There’s nothing quite as disconcerting on a Friday afternoon as hearing the guy on the other side of the room utter some choice four-letter words followed by “DNS”. This happened recently as I was packing up for the day at my favorite local coworking space, about to go enjoy the New Hampshire countryside on a beautiful summer weekend. I, however, in all good conscience could not ignore the plight of my fellow denizen of the local coworking space and reached out to help. Well, as it turns out, he was using the DNS provided by his web hosting company and waiting for a very simple A-record change to propagate. So why is it, in this world of supercomputers in our pockets and Silicon Valley-built electric super sedans putting Ferraris to shame, that we are still waiting with hopelessness on a Friday afternoon for 32 bits of information to push to a box out there?
Imagine for a moment, you’ve got a new app that you’ve launched for iOS and Android devices. You’ve built everything in such a way that more back-end databases and application servers can be spun up and down as needed to cope with varying levels of use. This is all well and good, but what happens when that new application server is spun up, but the DNS takes two hours to reflect the fact that there’s another A-record in the pool of available addresses? In the time that would take, the load profile on the application could have changed drastically for worse or if you’re really lucky, for better.
The software that the majority of the world uses, BIND, has managed to evolve over the years and provide decently quick and reliable service to those who employ it. There’s a tax with BIND, and that comes in the form of physical server resources as the application runs in memory. If the zone files start getting bigger, so does the amount of memory that the machine has to sustain. Thankfully, when zone updates are sent to BIND, they are enacted in a reasonable time frame without corrupting the information that’s already in memory. However, the speed at which the changes are then sent to all other machines in that network varies by operator and frankly, isn’t quick.
Think about it, we have all these great tools for orchestrating servers getting spun up at a moment’s notice, real time statistics and data about nearly anything, yet the venerable protocol that makes everything talk lags behind. This example is a common situation that many have experienced time and time again, and something at NS1 we’ve all but eliminated, in part, by developing our own name server software. We did this because the speed of confidence required to power the Internet requires changes to take effect quickly.
By being able to reflect zone and record changes in nearly real time, we ensure that critical changes in your stack are reflected where it matters most to the people who keep your lights on - your customers. Fast resolution of DNS is table stakes in today’s technology ecosystem, but fast propagation clearly sets NS1 apart. Granted, your changes will always be subject to the TTL value set for a given record, but when it’s time to make the change you can rest assured that NS1 is here to ensure that change is executed and propagated as quickly as possible.
In short, if you’ve been dissatisfied by DNS changes that seem to take forever and you’re looking to tighten up that piece of your stack, we’re here to help. As for my fellow coworking space denizen, I offered up our free trial to him after he waited two hours for that simple change to push to the edge with his hosting company’s DNS, and he’s asking for permission from his boss to make the change. Folks, it’s easier to apologize than it is to ask for permission - though no apology is necessary when it comes to making the switch to NS1. Don’t sit nervously at happy hour, staring blankly into your beer for fear that your changes haven’t taken hold - waiting sucks! Move at the speed of confidence with us, and you’ll be the hero of the office, and happy hour for that matter.