Perhaps it's unconventional thinking or perhaps it’s just that I am getting old and cantankerous, but there are some parts of the digital economy that do not wear well on me. In particular, the great granddaddy of all things shared - the internet – continues to come up regularly when I discuss the ClearSky architecture. We here at ClearSky Data have taken an unconventional and necessary view of network connectivity. Simply put, we leverage private transport networks for data. This approach still occasionally raises questions, and it’s not unexpected as public IP transit is the way the majority of people have connected for two decades.
The orthodoxy of the internet usually goes as follows:
- Everyone in the known universe is connected.
- The ways to secure enterprise-to-enterprise connectivity are well known and accepted (IP VPNs, IKE, etc.).
- So, just offer your network-based service to anyone who is connected to public IP transit and make it their problem to ensure their connection is suitably beefy.
If only that were a reasonable approach.
Public IP transit (AKA “internet”) is a wonderful tool precisely because so many entities (people, phones, companies, computers, clouds) are attached to it. If your goal is to reach a billion mobile phones, it’s the way to go. For what we are doing at ClearSky, however, public IP transit shares a lot of the disadvantages of, well, the subway. Just blocks from our offices, the MBTA serves its purpose and is a fantastic way to get lots of people to places that are nearby, but as a passenger you have to deal with all sorts of potential hassles. The trains run chronically late. They are overcrowded. They break down. The passengers may not be very nice to you. Maybe even the employees won’t be very nice to you. In the end, you will get to where you want to go as best as this disorganized mass of people and equipment can get you there. Most of the time it works out just fine. When it doesn’t, you are in a world of hurt.
It’s no wonder that folks who need to arrive on time (and not disheveled) will seek all sorts of alternative transportation. A taxi, an Uber, or even your own car can frequently get you to your destination in a better way. If you have real money, maybe one of these is within reach. With public transit as a base-case default, there are all sorts of improvements that are available to the interested traveler seeking to obtain a better experience.
So, why is it that the same does not seem to apply to data transport?
We provide access to the masses of data our customers count on when they want it, where they need it. For us, connectivity matters a lot, and the problems attendant with public IP transport will create a very unsatisfactory experience for our customers. We need guaranteed low, one-hop latency. In fact, we strive to bring link latencies below 1ms in most cases. We need guaranteed bandwidth to ensure that customers can predictably experience full throughput at multi-gigabit speeds. Our customers also can do without having to share links with Netflix downloads and mobile chat clients. While public IP transit will get the data to its destination, the best effort approach won’t cut it. Public IP transit can’t meet primary storage requirements today, nor will it in the near future.
To deliver the global storage network, we decided that we must start with “fast” and “good.” That meant building a private dedicated network for customer data. Like everything else in our tech universe, the costs are destined to go down over time, but in the meantime, we can deliver exceptional service without forcing everyone’s data to ride the subway.
Want to learn more about getting your data off the subway? Download ClearSky Data’s Global Storage Network white paper.