Fast Forward: What makes our network different?

Many view the internet as a nebulous cloud.  For the most part, it's invisible yet it enables us to do things that even just ten years ago we never would have dreamed of doing  (like VR in your own home, for example). At Google Fiber, we think the internet is pretty magical too — but we also know all the intricacies that it takes to make that magic happen. Though it’s often said a magician should never reveal their secrets, this is one (or a few) we’re happy to share.

Not all internet is created equal

There’s more than one way to get internet in your home or business — and not all internet is created the same. Traditional cable internet was built on copper wires that, by initial intent, was designed to bring you television. From the start Google Fiber was designed to bring you Internet - fiber optic Internet. Fiber internet is a series of ultra-thin glass tubes that transmit data at the speed of light. One of the benefits of fiber is that the material withstands the challenges of time, distance and natural elements, and it can be upgraded (higher speeds) quickly by changing the optical equipment at the end points.  In addition, Fiber is not as reliant on commercial power to keep signals moving through the plant nor as susceptible to commercial power outages taking down whole parts of the plant.

But not all fiber internet is built the same. What makes Google Fiber different from other fiber internet providers lies not in the types of fiber we use, but in the way we design our core network and the operational disciplines we use that focus on the Customer Experience.

First, we design a core backbone architecture (think city to city) that has multiple layers of redundancy to provide a high level of network reliability.  Keep reading for more details on how this works.

Understanding network redundancy

Reliability is a major consideration in what Google Fiber customers look for in their internet.  Google Fiber’s network offers 99.9%* reliable connection, which is in large part due to the way our backbone network is built and managed.  Our core backbone network is designed with multiple layers of resiliency.  This makes it a lot more likely that you’ll stay online, even if there is a failure or event in the backbone of the network. Making sure you stay connected is at the core of what we do.

Content delivery -  peer-to-peer and localized caching

Another way that we ensure our Customers get the speed and content they need is by large scale peering and localized caching within our network.  We peer directly with largest media and content providers (in fact, we’ll peer with anyone who is willing to peer with us, and we don’t use peering for revenue generation); this allows us to obtain and deliver content very quickly.  Minimizing latency is just as important as network reliability in our book.  In addition, our localized caching (storage) allows us to store popular content close to your home, so when you request a show or movie, you’re able to watch it in the highest quality without buffering or delays. 

Finally (and perhaps most important) are our Operational disciplines

At Google Fiber, we place the utmost importance on Customer Experience.  Changes, upgrades or expansion of the network are done with the Customer Experience at the forefront of our decision making process.  All proposed changes are pre validated in our labs.  Once those changes / improvements are readied for production, strict protocols are followed to ensure we protect the reliability of the network and most importantly, the Customer Experience.  We have a highly disciplined change management process and we continually measure, study and improve to drive higher reliability.

So no, the internet is not actually magic. In reality, it’s a bunch of people working together behind the scenes to create the technology combined with the operational disciplines needed to maintain the top performance and reliability. But at the end of the day, if your internet is working the way it should, it should feel magical. 

Posted by Thomas Feilmeier, Director of Network Engineering & Operations