This year, more holiday shoppers are expected to shop online, beginning before Black Friday through to the pinnacle of online commerce, Cyber Monday. While an increase in online shopping is good news for anyone looking to avoid the holiday rush, CIOs are rightly worried that their networks may get trampled in the online aftermath. The effect of millions of customers hitting your website at the same time can have a crippling effect on network performance, resulting in unplanned downtime, sales delays and other system hiccups at the most critical and costly time of year.
CIOs counting on the Cloud to come to their rescue may be in for a rude awakening on Tuesday morning. Cloud platforms are designed to optimize data center resources—servers, storage, applications—but not the network that interconnects those data centers. Just as a chain is only as strong as its weakest link, a chain of data centers is only as strong as the network that connects them. The reality is that many wide area access networks, connecting Data centers to each other and enterprises, may not be up to the task of Cyber Monday—particularly those that aren’t taking advantage of Software-Defined Networking (SDN) technologies.
CIOs have traditionally focused on the problem of bandwidth when facing Cyber Monday, as in: Do we have enough of it? The answer to that question was usually found in the form of additional leased MPLS lines that sat idle for most of the time and only came into play during peak workloads. This proved to be an expensive solution, however, leading CIOs to look for a better way to handle high traffic demands. Using SDN technologies, enterprises and Cloud service providers can now control the flow of traffic between data centers using application-aware intelligence that routes traffic based on priorities and existing service level agreements (SLAs).
The need for application intelligence in the wide area network (WAN) is based on the premise that not all business applications are created equal. For example, real-time Unified Communications in the Cloud has a lower tolerance for latency than a once-per-day data backup that might be a high-bandwidth consumer, but not a high priority. Because network bandwidth is not unlimited, enterprises and Cloud service providers need to intelligently manage how they allocate network bandwidth to different applications over a common infrastructure.
As a corollary to the idea that not all applications are created equal, not all enterprises give the same weight to the same applications. For example, let’s say a global retailer is gearing up for the Cyber Monday rush. Customer access to product data is clearly a priority for them, but so is bandwidth between the web ordering system and the back-office systems used for order processing. Yet, each of these systems may be in different data centers. Optimizing their data center interconnection using a solution that leverages SDN would allow the retailer to ensure that both types of traffic are given high priority versus non-critical data traffic, resulting in faster order processing and a better customer experience.
Network analytics has a role to play in this as well. The ability to analyze ongoing traffic patterns to alter routing intelligence in real-time offers a clear advantage, particularly from the perspective of SLA enforcement. As the analytics get better, network operators will get closer to a self-sustaining network that can police itself.
Of course, an optimized network isn’t just something for a CIO’s holiday wish list. Enterprise networks handle mission-critical traffic every day of the year, whether it’s customer service or e-commerce. In fact, it’s the mix of different kinds of daily traffic—such as balancing e-commerce alongside click-to-chat and WebRTC sessions—that most clearly demonstrates the value of an SDN solution that controls the underlying network resources.
But this time of year especially, enterprises and service providers need to step up their network performance if they want to avoid getting stepped on in the holiday rush.