a cost-efficient way for web site operators to run their sites and scale up with computing power and data storage as necessary. The key thing is to determine what makes sense to operate in the cloud. If an application is mission-critical, and losing it even temporarily might drive customers away forever, it may make more sense for a site operator to run things in his own backed-up data center or to arrange for special back-up from his cloud provider.
Then there’s also the management of how data are processed and transferred among multiple web servers and data centers. With effective technology process management, online retail sites can arrange to have cloud computing power automatically switched to an alternate source to minimize the chance of experiencing a service interruption. “As retailers build out computing architectures to meet evolving business challenges, I.T. complexity grows, and enterprises can’t count on EC2 or any one environment to be 100% available,” says Randy Clark, a senior executive at UC4 Software, a company that provides software for managing data flow in web technology systems. EC2, or Elastic Compute Cloud, is Amazon’s cloud-based computing service.
But with the right mix of strategy and technology on the part of users, cloud technology systems can serve retailers well and work through unexpected problems.