Let’s not apply the policies of the 20th century to the technologies of tomorrow
December 12, 2011
Today, broadband providers have to carry all kinds of traffic that they can't block -- content, Web pages, platform services and other modes of communication –- because of the principles of net neutrality.
In order to address net neutrality, let’s examine a bit of history.
The Pacific Telegraph Act of 1860 called for the simplest net neutrality scheme: All messages -- in this case, telegrams -- had the same characteristics. All messages were to be treated equally. There was the simplest possible traffic management (FIFO), and this simple traffic policy prioritized government messages. All these features were approved in Chapter 137 U.S. Statutes by the 36th Congress.Now it’s 150 years later.
The Internet is perceived as neutral -- a fundamental right of every person on the planet. Traffic management exists to manage limitations in capacity and different services through quality of service (QoS) and multiprotocol label switching (MPLS).The question is: Should all IP packets be treated equally? This is a difficult and important question.
Telecommunications law specifies distributing, preventing, restricting or delaying the transmission or delivery of a telecommunication message as a criminal offense. It’s punishable by up to three years imprisonment (although it is rarely implemented if at all).
Early disruptions of that model included attempts by some mobile service providers to block VoIP traffic on mobile broadband networks (as it did for wireline networks). This attempt was rejected by regulators and the FCC faster than a speeding bullet.Telecom laws were amended last year to allow principles of net neutrality, subject to fair and efficient traffic management. The regulator may now decide which traffic management practices are deemed “fair and efficient."
In the United States, the top two carriers have appealed this decision, and it is still being worked out. Meanwhile, the Internet continues to improve.
So, what does limited capacity mean? Simply put:
Bandwidth limitation, latency, jitter, packet loss
Temporary Congestion especially as HDTV and 3-D technologies spread on the Internet
Mobile Broadband networks are more vulnerable to this limited capacity and lack of spectrum.
What services are most sensitive to limited capacity?
VoIP telephony
Video conferencing
Gaming applications
Streaming video (which will truly take off in 2014)
Cloud services –- though the jury is out on the impact this will have on the network
Other services are less sensitive to fluctuations in capacity:
File Transfer/Sharing
Social Networks
Email
Normal Browsing
There is currently a huge conflict of interest when it comes to net neutrality, from small players to big players. Let me explain the benefits of net neutrality for the consumer and society as a whole. Net neutrality:
Helps encourage the free flow of ideas and content
Provides access to a wide variety of services
Provides access for the development of new services
Offers incentive to develop new services
Offers free choice for customers
The big broadband service providers have other concerns. They worry about:
Providing good services
Becoming more than just a dumb pipe
Reduced income from value added services (VAS)
Reduced return on investment
Losing captive customers
See the next page for some policy alternatives -- ideas that might make everyone in the industry and the world happier about the future use of IP.
Alternative 1: Simplify everything
Pros:
Simple traffic management
No ground for claims from service providers
Simple regulation
Cons:
Unacceptable degradation of real-time services
Damper on competition
Inefficient usage of bandwidth
Unreliable QoS
Alternative 2: Modify our approach to IP traffic and policy
In this scenario we would say that:
Real-time services (RTS) would be prioritized over non-RTS
Critical services prioritized over non-critical services
Low bit rate services (e.g., voice) would be prioritized over high bit rate services
Packets of comparable services are treated equally
So, now let’s look at pros and cons of this alternative:
Pros:
Promotes system efficiency
Provides benefits to consumers and service providers
Leaves lots of room for competition
Provides an incentive for innovation in services
Cons:
There will inevitably be debates about what fairness means
This alternative will require management by complex hardware and software
Leaving aside the hardware/software management question, how can this approach make everyone happy?
Ideally, we would like to make individual users (whether consumers or providers) better off without making any other individual worse. In the real world that won't work, but we can lessen the sting by applying policies intelligently.
There is room for everyone in the network, and we should spend our energy on innovations to the network rather than endless debates on issues, which most likely will not even exist in ten years.
So, if the FCC and others in the industry agree on either alternative, how does it get implemented? This is a big question.
Here are some of the challenges we face:
How do we make RTS users much better off by making non-RTS users a little bit -- but almost negligibly -- worse off?
How do we develop more in-house services versus competitors' services?
How do we eliminate the “not-invented-here” attitude in companies?
How do we discourage anti-competitive practices?
How do we block or bar service-degrading techniques that already exist?
How do we implement deep packet inspection without violating user privacy?
How do we watch carefully for handset and its applications on the network similar to what happened in the case of the iPhone?
I am proposing transparency instead of regulation for the Internet we all love and hold dear. History demonstrates that the Internet will improve with time, adding new helpful services and spurring investment.
Why traffic management is a must
The economic growth spurred by Internet technology would create jobs in the United States and worldwide and would have a positive effect on many deficit challenges faced by world governments. Even though I’m not a politician -- just a scientist -- that providing credible comparable information about broadband service provider traffic management policy will better serve end-users than more regulation.
In essence, let’s not apply the policies of the 20th century to the technologies of the 21st. I believe that the rate of innovation will be inversely proportional to rate of regulation in our new century, and that includes broadband wireless service.
Now, I can predict the counter-argument: “Do common customers have enough know-how to make use of such complex technical information?” My answer would be yes -- with time, they would. This would require aggressive education on the technologies beginning in the K-12 years.
I predicted almost more than a decade ago that IP-MPLS would eat everything and that broadband wireless would eventually eat IP. There is no question that this prediction has come true.
Here are some points I would like policy makers and service providers to consider:
Move to all-IP networks, and retire all public switched telephone network legacy services
LTE-A will finally push mobile broadband ahead of fixed broadband by 2015 or perhaps sooner
LTE allows the coexistence of voice, video and data in the same mobile broadband network
Traffic Management is essential for QoS even for in-house services
In the final analysis, traffic management is a must. It’s non-negotiable. A policy of optimized and fair traffic management must be established in policy and in the law -- especially in the regulation or licensing of ISPs and broadband service providers.
— Dr. Hossein Eslambolchi is the former CTO of AT&T and chairman and CEO of 2020 Venture Partners
You May Also Like