Net Neutrality: Are You Ready For It?
May you live in interesting times. I have seen the futures — both of them: one in which net neutrality isn’t adopted by the Internet carriers, and one where it is. I have a particularly bad track record of predicting how the majority of the electorate will respond to external stimuli. Since there continues to be no shortage of emotion surrounding the net neutrality debate, it’s reasonable to play out the scenario where we do have ‘free’ access to the network, where ‘free’ means that the carriers will do absolutely nothing — aside from queueing — to the packets coming and going from your network connection. I’ve touched on why this this ‘no management’ management strategy is insane from a fairness perspective in an earlier post (what ‘no management’ means when multiple subscribers are contending for an oversubscribed resource — or fairness between subscribers) and will come back to it another day. This post investigates what ‘no management’ means for a single broadband consumer in isolation (assuming no contention for resources upstream with their neighbors).
To play this scenario out, one needs to remember that broadband carriers are businesses. First fact: bandwidth demands are going through the roof. The Internet usage model has evolved from the pricing model used when almost all of the bandwidth was consumed serving either email and web pages. Video delivery has changed this forever. Second fact: the easiest way for carriers to increase their capacity to carry your video bits is to charge you more money. If the electorate votes that the carriers can’t touch their packets, then the carriers only have two options: manage the network via dropped packets with their increasingly under-provisioned infrastructure — not a way to make happy customers; or they can increase their capacity. The only way that they can increase their capacity is to spend a lot more capital (routers, cables, trenches, etc.) and spend a lot more on operational outlays like peering and transit costs — remember that carriers don’t have a fixed-price all-you-can-eat arrangement with their upstream network peers and transit providers.
Here’s the punchline: I’ve traveled the world talking to carriers and have seen the future: the only solution — they are going to stay solvent executing — to a net neutral policy is metered billing for both up and downstream packets to a subscriber.
No problem, right! Hands off my damn packets! Vive la revolution! Off with their heads!
What you may not have considered is that if the carrier isn’t managing your traffic, and you want control over your bill (you thought roaming charges were fun), let alone what’s happening with your household devices, you will need to manage your own traffic. Having built systems that help people manage packets from both a security and quality perspective, I can honestly say that it’s a hard problem. It’ll be like trying to manage your electricity bill without any idea what appliances you have, how much power they use, and when and how they turn on and off.
Some questions that you don’t care about when you have a fixed price network access suddenly become much more relevant when you are paying per bit:
- Did your son use up all of your bits downloading something to his laptop over a p2p network and you didn’t know about it?
- Do you have enough bits left over for that Hi-Def movie download?
- How big is that download going to be anyway?
- Are your teenagers (or adults) spending their evenings watching web videos instead of network television?
- Do you have a botnet on one of your pc’s that is injecting spam or generating scanning and attacks outbound?
Are you sure you can answer all of these precisely? Can your Mom and Dad? Grandparents? Do you really want to wait until the end of the month to find out? Knowing how many bits of certain applications you use would be pretty helpful, and maybe which family members are using them would be good to know too. You may want tools to help you police and manage your network so that you don’t have ‘roaming’ overage charges.
A second dimension to the network management problem you will face is making the network ‘work’ in your house: since you don’t want the carrier to touch your packets upstream, you’ve got to do all the management in your own network sphere of control. You may have all the bits you can pay for, but you are still limited by the bandwidth upstream in the carrier’s network. It’ll be your job to control and provision your converged IP network spigot: make voice, Hi-Def IPTV, p2p bulk file downloading, work on multiple concurrent devices with different traffic demands including realtime game consoles, settop boxes, voip handsets, laptops, desktops. Mom… put down that policy server!
Last but not least… it’s not just your own traffic you will need to worry about! Unless you want to have a lot of cars parked near your curb, you’d better make sure you’ve locked down your household network. You will pay for every packet that the neighbors (or those parked curb-side) use. Because security research moves pretty quickly, I hope you plan on keeping up with the latest and greatest here. Nothing’s more difficult from an IT perspective than rock-solid security.
From a company perspective, I’m agnostic about which way the electorate votes here. There is an equally large role for companies like Arbor to play either fork we take. Believe me when I say that it will only be more interesting for us if suddenly there are several hundred million more people interested in how they may need to manage and secure their networks. However, I personally would rather the carriers take on the role they prefer: building out their networks with increased capacity that is managed in a fair and transparent manner to the benefit of their subscribers. Managing a network is a hard job for actual network and security engineers; I’m positive that the majority of the people affected by the outcome of the net neutrality ‘debate’ don’t want to have their end-of-the-month bill depend on their ability to take it up as a hobby.