By Dave Andrews, Chief Architect
Massive enterprises and small businesses alike work with content delivery networks (CDNs). Despite the wide breadth of companies interacting in the space, customers have many misconceptions about CDNs. The biggest one is that we merely handle web caching or store documents such as HTML pages, images, and videos. This couldn’t be further from the truth. In reality, the tasks handled by a CDN are much more complex, nuanced, and dynamic.
The fact is, the internet in its natural state, is broken. For example, in developing countries, any number of infrastructure issues, such as unreliable power sources, or improperly protected network cabling, can cause an outage at any time. Around the world, many entities run old server technologies that work at snail speed and aren’t capable of delivering quickly or running higher-level applications, such as security and edge compute.
Our CDN professionals work extremely hard to manage around infrastructure and technology failures and human errors so that our customers won’t have to. We have years of monitoring, alerting, and operational procedures in place to notify us immediately when problems occur. Whenever we observe a new type of problem, we add new monitoring metrics, alerting thresholds, and operational response procedures, which are often fully automated, to deal with the issue going forward. This isn’t a trivial amount of work when you consider we’re running a CDN covering six continents.
Imagine having thousands upon thousands of active customers and over 30 independent development teams deploying new features, optimizations, or making configuration changes that all interact in some way on the network. That’s how things are at Verizon Digital Media Services, and as you might suspect, ensuring people and technology work seamlessly together, is tricky.
We continually invest in tools that ensure the actions noted above are made safely, enabling everyone on our team to make changes without negatively impacting the system. And we’re continually improving our tools’ supporting systems to increase the performance, safety, and reliability of our customers’ delivery traffic. For example, we’ve built security systems from the ground up that do not slow performance for legitimate users. We’re also in the midst of productizing a subset of these systems, which we are calling EdgeControl. This will provide a new paradigm for managing your CDN configuration, one that adheres strictly to the principle of least astonishment for DevOps, which will allow for total control via simple and powerful APIs that provide integrated testing, fast and controlled deployments, and real-time analytics. EdgeControl allows our customers more visibility and control over their configuration than ever before. It’s an opt-in product, meaning that our customers will be able to choose whether to use it. We expect the enhanced control will enable and encourage a rigorous operational stance in our customers, where issues are caught early and resolved quickly, to keep their content delivery smooth.
Giving customers the ability to make changes with EdgeControl doesn’t mean that we’ll no longer be working behind the scenes to ensure our network is running at its best. Our internal team works around the clock to go beyond simple “does-it-work?” configuration settings to ensure our customers extract the most value and performance from the network. In most cases, we’re able to solve any network issue before it impacts our customers, meaning they never have to think about the ongoing complexities of keeping the entire system running smoothly.
To further improve performance and reliability, we’ve collected a tremendous amount of feedback from our professional services team and integrated a lot of it into these new systems to ensure we have compelling products (and future products!) that allow customer control over how they use the system and to see the impact of their use.
Although we’re empowering users to get more visibility and have more control, there’s still a lot they don’t know about what happens behind the curtain. One example is our sophisticated DDoS detection and alerting system that shows when and where DDoS attacks are happening and their attack profile. The biggest reasons we don’t bring this up often is because we don’t want to give attackers insight into the impact they might have on the system, where their attacks are landing, and how much capacity they might be absorbing.
So yes, while we strive to create products and systems that offer our customers transparency and visibility, we also create clandestine “silent systems” to keep attackers in the dark.
Speaking of “silent” systems, we also keep relatively quiet about the various components of network infrastructure within our company. The internet is, in many ways, like a power grid in that everyone expects it to work without ever knowing exactly how it works. How many times a day do you think about how electrons flow from a power station to your house? Probably not very often. The same goes for the internet. It’s not the job of our customers to worry about how the internet works, it’s our job and that of our vendors and providers.
Despite CDNs being viewed as simple web caching networks, there’s far more to them than meets the eye – even if you can’t see it. The work we do and will continue to do helps keep a growing percentage of the internet up and running at peak performance.
And while we’re constantly developing new ways to involve customers in the running of their system and enable them to get even more benefit from our global network, customers can be assured that we’re always working behind the scenes to make our CDN perform better and safer for them.