We’ve all been there, ambling through an online checkout process card details in-hand, ready and primed, before you’re given a server error and 404 page. As a customer it’s infuriating and for the business owner it is even more so. Site downtime costs money whichever way you look at it. Recent rough estimates point to around $26.5 billion in revenue lost a year from IT downtime. So, how do businesses prevent and manage server and website crashes?
Before digging in it’s important to realize that the root cause of downtime can be a variety of things; ranging from poor hosting to crash-inducing traffic spikes (not necessarily a bad thing). This was famously the case with Apple who’s website went offline owing to the enormous volume surrounding the release of the iPhone 4S.
With Black Friday and Cyber Monday recently passed and Christmas on the horizon adequate preparation is essential in the avoidance of site downtime. Retailers and Managed Service Providers alike will be gearing up for the expected upsurge in traffic, so let’s take a look at the ways they negate possible downtime cutting into their profits.
1) Good hosting
For hosting, key stakeholders and business owners are required to take stock and understand the exact demands placed on their e-commerce. Correct planning and foresight will make them acutely aware of the precise limitations and boundaries of their existing hosting package. Particularly successful businesses can experience such monumental growth year on year that they soon discover they have considerably outgrown their current hosting requirements. The problem with exceeding capabilities lies with the shrewd hosting companies imposing hidden caps and clauses for such events. This results in spiralling hosting costs or site downtime; neither of which are desirable options.
Smart businesses will avoid getting caught out in this trap by negotiating with the respective web hosting company to secure the best deals on their uptime policies and guarantees. Terms for negotiation will include credit towards hosting fees if the site goes down, expected usage caps and clarity around the much mooted term ‘uptime’.
Tying into hosting service capabilities is that of servers. It may be true for many online companies that for 80% of the year their shared server hosting capabilities are more than enough for their customers’ needs. Some businesses opt for shared servers because, whilst they may be slower than dedicated servers, they are also much cheaper.
Where the critical decision lies, is in deciding whether the saving is worth it when the peak period hits and sales begin to flag due to technology issues. CS Monitor reported that some retailers can make up to 40% of their annual revenue in the Black Friday/Christmas peaks of November and December. Companies with comparable annual peaks may well see reason to change from shared servers.
3) IT Monitoring
To IT managers out there the saying ‘take your eye off the ball’ may well be a familiar one. One minute everything is in a state of Zen, the next all hell has broken loose. Luckily, there is software in-place to prevent such apocalyptic scenarios from coming to fruition.
The tools in question are IT Monitoring Systems and application performance monitoring tools. Proactive monitoring enables the best businesses to head off issues before their customers notice and profits begin to nosedive.
An effective monitoring strategy allows the stakeholder(s) knowledge of how each one of their business critical IT components are performing and how their availability will impact the business. Top-of-the-range monitoring software encompasses custom fields, customisable dashboards as well detailed up-time reports, regular status updates and of course failure alerts. If a company is having an influx of traffic on Cyber Monday and their servers are on the edge taking the strain; their monitoring software is going to let them know well in advance. This drastically minimizes the risk of potentially critical problems putting their margins at risk.
Yes, businesses need to do housekeeping too especially if they host on-site servers. A fairly traditional site in the common server room tends to resemble something like this:
Now, let’s compare that with Google’s data centre:
Google is an extreme example that illustrates the complexity yet necessity of organizing servers. At the best of times server rooms are hot, stuffy and complex places to be. By merely keeping an area well ventilated and circulated companies can stave off potential losses.
Simple steps taken to prevent overheating and crashing include:
• Keeping the server room tidy to increase airflow
• Arranging in rows so that hot air is expelled out the back
• Ensuring cold air reaches the equipment whether with an A/C unit or enclosure
• 77. 77 degrees is regarded as the maximum temperate a server room should reach
• Dusting and cleaning prevents overheating
Content Delivery Networks consist of a distribution of servers deployed in multiple and various data centres all over the web.
Why do businesses use CDNs?
Because time is money. In the online world their content is just milliseconds away from users across the globe, however online customers aren’t the most patient of people. Any delay in loading and performance issues can be enough to turn a potential customer off. CDNs help to serve content to end-users with high availability and equally high performance. This improved user-end performance manifests itself in a variety of mediums. Some files may be pre-cached meaning that there’s a high probability that when visiting a page someone will have already visited the site using the Google CDN. Therefore the file is cached and the browser won’t need to be downloaded again. Other benefits include a distribution of loads, saving greater bandwidth and reduction of existing hosting costs. In the competitive sphere of online retail every advantage helps, utilising CDNs enable many websites to perform to over and above levels expected by their visitors.
The thought and planning that goes into creating seamless online business systems is staggering. Successful companies are as well prepared as they are smart; any potential event will have set testing schedules and simulators in advance of launch to mimic worst case scenarios. Every minor detail will be scrutinised with the financial heads to calculate how much a minute/hour of downtime would affect the business. Amazon’s two hour downtime in 2008 not only cost them a reported $3.6 million but their share price also fell by 4.1% on the day- the consequences can be huge.
Comments on this article are closed.