Grant Gross of PCWorld did a great story on “Mobile groups protest proposed Net Neutrality rules” yesterday and covered a wide range of opinions including myself.  The story started off quoting Chris Riley from Free Press arguing in favor of applying wired Net Neutrality regulations to wireless networks.  Riley’s justification was that “Whether you’re connecting to the network through a wire or wireless airwaves, it shouldn’t be any different”.  The problem with this assertion as with other Free Press assertions is that it isn’t backed up by data.  Just saying that wired and wireless networks are the same doesn’t make it so because the networks are different in technical terms, in legal terms, and in economic terms.

Among the more troublesome things about the proposed FCC regulations are the expansive new regulations being applied to the wireless networks.  The FCC has never applied Net Neutrality rules on mobile wireless networks.  Some have suggested that Net Neutrality had been successfully tested for a period of two years during a 2006 AT&T merger settlement and that this had no negative impact on investments, but those regulations were not nearly as expansive and they were only applied to wired and fixed wireless networks which are significantly different from mobile wireless networks .  Since that time, nothing has changed to justify expanding those rules to the wireless network and prohibit existing business models.  The reality is that there are technical, legal, and economic reasons why large-scale regulatory changes to the entire wireless industry are such a bad idea.

The technical differences between wired and wireless networks

As the “grandfather of the Internet” David Farber pointed out, the Internet began from two completely separate experimental networks – the ARPANET wired experimental network and the PRNET packet radio experimental network.  The differences between these networks were was so great that it required the Internet to be flexible enough to accommodate fundamentally different types of networks.

Transmitting signals on a single wire bears some similarities to a wireless network but the resemblances stop there.

• Wired networks run on a lot more than a single wire, and each wire is a completely new and separate transmission medium which multiplies capacity.  For example, a DSL network has a separate wire going to each home while a wireless mobile network has only one transmission medium.
• The propagation characteristics are vastly different between wired and wireless networks.  When you try to send electrical or optical signals down a cable in a single direction, the signal doesn’t disperse which means it stays strong.  Wireless signals disperse in every direction possible which is what gives it its wide coverage area, but the signal weakens very rapidly with the square of the distance.  Wires poke through the wall which results in no signal loss when penetrating walls, but wireless signals are typically weakened another 10-fold when going through the walls of a house or building.  The result is that the throughput per hertz of signal is typically much lower on wireless networks than wired networks.
• On cable broadband networks where the wire is shared, the number of people sharing the same wire is significantly less than any wireless network service covering a few square miles or more.  Furthermore, the differences in the transmission medium means that on average, we can get a lot more bits per hertz over a shielded coax cable than any wireless medium.  The total amount of available spectrum on a coax cable is also several times higher.
• A single strand of single mode fiber has many terahertz of capacity.  All the usable wireless spectrum (for mobile applications and not line of sight) in a given transmission zone amounts to roughly 5 GHz.  The difference is several thousand times.  Furthermore, Commercial Mobile Radio Service (CMRS) spectrum is extremely expensive and scarce and it is limited to a few hundred MHz.
• An HSPA 3G cell might have 7 Mbps of total capacity shared between hundreds of users. For low-duration bursty bandwidth applications like web surfing (which is what the vast majority of wireless customers want), it can multiplex (share) very well because each user is only using the bandwidth a small percentage of the time and for a few seconds at a time when they are actually loading the web page but remaining idle most other times.
• On the other hand, P2P or video streaming not only uses a lot of bandwidth, it also sustains usage continuously over several minutes or hours which does not multiplex (share) at all.  This means that very few P2P users can quickly take over large portions of the spectrum by opening up hundreds of simultaneous connections.  Wireless technology not only has a bandwidth limitation, but it also has a packet per second limit that gets pushed to the limits with just a few P2P connections.
• To strike down existing Terms of Service (ToS) restrictions on wireless networks against heavy bandwidth and heavy duty cycle applications is not practical on a technical level.

The legal issues

Wireless spectrum is sold for very specific purposes e.g., mobile telephony services.  Last year during the 700 MHz auction, AT&T nearly paid three times more per MHz of spectrum for unencumbered spectrum than Verizon which bought much cheaper spectrum with the understanding that it had to operate under “open” Net Neutrality rules.  For the FCC to come back one year later after the auction and declare ALL wireless spectrum to be under Net Neutrality raises some serious legal and ethical questions.  Does the FCC intend to give the price premium of billions of dollars back to AT&T now that they plan to change the spectrum to a regulation-encumbered block?

The economic differences

For one thing, the wireless industry is far more competitive, and the United States is one of the few nations with 4 solid competitors in the Wireless space and Americans enjoy the lowest per-minute cost in the world.  The wireless market is also far more diverse than the wired market in terms of business models.

Another major difference is that content/application providers like Amazon actually pay Wireless access providers like Sprint to enable mobile connectivity to eBook readers like the Amazon Kindle.  The access being provided is limited to the websites and services that Amazon permits, but Amazon is paying for connectivity and not the Kindle user.  Would Net Neutrality apply to the Kindle and devices like it?  Would the restrictions on the Kindle be prohibited?  Does this mean the free but restricted network access on the Kindle go away?  We don’t know for sure because the proposed FCC regulations are very open ended, and it will be a field day for the lawyers and bring years of uncertainty to the market which will stifle investments and innovation.

Finally, one of the biggest problems with Net Neutrality regulation is the requirement to allow for attaching devices.  Today you can get 5GB capped (which is effectively unlimited for web surfing) services for a smart phone for $30/month.  If you want to tether a computer, the plan costs $60/month.  Since Net Neutrality regulation sets a minimum service requirement that gives all subscribers the right to attach devices, the $30/month plan would have to go away.  That means the majority of mobile data users who don’t need or want to use a laptop would be forced to pay for capabilities they don’t need.

Usage caps would need to be finer grain under Net Neutrality

Wireless networks are already stressed to the limit today.  Imagine what would happen if we multiplied the load by 10 or more. There are already 5 GB caps in place, but it is with the understanding that users won’t use certain applications which effectively holds the average usage much lower than 5 GB.  If Net Neutrality rules apply, there would need to be finer grain caps which haven’t been implemented today, and these will be far more complex and confusing for the mobile subscriber than necessary.  For example, we may need hourly and peak usage rate along with packet-per-second limits if P2P usage becomes rampant.  Just because someone stays within an hourly limit doesn’t mean that they can’t congest the network on a minute by minute or second by second basis which renders the network unusable for hundreds of other people.

Conclusion

The FCC needs to consider the substantial technical and economic differences on wireless networks before they impose regulations that are already too heavy handed for wired networks.  It will be difficult enough to get true neutrality on the wired network and technology might someday allow wireless networks to support a wider range of applications that we are accustom to on wired networks today.  Until we get there, wired and wireless networks need to be treated differently.