ZeuBlog

Really Big Numbers

By Kevin Walsh

With really, really big numbers being casually bandied about in both government and networking circles, it is instructive to put these numbers in context. A recent article on GigaOm by Stacy Higgenbotham does just that with regard to uploading massive databases to cloud computing sites. Since most non-math majors have difficulty comprehending the scale of really big numbers, a useful tool is to express it in terms of something they can grasp, and time is one of the easiest ways to do that.

It appears that the smart people at Amazon, which is investing heavily in cloud computing, have determined that even a state-of-the-art (10 Mbps) broadband connection would require 13 days to upload a 1 terabyte database. Since even the US Postal Service can easily beat that time, Amazon has launched a service called AWS Import/Export that uses, yes, regular mail to upload large databases to cloud-based storage facilities. Sounds like one more reason to engineer better broadband networks.

Pop quiz: the US federal government is expected to spend $4 trillion dollars this coming fiscal year. If they could spend $100 every second, 24 hours per day, 365 days per year, how long would it take to spend that $4 trillion (you can disregard the extra day in leap years, it’s not going to help that much)?

The answer is here.

It’s a TV Thing

By Tom Meehan

There is currently a good deal of debate in regulatory circles concerning the differential treatment of certain traffic types—video being the primary example—over broadband connections. On one hand you have broadband service providers (BSPs) claiming that unless they step in and provide service quality guarantees to traffic that is currently traveling over best-effort networks not built for video, the consumer viewing experience will be decidedly dissatisfying. On the other hand you have a mixed group of people (academics, non-profits, axe-to-grind think tanks, etc.) that insist all IP packets traversing the internet should be handled in precisely the same manner, even in the face of congestion.

Much of this debate concerns itself with web traffic delivered to computing devices (and I’d include smart phones in this category). But what we actually have today, with web-based content storefronts such as Apple iTunes, Amazon Video on Demand, and Netflix, is a video service delivered to consumer TVs. It’s a TV service, not an internet service. In this context the regulatory treatment of video traffic traveling over content delivery networks (CDNs), through BSP-provided broadband video channels, into customer premises digital media players (such as Roku), and onto 50-inch flat panel HDTVs should be no different than IPTV, cable TV, or direct broadcast satellite (DBS).

A New Game in Town?

By Kevin Walsh

OnLive Gaming, which has been in stealth mode for seven years, recently unveiled what it bills as “the future of video gaming.” Although the service won’t be available until this Winter, it has already grabbed a lot of attention due to its radical structure—all game computing and graphical rendering is done “in the cloud” with encoded video streamed to computer- or television-based players. This approach creates an environment in which OnLive gamers will never be compute bound. With, presumably, massive cloud-based data centers crunching away on the latest version of, I don’t know, F.E.A.R. 2, OnLive can allocate as much compute power as required. While this neatly fixes one problem, it introduces another—indeterminate network performance.

(Click image to view full size)

Not unlike IPTV environments, OnLive encodes the graphical output and streams video, SD or HD, to the gamer. OnLive claims that 1.5 Mbps is sufficient for SD and 5.0 Mbps for HD, which should, in theory, meet most definitions of “modern” broadband connections…

Analyst Guest Blog: What Broadband Crisis?

By Teresa Mastrangelo, Principal Analyst, broadbandtrends

Much has been written about this apparent broadband crisis occurring in the United States.  Not only are we supposedly falling behind other countries in terms of broadband penetration but apparently in speeds and pricing as well ($ per Mbps).

Let’s set the record straight.  The United States is the 2nd largest fixed broadband market in the world, behind only China and well ahead of Japan.  Broadband penetration, as a percentage of households (the better method of calculation, rather than using population like the OECD), is at nearly 60 percent.  According to the FCC’s most recent High Speed Internet report, more than 65 percent of subscribers receive speeds greater than 2.5Mbps, with the majority falling into the 5Mbps category; and let us not forget that broadband is available to approximately 88 percent of households.  If we include mobile broadband as well as WiMAX and Satellite broadband, availability is significantly higher, probably in the range of 96 to 98 percent.

Is speed a problem?  I guess that depends where you live and what level of competition is in your market.  In high competition markets, speeds of 50-100Mbps are now commonplace, while low or no competition markets are likely to only see speeds of 5-15Mbps.  The same is true for price – in markets with heavy cable competition Verizon offers its highest tier at $89.95, while in less competitive markets the same product cost $144.95 per month.