in reference to: FCC member blasts Google-Verizon plan, says it favors the rich - The Hill's Hillicon Valley (view on Google Sidewiki)
Saturday, August 21, 2010
in reference to: FCC member blasts Google-Verizon plan, says it favors the rich - The Hill's Hillicon Valley (view on Google Sidewiki)
Last week the FCC published its’ report on U.S. broadband Internet usage entitled Broadband Performance: OBI Technical Report No. 4. The press chose to report on the sensational claim in the Executive Summary that actual measured bandwidth was half of the advertised bandwidth. If they would have taken the time to read past the Executive Summary or not copy the other articles written about the report, they would have noticed that the report supports Quality of Service (QOS); thereby, implicitly endorsing differentiated services. They even dedicated Appendix 3 to a cursory discussion of QOS.
In Section I, the concept of QOS is first mentioned when profiling the different types of traffic users download. In the quote below, The FCC states that high definition video needs bandwidth and QOS.
At the high end of the range, an application such as enhanced high definition (HD) video teleconferencing could require 5–10 Mbps, or more along with significant quality of service (QOS) performance (see Exhibit 9, where “Symm.”—short for symmetrical—indicates that the download speed is also required for upstream traffic).
In the next paragraph they reveal the other parameters that are required for HD video conferencing.
Download speeds are only one measure of broadband performance.
For example, HD quality videoconferencing requires very fast upload speeds to allow a person to transmit her image and voice while simultaneously receiving the image and voice of another person. In addition to upload and download speeds, measures of QOS such as availability, latency and jitter (variation in latency among different packets) may be important. Some applications, like e-mail or text-based Web surfing, are generally insensitive to these other measures of network performance, but for other applications, such as videoconferencing, these measures may be important (see Exhibit 10).
These statements introduce the reader to the concept that bandwidth alone may not be sufficient for certain types of services. Later in Exhibits 9 and 10, services are classified by their need to be experienced immediately along with the need for QOS to determine user experience. The FCC is unequivocally stating that all bits are not created equal. They identify real-time and near-real-time traffic as needing lower packet loss, latency, and jitter from typical web browsing or e-mail reading. The FCC’s quiet endorsement of differentiated services came in the beginning of Section III by stating:
The NBP therefore relies on a National Broadband Availability Target defined in terms of quantified download and upload speeds, with qualitative reference to a QOS consistent with the delivery of voice and video applications.
Perhaps the reason why the FCC was dragging their feet on net neutrality regulation was that internally they actually support differentiated services.They realize that it can improve overall user Internet experience and provide real competition to the incumbents. By letting Google and Verizon publish their principles of net neutrality, they let those two take the flack for supporting differentiated services instead of staff having to deal with the political fallout. Whatever the reason, I am glad that the bureaucrats recognize how the application of QOS will benefit the Internet. Too bad the press missed it.
Thursday, August 12, 2010
Out of the hundreds of articles written about the recent set of principles released by Google and Verizon, this article written by Richi Jennings correctly points out two reasons why the egalitarian net neutrality concept will not work. He realizes that packet prioritization is necessary to create a good experience for all Internet users. The Internet’s architecture has changed much from the early days when it just pushed around short notes, e-mail, file transfers, terminal sessions, and net news. The services have increased to real-time and near real-time traffic like voice, video, instant messaging, and webcams. The backbone has changed from cobbled together private line connections to a complex backbone connecting many different carriers with both paid and neutral peering points. Its’ ability to adapt to the increase of users, services, and complexity is a credit to the many people that contributed to its architecture.
One of the reasons that the Internet has been so successful and extensible has been an absence of governmental interference. True that some countries have outlawed it or prevented access to certain types of material, but they have not meddled into its fundamental design. If the U.S. government intervenes by not allowing companies to add new capabilities or manage their networks, then the Internet as a whole will suffer. Every time the government creates a new law there are always unintended consequences. This time we can see them before they happen (is that possible?). Any carrier should not block or inhibit access to any lawful content provider. That is/should be a basic Internet tenant. That tenant should not exclude the classification/prioritization of traffic types as long as it is applied equally to any content provider.
Richi is one of the few people in the technical press that has enough of a technical understanding of the Internet to acknowledge that Google’s and Verizon’s principles are a good start. Holman W. Jenkins Jr. wrote an op-ed piece in today’s Wall Street Journal that acknowledges the need for traffic differentiation and recognizes that wireless is the true battlefield. I wish more of the technical press was as enlightened as these two.
Wednesday, August 11, 2010
Monday Google and Verizon released their joint framework for an open Internet. While the press and digiratti expressed their outrage that two leaders in the industry too the initiative to make a proposal, it was a major step forward in reaching consensus by the largest access and content providers. Condemnation was generally universal from expected sources like Wired, This Week in Tech, the George Soros funded MoveOn.org, the technical press in general, FCC, and even the Wall Street Journal. Other carriers, such as AT&T and Level 3, were generally timid in their acceptance of the progress. The reasons for the condemnation were varied and often overstated. Most of the denunciation resulted from a lack of knowledge in how the Internet and telecommunications networks work. This ignorance causes fear that results in calls for government intervention even though there is yet to be a problem.
Presently all Internet traffic is classified as a “best-effort” service. That means that all packets travel independently through the Internet with no preferences given to them. The packets are not even guaranteed to arrive at their destination if there is congestion at network nodes which is why TCP was invented to tell the sender to resend lost packets. The truth is that most non-business Internet traffic uses the same equipment and bandwidth as business services that are already prioritized. Carriers may already be slightly delaying best-effort Internet when their router becomes congested with business services. Therefore, the concept of “all packets being equal” is already a misnomer. Even content providers like Google prioritize their internal traffic over Gmail sessions or search queries when they may be comingled. Carriers have been selling “differentiated” services to businesses for several years to ensure the quality of video and voice traffic over e-mail, web browsing, and other applications that are tolerant to a few hundreds of milliseconds delay. It is about time that residential users have access to those same capabilities if innovative services like Over-the-Top Video (Hulu Plus, Netflix, etc.) and VoIP are going to compete with Cable TV and “digital phone services."
Several articles prophesized that the introduction of differentiated services would create two Internets because carriers would dedicate almost all of their bandwidth to the more profitable differentiated services or split networks. This statement is false because operating parallel networks is not an economically viable solution. Duplicating equipment would drive the costs up of all services. In reality network operators are always trying to converge their networks to a single network for economies of scale. We are reaching the 1950’s goal of the Bell System that all services will ride over a single network. The Internet Protocol (IP) and inexpensive Ethernet has made that goal possible. Carriers are increasingly migrating their network to IP to yield both capital and operational cost savings. As these networks merge, Internet access will ride right along with the voice and video services also offered by the carriers. Services like FiOS, U-Verse, and Xfinity are based on this shared network concept. Carriers of these services dedicate bandwidth to each service. If a competitor like Netflix wanted to offer a service to you they could not purchase bandwidth from Comcast, AT&T, or Verizon. They have to do it over the Internet where their movies are subject to potential degradation through the normal course of transit.
Differentiated services would simply introduce priority bits to the traffic that may be sensitive to delay and jitter like a Netflix movie. Best-effort traffic would continue to have no priority. In most cases the end-user would not even know that their traffic was being delayed. If that same end-user purchased a Hulu Plus Plus service that was differentiated, they would no longer see the excessive buffering, stopping, and block errors that they typically now see. The result is a better Internet experience for all. Throwing more bandwidth at an end-user does not necessarily solve the problem as Molly Wood of CBS Interactive (CNET) and others have postulated. Routers, servers, and other transport equipment can slow down traffic if too much hits it at one time no matter how much bandwidth is available.
Net neutrality is a vague term that means different things to different people so saying someone is for or against net neutrality means nothing. Most everyone including the carriers will agree that they want an open Internet, but what does that really mean? Verizon and Google each have some of the Internet’s pioneers working for them, and they have the technical know-how to best evolve the Internet. Al Gore, Ted Stevens (RIP), George Soros, Rupert Murdoch, Jason Calacanis, the FCC, and many others did not invent the Internet nor do they engineer it on a daily basis. They should leave the engineering of it to the many companies and people that build and manage it. The Internet Engineering Task Force (IETF) is composed of those individuals which make it a perfect body to decide the tenants of net neutrality. The Internet is a global phenomenon because it was designed and built by these people, not bureaucrats, politicians, and do-gooders. By deviating from that formula we risk polluting the Internet with restrictions that will stifle innovation and the entrepreneurs that net neutrality proponents are claiming to represent. That is why I believe that this shared statement of principles should not serve as a basis for legislation by Congress or regulation by the FCC, but as a request for comment (RFC) by the IETF so the Internet can truly be left free and open globally.
Verizon and Google arrived at a major milestone by agreeing that they will not block or impede traffic/services from any particular lawful source. Both parties agreed that network management techniques or traffic prioritization was allowed by service type and not service provider. The implications of this agreement is that YouTube could compete against FiOS on a level playing field. The small additional charge for a guaranteed quality of experience is well worth it. Revenue gained would be used to continue to invest in continued Internet improvements and a better user experience. It is a pity that the technical press does not grasp this fact. Instead they are intent on listening to governmental organizations trying to expand their control over the Internet or third-parties spreading their socialist agenda. None of these groups have a true financial interest in the Internet. The astute observer will notice that true content providers (mainstream media excluded) are absent form the cacophony of descent. Could it be that they are siding with the largest content provider in the world? Perhaps they realize that these principles could allow them to compete with the large carriers.
Verizon and Google took the initiative to form an agreement between content provider and carrier. The media assumes that any time private companies get together that they are doing evil. They would be wise to keep in mind that private companies, research institutions, and universities built the internet, not the FCC or Congress. Google and Verizon did what bureaucrats and the politicians could not do: take an initial step at defining net neutrality for wireline networks. There are some items that could be refined a bit as other content providers and carriers come on board, and they must tackle wireless networks. Exempting them was their biggest mistake, because they are just as an important Internet access method for many Americans. The IETF should take this draft proposal and expand it to cover all access methods including wireless. Global adoption by carriers and content providers is crucial so innovation and competition can continue on the Internet.
Note: The author does not work for or compensated by Verizon, Google, or any other carrier or content provider. These opinions and facts are based on my two plus decades of experience in the telecommunications industry.