On the internet a certain piece of information published may go viral when one least expects it to. This phenomenon, known as the 'slashdot effect', may lead to an unprecedented hike in web traffic influx. This may cause the website to slow down or make it inaccessible. If a digital publication's infrastructure is not equipped to cope with this sudden surge in traffic, their readership will suffer.
Auto-scaling: The need of the hour
Modern digital publishers need to reconsider the tools they currently work with. As a famous DevOps quote goes, “Think of infrastructure as cattle, not pets”—it holds all the more importance for the digital publishing domain. To put it simply, the moment a certain infrastructure starts looking unhealthy or outdated, regardless of whether it is the TATs, cost-effectiveness, scale, or analytical capability etc., it needs to be replaced immediately to meet the current demands. In such a case, everything needs to be duly backed up to ensure a seamless transition to a newer version. An auto-repairing infrastructure can therefore be an effective solution.
Publishers need to be well-equipped with the right tools to constantly monitor their system’s health and exercise their ability of automatically setting up new infrastructure. For instance, the system needs to be designed in such a manner that will avoid cascading failures. This means that in case a problem arises in a particular search page, the user should be able to access the rest till the problem gets sorted.
Therefore, even if the trending stories section remains down owing to an error or a massive user traffic, the system should allow users to see other sections of the page instead of killing the page entirely. This would allow publishers to fix the issue and manage the traffic seamlessly without losing a major chunk of their audience.
The ‘Chaos Monkey’ model
In 2010, Netflix brought forth a model wherein it randomly created a ‘chaos’ by crashing its own hosts to validate whether its auto-remediation worked. This pattern of enforcing failures via random termination of services within the architecture, known as ‘Chaos Monkey’ now, was pioneered by this America-based media service provider to test the stability of its system upon its migration to the cloud.
The model guided Netflix towards better automated solutions to cater to situations wherein critical components of its production service infrastructure could be taken down. Following suit, many organizations across the globe have had a positive experience with the Chaos Monkey pattern, making them better-aware about all areas and the extent to which their systems can possibly be impacted.
Therefore, this approach has been significantly helping publishers cater to the issues that they encounter in the production service infrastructure, because that is where the real problem lies!
Today, publishers need to be fast-paced and must possess the ability to handle massive amount of user volume and simultaneous log-ins, etc. For this, their Content Delivery Network should be able to absorb maximum load and consequently, it needs to be scaled first. CDN's are a very essential aspect of differentiation for any publication, and equipping the network with new-age tools such as Data Analytics, AI etc., will give it more strength and flexibility to handle sudden surges.
For instance, some renowned media companies have recorded a cache hit ratio as high as over 95%. This enables a piece of viral content to keep going higher, while bringing the server load down to as much as 1/20th of the load. The cache hit ratio typically varies from 60% to 99% in some cases, increasing to reach an average peak of 90-95% with automation while bringing the backend load down by half.
Equipped with tools like Kubernetes Cluster Autoscaler deployed on cloud computing platforms, certain third-party service providers can help publishers who are still trying to learn the tricks of digital publishing to auto scale their infrastructure. These providers have their own sets of customized and optimized cost plans, allowing publishers to pick and choose the services they need, even for particular periods of user traffic upsurge, for instance.
Audiences will jump to whichever avenue delivers content faster, without any lapses in engagement. Digital content publishers need to develop a smart infrastructure and keep monitoring and updating their systems continually to cater to the digital generation, or risk becoming ‘old’ news. With its promise to deliver both 'speed' and 'scale', an automated infrastructure is now an indispensable part of digital content publishing.