Yes, it’s that time already.
You may not be ready to think about holiday shopping season yet, but if you’re an online retailer or another business that may experience web traffic spikes in a few months, you should review your strategy now. Training recovery staff, testing and fixing any unexpected snags may take longer than you expect, and procrastinating now could mean disaster later.
Anyone can be a victim of a traffic-related crash, especially during the holiday shopping surge. Nearly every year, a high-profile crash takes a retailer offline in the days following Thanksgiving. Last year alone, Macy’s, QVC, Walmart, Victoria’s Secret, and the Gap all saw web outages that likely cost them millions of dollars in sales.
Even if you aren’t expecting a major bump in your regular traffic and your site is usually stable, there is always a risk of experiencing a website outage due to unforeseen reasons. For example, if a publication with a high readership links to your site, the resulting clicks could overload your infrastructure.
What can we learn from these retailers’ mistakes? When you direct people to your site, you must be ready to deliver. You may think you’re ready for a surge in traffic. But are you? Whether it’s a sudden viral surge or peak traffic on Black Friday or Cyber Monday, a frenzy of customer interest can quickly become your worst-case scenario.
Know your risk and limitations
Scale breaks things: This IT expression reminds us that what works on a smaller scale may not translate when stress on the system is multiplied. Your site may be running fine now, but if you’re only accounting for the traffic you expect, the structural integrity of your site may not hold up if your Cyber Monday deals go viral.
Site-rattling changes in traffic aren’t always sudden. If your organization is growing steadily, your traffic can increase steadily, too. While this is a cause for celebration, it’s also a reason to re-evaluate. You may hesitate to put money back into your servers and invest in the operation of your expanding business instead. But by underestimating your projected traffic, you may be flying too close to the sun.
Even with the risk involved, there are plenty of reasons why you may not be prepared. You could be reluctant to update a system if it’s currently working just fine. You may be underestimating your risk or not accounting for future growth. Or you may not want to spend money to expand your capacity beyond the minimum in the outside chance something could go wrong.
The key to preventing sudden IT failures like the ones above is to acknowledge and understand your risk. When you take the measures necessary to anticipate and prepare for traffic surges, you won’t be left with a timeout message for all your visitors to see.
Does this mean you should sink tons of resources into an excess of physical server space just in case? Maybe not, but you should use what you have more efficiently and make it work better for you. By adding flexible options to your existing systems, you’ll have backup when your site is under stress without draining your budget all year long.
Here are some ways to plan for holiday traffic:
1. Make sure your servers scale.
You might be able to handle the active sessions you predict you’ll get on a regular day. But if your content goes viral, you may not have the bandwidth you need depending on your active session capabilities, and you could fumble once-in-a-lifetime opportunities. Auto-scaling elastic computing for your website’s front end can help handle unexpected surges by growing and shrinking on demand to fit your needs. This way you’ll only pay for what you need, when you need it.
2. Diversify your locations.
Consider building a content delivery network (CDN) from a provider that allows for automatic scaling services. A CDN will diversify the locations of your servers (when it’s closer to your visitors, it limits geographic hops and they’ll get your content faster), and it will cache your static content for easy access.
3. Prepare for more concurrent sessions than you’d ever expect.
Putting too much pressure on one server can lead to crashes or slower response time. Load balancing software can help reroute lots of traffic to servers that can handle it. It acts as a hall monitor, making the flow of traffic more even. This will keep things level and won’t put too much stress on one particular server.
4. Test under pressure.
Want to see if your site can withstand a surge? There’s no better way to check than actually trying to break it. By taking the time to put real pressure on your servers and push their limits, you’ll see how your site reacts in real time without taking the risk during peak traffic times.
Invest in peace of mind
When your site is ready to handle the traffic you’ve attracted, and when you’re ready to adjust for more growth down the road, you’ll be able to enjoy it and not live in fear of a crash. That way you can spend your time delivering your content seamlessly—and focusing on your business during the holiday season.
Sungard Availability Services specializes in business continuity, information security, IT disaster recovery and cloud systems.