As U.S. consumers wind down from work for the Thanksgiving holiday this week, many of them are also gearing up for mania that is Black Friday and Cyber Monday.
Cyber Monday, in particular, is expected to draw its biggest online crowds yet this year as a new survey from PriceGrabber reported that 41 percent of consumers plan to shop online one week from yesterday -- up from 37 percent in 2011.
"What matters is being prepared and keeping things as simple as possible -- complexity is the enemy of uptime," Lodge said.
While many retailers are competing with each other to offer the best deals and the best customers, underneath all of that, they really have to ensure that their back-end infrastructures can withstand such heavy loads and pulses of traffic.
As many businesses have jumped to cloud adoption in the last year, Cyber Monday this coming November 26 could be the biggest test yet for many of those cloud infrastructures. Even though many of these systems should have been securely in place by now, experts do have some last-minute tips and advice about how to keep the cloud afloat during the holidays.
"What matters is being prepared and keeping things as simple as possible -- complexity is the enemy of uptime," said Mathew Lodge, vice president of Cloud Services at VMware.
One strategy being employed increasingly to handle sudden large amounts of sales activity is cloud bursting, in which an organization uses temporary capacity from a cloud provider to augment their own infrastructure.
Chris Sharp, general manager of cloud and content at Equinix, noted that "you have to look at where the traffic is coming from" when you are crafting a inter-connected cloud strategy.
"You don't have to build everything yourself. Now you can pick and choose partners," Sharp reminded, adding that having an ecosystem of partners is a major competitive advantage.
"You don't have to build everything yourself. Now you can pick and choose partners," Sharp reminded.
Lodge suggested to look for cloud providers that can guarantee server and storage resources for every new virtual server while offering high-bandwidth I/O interconnect. He also posited that running the application across several providers is a useful hedge against capacity and performance issues with a single provider.
"Because many retailers are likely to need high capacity at the same time, it also makes sense to use several compatible cloud providers," he added.
Sharp emphasized the importance of distribution, explaining that leveraging a multitude of networks should be able to both reduce latency and and redirect traffic to a fail-over location in another geographical location when spikes occur.
"Distribution is king, and distributing your infrastructure is one of the more critical parts of disaster recovery," Sharp asserted.
Equinix reps also cited a Forrester projection that retailers will pull in approximately $68.4 billion in online revenue -- a 15 percent increase over last year -- this holiday season. Thus, any cloud down times could cost retailers big, which is why it is so important to invest in solid and scalable cloud infrastructures now.
We'll certainly know more about how well current cloud strategies can withstand sudden and large bursts of traffic after this weekend, which could provide lessons for the rest of the holiday season and all-year round going forward.
"Every industry has its spikes, whether it is Black Friday or tax time," Lodge remarked, "While public cloud is particularly effective for unpredictable demand, remember that lots of other tenants have the same spike issues as you do."