Headline for article with cartoon and scary text saying Website Outage

Low-Cost Load Distribution for Static Content

Ever wanted a more reliable website but couldn't afford to upgrade your static servers? Here is a straightforward approach for distributing server load across multiple data centers to supply greater business website availability. Notice, this is a low-cost load distribution system, not load balancing.

What's the Foundation of Any Content Distribution System?

What this method does is divide incoming traffic between several servers using Domain Name Services (DNS). DNS is like a phone book converting common domain names into Internet Protocol (IP) addresses that stand for your server.

For example, www.adbriefing.co.uk resolves to n.n.n.n address via DNS, where n.n.n.n is an IP address owned by our hosting provider. [In reality, this website resolves to both IPv4 and IPv6 addresses, which is beyond the scope of this article.]

By serving your static content from different servers, there is less load on any one server. Too many business websites go down when they are fortunate to have a surge of traffic. When internet marketing works, it only works when a website can manage it.

Because you really can't control how much bandwidth is available at your hosting provider, this method supports a data center level distribution of requests. If you have significant volume, you may still need host-aware load balancing at each data center.

It's a Poor Boy's Content Delivery Network Across Static Servers

This model is great when you want up-time and aren't ready to invest in hardware. Software solutions are available for under a thousand a month to manage 1,000,000+ impressions, depending on the density of your content.

Over the years of hosting hundreds of websites, I've used this with clients who didn't want to invest in content delivery networks. It was more efficient to synchronize files between multiple physical servers for static assets like images, JavaScript, and Cascading Style Sheets (CSS).

Using DNS round-robin with failover and monitoring, you can improve your overall availability while spreading around requests across several data centers. Think of this as the step before using a content delivery network with multiple service points.

Editor's Note: At the time of first publishing this article in 2008, AdBriefing Copywriting Tips used DNS Made Easy for static pool round-robin with failover and site monitoring. It's a cost-effective solution that still works in 2021. — Samantha

You'll find this the easiest to manage with primarily static content. The configuration can use a script that does file distribution on the back end, redirecting to a primary site if files are not found. Your own private cloud service like Resilio File Sync can move even multiple gigabyte media files.

Core Concepts Around High Traffic Media Sites for Reliability

While this is not a technical article, this method works on both Apache and IIS virtual servers (dedicated servers are faster but will cost more.) I've even used it for migrating from shared servers. The key to understanding is DNS round-robin, or anycast is the core of supplying a clean IP address to a DNS request for your domain.

Many low costs, even free solutions are available, including CloudFlare for low-traffic websites. The problem comes when you have significant volume, AND you want to prevent a third party from throttling your users. If not set up correctly, CloudFlare will block secure, proxy, and foreign traffic.

A great case study is how Amazon AWS suspends Parler (a social media network) running up to the 2020 US Presidential Election. Even Parler paying $300,000 a month for hosting, Amazon AWS bans them at the whims of a Change.org petition.

Shadowbanning or throttling traffic is another real possibility that affects marketers. For years Google has omitted direct marketing style health, fitness, and income opportunity information publishers. Thousands of quality publishers find Facebook, LinkedIn, and Google not serving their advertisements.

As a copywriter, you don't want your client to be invisible. Especially when it makes your copy look like it doesn't perform. DNS load balancing may be a foreign concept, yet this article aims to give you some background to ask questions to know your content gets delivered.

What's the Right Design When You Are Starting Out?

Delivering static content via a separate set of servers or content delivery network requires your website to have the right kind of design. The WordPress publishing platform helps with this; however, most shared hosting services don't make it easy.

If your application separates static content from dynamic, you'll be able to still handle database transactions. For example, this low-cost load balancing delivers images, CSS, and JavaScript, while a single dedicated server manages requests.

My clients with millions of unique visitors daily clone both front-end servers and back end. They may have 5 to 10 static server clones behind a load balancer or DNS round-robin. Each server is configured to pass requests to a single primary copy if the content isn't found locally.

Meanwhile, every hour servers synchronize while raw logs merge onto a monitoring server. It doesn't have to be complex to work, but it could mean hosting across multiple service providers globally. Every part of this is transparent to the end-user of the website.

With Resillo Sync, you can have real-time read-only content on your forward-facing servers while that master server is behind a firewall. Remember, a surge of traffic isn't your only concern these days. A competitor can start a denial-of-service activity as quickly as a malicious hacker.

Because this approach doesn't require special software or hardware, you'll be able to implement it starting small. As your site grows, you can include up to 12 locations with DNS round-robin, each with load-balanced back ends.

Questions For the Webmaster to Make Sure Your Content Gets Delivered

For publishing start-ups and existing media operations, you need the right questions. These questions help you stay ahead of both changing technologies, as well as changing political climates. Here are critical questions for your webmaster and hosting services.

  • If a critical service fails, then what is your escalation path to delivery business continuity? Meaning, if anything fails, what is your webmaster doing so you know about the failure. Still, the end-user doesn't have any impact.
  • Suppose a third-party vendor goes under or decides to stop serving your business. What is the plan to migrate to an equivalent service? No-fault of your own, a vendor can close, leaving you with a gap in service to your visitors.
  • What is the most significant surge of traffic these servers can manage with the size of your marketing campaigns? Most servers cannot handle a campaign that goes viral, especially since paying for surge volume you don't use can be cost-prohibitive. Know your limits.
  • Are there seasonal variations in traffic that are normal? What would extraordinary traffic look like? Knowing these thresholds helps you plan added service levels during peak season, allowing you to scale back to reduce costs. We did this at Lillian Vernon to prevent their call center and catalog website from failing during the holiday rush.
  • What is the evidence that no single point of failure exists in your hosting environment? Don't take your vendor's word for it; evaluate your publishing platform often. It only takes a single rogue plug-in to compromise the integrity of your site. Plan for these failures in advance.

To be fair, there are hundreds of questions that form your Business Continuity Plan (BCP). These include questions about credit card merchant providers, consumer privacy, and other risks associated with online publishing. It's easy to start a website; the hard part is keeping it fully available to your customers.

Contact our offices when looking for low-cost distribution services that improve up-time while enhancing visitor's' experience. There is a bridge between high-cost content delivery networks and where you are now to save you money.

© 2008-2021 Hitt Publishing Direct & JWH Consolidated LLC, All rights reserved.

Justin Hitt has more than twenty years of experience with technology risk management, including earlier certifications with operating systems and load balancing hardware. Clients include IBM Global Services, AT&T, Lillian Vernon, Dominion Enterprises, and many high-volume hosting providers. If your publication, social network, or media platform is at risk of outside attacks, banning, or service disruptions, then schedule your no-obligation paid consultation today.

005_I0720A