Home / Blogs

Beyond Neutrality - Enabling a World of Connected Things

Video Recording of “Beyond Neutrality – Enabling a World of Connected Things” talk at a joint meeting of the IEEE CE, Communications and Computer Societies and the ACM.
(Click to Watch)
This article is based on the talk I gave at a joint meeting of the IEEE CE, Communications and Computer Societies and the ACM. It is based on my Connectivity Policy essay as well as the column I wrote for the IEEE/CE Magazine.

The growing interest in the “Internet of Things” is forcing us to think beyond the web to a much larger world of connected devices. We can tolerate the many barriers to connectivity because we expect that someone can provide the necessary credentials to log in to the providers’ services and to adjust Wi-Fi access keys whenever the access point changes or simply to click “agree” at a hotspot.

This doesn’t work for “things” which can’t recognize a sign-on or “agree screen”. This may not be obvious at first because it’s easy to demonstrate an automated house if you set “things” up just right and don’t change anything but the illusion quickly disappears after the cameras are turned off and real people try to live in the home. Many of the wearables on the market stop working as soon as you walk away from your phone or change phones.

A New Paradigm

The Internet represents a fundamental shift from thinking about services inside a network to services created outside the network by users with their intelligent (AKA programmable) devices. By using the intelligence in our devices we are able to create solutions that do not depend on a provider.

VoIP works by taking advantage of opportunities and as the capacity of the Internet grows (thanks to applications like the web) high quality voice with video become more likely. In the talk I cite an IEEE article on VoLTE (Voice over LTE) which depends on every carrier in the path doing the right thing—very old paradigm.

This is why it is important to understand what I call the three stages of digital. With telegraphy we could carry messages over any distance but the introduction of analog telephony made distance difficult and required a different infrastructure for each kind of content.

In the second stage communication technology addressed this issue by encoding the analog signal digitally. As with any new technology it emulated the old technology and its business model. More subtly, it kept the assumption that speech was maintained within channels and could be measured in bits. We assumed that communications in the sense of “speech” and communications as a technology were the same.

Today we are in the third generation of digital in which we use intelligent devices to create our own solutions without depending on providers. We are no longer emulating the old phone networks.

Yet the assumptions of analog telephony are still implicit in today’s polices even as the intelligence is in our devices rather than in the network and speech is no longer confined to channels. Communications technology and communications as speech are no longer the same.


Our public policies have yet to catchup with this fundamental change. The very term “Federal Communications Commission” shows the confusion between the technology and speech. Existing regulations continue to be based on old paradigm and assume that speech is contained within channels (wires and frequency bands). The business model of telecommunications depends on carriers making a profit by using their expensive infrastructures to assure the delivery of content as if it were freight.

This framing of speech as freight is the reason we seek to extend common carrier rules in the form of network neutrality. And, indeed, to the extent that we rely on carriers we need rules to assure they play fair.

But the world has changed and with the Internet we no longer depend on carriers. As we connect “things” there is no “content” as such—just devices exchanging packets.

This has always been true with the Internet but as long as we see the Internet through the lens of the web it seems as if it’s the next iteration of the phone network. Connected things force us to confront the mismatch between the carrier-centric view and one in which value is created in devices.

Broadband Infrastructure

If we look up at the poles in our cities we can see the television-age technology with a separate analog pipe going to each house. No wonder it is expensive and difficult to maintain this infrastructure.

Yet the power of the Internet comes from our ability to think about the relationships between devices with nothing between the two end points. The value is now created using the networks and not by owning them.

Even more of a problem for the current business of broadband is that there is no longer any differentiation between services. In this illustration (see image on the right) all it takes to switch between Comcast and Verizon is unplugging the white Ethernet wire and replacing it with the red wire. There is no essential difference!

With HBO and CBS announcing that they are going to be offering their content, as Netflix and Hulu do, without requiring a cable TV contract, we’re going to be forced to come to terms with a new landscape.

All our devices are now simply computers. This is a diagnostic screen (see box below) from my Verizon Set Top Box (Arris VMS1100).

It reminds us that indeed the STB is just another computer. It’s time we recognized the reality and move from a Federal Communications Commission to a Department of Connectivity that focuses on infrastructure. Other federal agencies (such as the FTC) would be responsible for what we do with the technologies.

Paying for our Infrastructure

This common infrastructure would be like roads and sidewalks, funded as a common infrastructure. The idea that each wire must be a profit center doesn’t work because the value is in the whole and not the parts.

We’re not going to see people embrace this approach without examples. Instead of focusing on cities we need to start from places where a small number of people can cooperate such as an apartment house or housing developments (Known as Multiple Dwelling Units or MDUs) where a small group of individual or a single owner can set policy.

We have real world examples such as “Castle Village” in NYC and other living groups where people pool their resources to pay for common infrastructure and for a single fat “broadband” pipe.

These connected zones will set examples for the future. These zones serve as the seed for wider connectivity and at some point we’re going to see such common connectivity as the norm.

Mobile and the Internet

The Internet architecture needs to evolve. There isn’t a mobile and an immobile Internet. The relationships between end points shouldn’t be dependent upon where they are at any given moment. The accidental history of the Internet going back to the days of large mainframes has left us with a legacy of expedient design decisions.

We assumed machines didn’t move around much. They stayed in the same place or “address” so we didn’t need to distinguish between their network address and the machine identifier. The IP address could serve both roles. As machines became more mobile this no longer made sense.

The legacy of “long distance” being difficult has led us to focus on accessing distant web sites as a prime example of what we could do with connectivity.

For connected things we need an architecture which is about relationships that are indifferent to place and what’s between the two end points. We need to avoid central dependencies. If I want to turn on the light in my house I shouldn’t need to be connected to the DNS just to find the address of my door lock and the ability to control the lights in my house shouldn’t be dependent upon a single hub working perfectly year after year.

The most mobile devices tend to be small so we use the term mobile as a shorthand for programming for small screens. This is just one possible form factor and not mobility.

By recognizing making this assumptions explicit we can rethink our “givens” and move on to take advantage of connectivity. We can look beyond neutral treatment by providers to empowering each of us to create our own solutions.

One other bit of late breaking news: Android 5.0 (Lollipop) will enable people to pick up any device and use their applications. This challenges policies that presume that carriers provide devices as a way of delivering and billing for the services. It’s more reason to shift to a model of funding common infrastructure rather than trying to maintain the fiction that we communicate within pipes.

Protocols for Connected Things

A new generation of protocols can be built using the existing Internet just as today’s Internet repurposed our telecom infrastructure.

Self-Chosen Identifiers: Use identifiers chosen by the devices themselves without any central function. We can use methods such as GUIDs (Globally Unique Identifiers—random numbers) as one mechanism. We don’t have to depend on having a central provider of identifiers in order to assure uniqueness. But uniqueness is not always necessary. We deal with ambiguity each day as when we are looking for a Starbucks but don’t care which one. Or we may have multiple John Smith’s and we don’t need a third party to tell us which is which.

Place and Final End Points: We would distinguish between places and final end points. Places don’t move much. This means we can have a relative stable and simple routing tables. End points would be responsible for telling their counterparts where they are at a given time just like we tell our friends our current street address.

Economics for Simplicity: Perhaps most importantly, by funding common facilities we greatly simplify today’s routing by avoiding the twisting and winding passages that are more about the business model of telecommunications than about providing connectivity.

Making every wire a profit center means that the carriers have to prevent unbilled bits from passing. This works at cross-purposes with connectivity and is an impedance mismatch for best efforts connectivity in which raw packets have no intrinsic value.

The Business of Things: Very simply billing for packets requires putting a border around the facilities we use to communicate and makes failure the default. If we are to realize the potential for an “Internet of Things” and enable a connected society, we need policies to make successful connectivity the default and not the exception.

By Bob Frankston, IEEE Fellow

Bob Frankston is best known for writing VisiCalc, the first electronic spreadsheet. While at Microsoft, he was instrumental in enabling home networking. Today, he is addressing the issues associated with coming to terms with a world being transformed by software.

Visit Page

Filed Under


Not that I would necessarily be against Dan Campbell  –  Oct 27, 2014 2:32 PM

Not that I would necessarily be against my tax dollars going to assist my Internet connectivity, I think you need to realize that the whole thing started as a Government project that never took off commercially until, well, it was commercialized, and many saw the opportunity to make a profit off it all, propelling us forward to where we are and continuing to propel us.  The carriers are no angels and certainly not the most efficient in the world - there are a ton of issues particularly around the US duopoly in broadband space or even the backbone that are still somewhat grandchildren of the original long distance carriers and baby bells.  But injecting the Government into the middle and all of its bureaucracy and inefficiencies may not exactly be the panacea.  Wouldn’t we be backtracking, and what are the downsides?  There are always implications.  Ultimately, once the Internet was commercialized it evolved much faster than the legacy phone network did over the previous (nearly a) century, as did the intermediate infrastructure products, end user computing devices, consumer products that traditionally were outside the computing or telecom realm, and general technology development to support desires.

As for some of the other points in the article, there are already unique global identifiers and always have been really, with some limitations, however that does not mean computing is ready to handle them all without some level of aggregation. Hell, we still have zip codes for goddsake!  Internet routing tables are already hovering around half a million routes, which has taxed even larger backbone network hardware, necessitating upgrades, but certainly puts a strain on even large enterprises much less the end user computers or broadband devices and STBs for the home that need to stay somewhat inexpensive.  Most enterprises won’t run BGP, or at least not while accepting full routing tables from its upstream ISP - there is seldom a reason to in most enterprises, and certainly less so in the home (if at all ever), but you couldn’t if you tried.  No more so than I could actually have a database at home of every possible physical address or phone number on the planet and somehow keep track of that while it changes by the minute.  There can’t be “relative stable and simple routing tables” if you must create unique IDs for every single connected device that must somehow maintain that unique ID and communicate its location constantly as it moves around.  Of all that has changed and evolved through the last two decades, one need only look at the incredibly few changes within routing protocols like OSPF and BGP and their overall stability to see how critical they still are and how well they work, thus there must be a major necessity.

The essay is only an overview. I Bob Frankston  –  Oct 27, 2014 3:05 PM

The essay is only an overview. I go into more of these issues in the talk where I try to explain how the Internet is a discontinuity from the concept of networks as a service. The FCC is very much in the middle today – by starting with connectivity from the edge we reduce the dependency on a center. I use the term DIO for Doing It Ourselves as a local community such as in an apartment house. We use the term “government” for such cooperation but it’s very different sense that a rules imposed from on high.

I can’t go into all the design details in this short note but I contend that routing would be far simpler and more stable if we could take advantage of geographic identifiers like zip codes rather. The network itself doesn’t need to track or even know about each unique address. The devices (and users) would be responsible for telling others how to reach them. You can compare it this with Skype (or, now WebRTC) – the network doesn’t know that there is a conversation progress. That’s purely between the two end points.

Well of course things would be simpler Dan Campbell  –  Oct 27, 2014 3:56 PM

Well of course things would be simpler if routing tables had some sort of regionalization, and there were attempts at that even in how IPv4 allocations were made from IANA to the RIRs, but no one really adhered to those artificial boundaries.  You can’t.  And things are better ultimately that they didn’t, like we do for zip codes.  Companies and network providers are global and, more recently, mobile, so it’s not easy to maintain regionalization without some other consequence that’s just as bad.  Ultimately, all the end users really need to know is the next hop to send something, the default gateway, and get it out the door.  Thinking as an end user / residential subscriber rather than an IT person, I’ve always viewed the backbone provider network as a cloud on which I don’t have any direct dependence to one entity providing service.  It’s always there and I don’t really pay for it directly, although those business certainly make money selling transit service or broadband or everything in between.  At the end of the day, broadband is relatively cheap for me and it’s faster than I currently and probably will ever need, covering multiple computers, phones, automatic backups to the cloud while also running one or two HD streaming channels at the same time.  If anything, advances in wireless and cellular technology are needed more so than a common core backbone that is somewhat “free”, since for mostly people it kind of already is.  Yeah, the price you pay for broadband indirectly includes everything that goes into the cost including upstream transit and on and on, but it’s still cheap in the long run.

Also, aside from the routing protocols I mentioned, IPv4 has shown to be one of the greatest of all technologies and incredibly resilient, so much so that it not only won out about two decades ago over a host of other layer 3 protocols, but it still hampers IPv6 deployment, even with the limitations that NAT/PAT places on applications and even with the exhaustion of IPv4.  It’s almost 2015 and it’s still going strong and IPv6 is still slow.  In the mid-to-late 90s there were some comical talks about ATM replacing IP and having ATM to the desktop; that was blown away.  There really hasn’t been a viable reason to change from IPv4 that provides enough obvious benefits to overcome the cost and pain to migrate.  IPv6 isn’t really a bit leap, it’s like a cousin, and it’s still stunted by IPv4 today.

As I said it is difficult to Bob Frankston  –  Oct 27, 2014 4:03 PM

As I said it is difficult to do network design in this forum. The continued use of IPv4 is indeed a testament to the power of “routing around”. This is why we need to go back to first principles and rediscover the Internet from the edge without confusing words like “broadband” or “backbone” which are terms from telecommunications.

There's a number of different ways this could work wolfkeeper  –  Oct 28, 2014 11:03 PM

If we assume that the basic idea is: I wander around, and it connects automatically: it works, wherever I am

Then that’s only half the equation. Some of the other things that go into the rest of the equation are:

* who pays what to who for the bandwidth you use
* who pays what to who for the access of equipment
* who originally pays for the purchase and installation of the equipment you use
* what (if any) restrictions are there on use of equipment

If the per byte cost is very low (and it’s tending towards zero), people could just buy the hardware and install it for their own use, and let other people use it as well, because they’ve already paid for it, it doesn’t matter to them, it’s a flat fee. But then what do you do about next door neighbours who won’t or can’t get their own line installed? The hardware has a maximum capacity. But perhaps the capacity may eventually be so very large, you wouldn’t normally hit it?

Another perhaps better example, conceivably it could go like a razor blade model; the ISPs give you the line and hardware for a very modest price, and then charge you per byte for accessing it. As in, not for using the line, but everyone has their own central account with an ISP and all their own equipment uses that account wherever they are in the country. The ISPs could even fractionally cut you in, if somebody else uses your line, your line gets cheaper for you. That way it incentivises hotels (for example) to install good equipment, because the line would pay for itself, and so you wouldn’t want it to choke, because then you’d lose money.

Those kinds of options and more may happen.

You need to be careful about treating Bob Frankston  –  Oct 29, 2014 5:54 AM

You need to be careful about treating bits are a consumable and that we have pipes. I explain more in the video and in another CI post—http://rmf.vc/PurposeVsDiscovery. Think about sidewalks as common infrastructure, albeit far more expensive than wires. We don’t have a complex mechanism to charge people for each step then take and the amount of sidewalk they consume in the process. We just pay for concrete or copper or fiber once and a small fee for maintenance as a shared facility.

They are still finite resources wolfkeeper  –  Nov 5, 2014 10:12 PM

Both sidewalks and pipes are finite resources.

I do agree that we don’t meter each footstep on sidewalks.

However, in the real world there are actually mechanisms that try to make sure that people don’t monopolise a sidewalk; if you pitch a tent on a sidewalk, you will typically get removed in no uncertain way.

Likewise, if people just connect to your broadband and start file sharing, this can potentially be a big issue.

I suppose there are obvious ways to mitigate this, such as deprioritising their traffic.

In general, the lack of peering in WiFi is quite annoying; I’ve noticed situations where my equipment could see neighbour’s base stations, and base stations could see each other, but my equipment couldn’t see my own base station. And the inability to join broadband connections together means I get a fraction of the peak bandwidth I otherwise could be enjoying.

After thinking about it carefully, a lot of these problems, which are a subset of what you are trying to do, seem to be ultimately down to NAT, without this, this could be configured in with suitable routing tables, but NAT makes it all but impossible to do that.

Additionally, there’s no reason in principle that a laptop couldn’t log simultaneously into multiple base stations. But given the NAT problem, there’s little point. The packets for any given stream couldn’t use both routes simultaneously since they would have apparently originated in different places. Even though IPv4 permits that.

I've addressed these issues in the talk Bob Frankston  –  Nov 5, 2014 10:21 PM

I’ve addressed these issues in the talk and other essays—that’s why I provided the pointer to http://rmf.vc/PurposeVsDiscovery. As to NATs - they are symptoms of larger protocol problems that I discuss towards the end of the talk.

Comment Title:

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.

CircleID Newsletter The Weekly Wrap

More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Co-designer of the TCP/IP Protocols & the Architecture of the Internet



New TLDs

Sponsored byRadix

IPv4 Markets

Sponsored byIPv4.Global


Sponsored byVerisign


Sponsored byDNIB.com

Threat Intelligence

Sponsored byWhoisXML API

Domain Names

Sponsored byVerisign

Brand Protection

Sponsored byCSC