If You Build It, They Will Come. But Are You Ready?

    August 8, 2015
    avatar

    By Wyndham Sellers

    When you’re launching a new internet service, increased site traffic can crash your site and hurt customer experience. Learn how to handle the overload from these inevitable spikes in demand.

    In the 1989 film Field of Dreams, an Iowa farmer hears the voice of God tell him, “If you build it, he will come.” Following this strange, but compelling message, Ray spends months building his dream, a baseball diamond in a cornfield. But what if, after all that work, there were no seats for the fans, or dugouts for the players? Would spectators stick around with nowhere to sit and watch? Would players be content to sit in the dirt, waiting for their next at-bat?

    When you build an incredible online service, you can expect increased traffic to come your way — what’s critical is that an overflow of site visitors doesn’t hit a capacity wall, downing the webpage. A broken site is the fastest route to a bad digital reputation — the stadium, in other words, has to have a place for everyone to take a seat.

    Planned Recoil

    History is filled with stories of powerful organizations coming up short in this department, subsequently losing revenue and credibility due to huge, unmanaged traffic spikes.

    In October of 2012, the Google App Engine suffered through debilitating downtime as soon as it launched, to the disappointment of users and the embarrassment of the search engine giant. More recently, Obamacare’s healthcare.gov sweated an unconscionable 16-day shutdown as opening-day users poured in.

    Similarly, there’s been plenty of recent news coverage of so-called Distributed Denial of Service Attacks (DDoS), a form of hacking in which assailants crash websites through an artificial site overload, according to Digital Attack Map.

    Generally, large firms have the infrastructure in place to deal with downtime quickly, but it hurts their reputation and revenue stream, and it can take small companies off the map.

    It’s therefore critical to deal with these problems with speed and agility. But ironically, increased exposure can actually cause these issues in the first place. It’s a problem known as the Slashdot Effect — sites can crash when a news story or popular social post unexpectedly directs huge numbers of visitors to a site that otherwise experiences relatively low traffic.

    Traffic spikes like these are difficult to predict and prepare for. However, there are certain guideposts: in the case of most companies, the initial launch of a service will bring a predictable surge in traffic to the entire site. Generally, this will taper to consistent levels, save for marketing campaigns and special news coverage.

    This means preparing your IT infrastructure for the initial blow before throttling back for usual traffic. That way, you avoid hitting a capacity wall and sullying your reputation.

    Bend Without Breaking

    However, the single-most important aspect of preventative IT infrastructure is elasticity. The term refers to the margin of traffic a site can handle beyond everyday levels — the higher the elasticity, the greater the protection. So before launching a service, it’s important to prepare a mature analytics system to gauge expected inbound traffic and determine the desired elasticity. These values will naturally fluctuate over time, and should be adjusted accordingly.

    But, as with any business, it’s best to avoid waste, and it’s indeed wasteful to expend resources on a massively elastic infrastructure that normally sees lower, steadier numbers. As always, flexible processes are better than static ones — the best practice is to re-deploy an elastic infrastructure during spikes, but to put it on hold during periods of routine traffic.

    Getting this right requires a certain degree of analytical power, and is perhaps the most challenging aspect in preparing an IT infrastructure.

    For the analytical and technical expertise you need to carry out this mature plan of attack, contact the seasoned group at TeamQuest. We specialize in simplifying difficult tasks, like calibrating and implementing effective site elasticity. Specializing in high-capacity management services, TeamQuest can optimize your IT infrastructure to industry-leading standards.

    If you build it, they will come — TeamQuest is here to make sure that they’ll keep coming back.

    Image credits: Vicki Timman/flickr)


    Category: capacity-planning