Thursday, April 23, 2009

lack of government rules for Reliability of Internet Service

The National Telecommunication and Information Administration (NTIA) is considering criteria for evaluating the proposals for the broadband grants ($4.7b). I think besides the peak data rate as a main factor, it is essential to include some reliability measures. Various factors may play into the reliability of an ISP. I name two. First, every ISP has some down time during which users cannot access the net. Ok, If this averages to 0.1% in a year, that's fine. But what if the network is down 4 random hours a day every week. Can a bussiness rely on it even if when it is not down, it provides extremely high-speed. Second, the nature of the TCP/IP protocol introduces a random delay to the network access time. You would think this delay is hidden from the users senses. True, only if the ISP meets a minimum set of standards in quality of service. Imagine the internet packet that contains your search request to a google's server takes 30 minutes to get there because AT&T, your ISP, does not have enough routers due to high traffic. Is it acceptable, even if your search results are downloaded at lightening speed? Would AT&T stay in ISP business? Evaluating a company's proposal merely based on the peak data rate they can provide and ignoring the quality of service provided might lead to a system that is not reliable. If we want to promote providing broadband Internet as a tool for economical growth, i.e., if we want people to make money by having broad Internet at their disposal, we need to provide a "reasonable" level of reliability with the service. I think the appropriate metaphor here is a "road". If you want to sell your product to the neighbor city you need to be sure that the road between you and that particular city is not closed. If the road is closed for road work half of the time, or if it is always rush hour on that road, you will probably move out of the area, or move out of that bussiness. If you want to sell your product through Internet you need to know that when you need to be connected, there is a high probability that you will be. In metropolitan areas, or anywhere with enough demand to stimulate competition, the services will evolve to becoming more reliable because the users will naturally choose the more reliable solution. However, the recovery act is targeting un-served and under-served areas, mostly rural, that does not have similar level of demands. When the government is investing in such rural broadband infrastructure, it must enforce regulations that guarantee reliability similar to what is available in metro areas. I found it interesting that the Italian government has already established reliability measures to evaluates ISPs. Some examples include “activation time”, “repair time”, “availability”, “real line speed”, and “packet delay”. So far, I could not find any government guide lines for reliability in the United States. The FCC does seek nominations for a new committee called Security, Reliability, and Interoperability Council. The new committee is going to focus on the reliability of commercial communication systems, but only in cases of emergency and disaster. A similar committee is needed to at least oversee that the large amount of money spent in developing rural broadband access results in reliable broadband internet for America. edited by Behnam Analui


  1. What is the definition of "broadband"? What BWs are these guys talking about? Is this by any chance like a fiber-to-home thing or no? -HH

  2. Based on the discussions I heard at the NTIA public hearing, NTIA is mostly interested in fiber and wireless, for example 4G, solutions. There were two suggestions for the rates:
1)The first suggestion from the FCC/NTIA advisor in the public hearing was to define broadband as more that 5Mbps (down link) and 2Mbps (uplink) for new players who like to apply for grants. 
2)The second suggestion was to define "basic broadband" as a rate of 400kbps-1.5Mbps for downlink and "advanced broadband" as a rate of more than 1.5Mbps. 
These are still suggestions. But if they have proposals for more than 5Mbps, fiber is a good choice.

  3. Just one more comment:
5Mbps might sound unnecessary for the rural areas today. But the government is planing for deployment of a system that will be sufficient for the next decade. Think of 10 years ago. Wasn't 100kbps too good to be true?
    The government funds are going to be granted from now until September 2010. The projects are going to be completed in two years, i.e. 2012. So we are talking about a deployment that is ready to be used in ~4 years. 
One of the proposed ideas that was received very well in the public hearing was to fund projects that support dynamic increase in the data rate without huge cost.

  4. Clearly I have a bias here....but really the most cost effective way to provide quality broadband coverage across the country is via satellite. Yes..I know geo synch satellites have a 0.5sec round trip delay, but thats just latency, doesn't hurt streaming media (youtube, netflicks etc) at all.

    Read about our latest venture...

  5. I think the NTIA has to realize that with $7B you are not going to provide fiber to every home or even WiMax/LTE to everyhome. If the goal is to provide good broadband service to every American home with a total budget of $7B, Satellite Based broadband has to play a prominent role. What do you think?

  6. I can, at least, tell you that I have not met any lobbyist from Satellite companies in the NTIA public hearing. You might want to be the first one!

    At least $200 million will be available to upgrade technology and capacity at public computing centers, including community colleges and public libraries. But in my opinion Satellite does not have enough bandwidth for those centers. (why do you think?)

    The remaining applications will be residential buildings and small farms. For them, I think, the main concern is the cost. What is the cheapest price tag you can imagine for the Satellite service that your press release explains? (both upfront and monthly)

  7. Besides, the purpose of the stimulus package is to reduce the cost of broadband by promoting competition. The funds are going to be allocated to small companies to generate competition for large corporations. Satellite providers do not fit in the category of small businesses.

  8. you can check out today's prices at:

    the next gen is far more cost effective....but today's prices can serve as an upperbound. The next gen satellite that we are launching will have higher rates for cheaper prices...exactly what the price will be, I don't know.


    So the terrestrial Fiber to the Home solution is prohibitively expensive. Note that the above estimate is for UK..a much smaller country than the US.

    Do you have an estimate on how much it will cost to deploy WiMax/LTE to everyone in the US?

  10. Not really, but what know is:
    Open Range Communication ( is planning to build 4G (not clear which technology) for over 500 communities in 17 states. They are announcing on their web site that they got $267 million in loan from the federal government and $100 million from a private investor. So the only estimate I can give you is $367 million for 500 communities or roughly 1 million for each community.

    check out this map