Two professors from the Department of Electrical and Electronic Engineering at the Technical University of Bari, Italy have released a white paper detailing their testing of Akamai's HTTP-based adaptive video streaming technology. The paper includes a lot of charts and graphs on their testing methodology as well as the results. They analyzed the client-server protocol employed in order to actuate the quality adaptation algorithm and they evaluated the dynamics of the quality adaptation algorithm in three different scenarios. Some of the highlights of their findings are:

• video is encoded at five different bit rates and each level is stored at the server
• the video client computes the available bandwidth and sends a feedback signal to the server that selects the video at the bitrate that matches the avail- able bandwidth
• the video bitrate matches the available bandwidth in roughly 150 seconds
• a feedback control law is employed to ensure that the player buffer length tracks a desired buffer length
• when an abrupt variation of the available bandwidth occurs, the suitable video level is selected after roughly 14 seconds and the video reproduction is affected by short interruptions.

The same authors have also released their findings in another paper on the subject of Feedback Control for Adaptive Live Video Streaming. I am not a networking expert or engineer and don't know the technologies down to this level. So while I am not endorsing the findings of either of the papers I'm sure others who know more about this than I do will have some comments to share.

Netflix's domination is about to truly go global. While Netflix has been saying that they will expand internationally in 2011, the company has yet to release any info about which location they plan to target next. While I don't have all the details or know all the specifics, it looks like South America is the next territory that Netflix will launch their streaming based subscription service.

In order for Netflix to expand their business and keep up their current rate of growth, the company needs to enter countries that have a large population capable of getting a 2-3Mbps video stream. While not every country in South America fits that bill, many do, and they have huge populations with large Internet penetration rates:

• Argentina: As of March 2010 the country had a 64% Internet penetration rate with 26.6M users with an average broadband speed of 3.33Mbps. 
• Brazil: As of Dec. 2009 the country had a 37.8% Internet penetration rate with 75M users with an average broadband speed of 4.46Mbps.
• Chile: At the end of 2009 the country had a 50% Internet penetration rate with 10M users with an average broadband speed of 6.62Mbps.
• Colombia: In mid 2010 the country had a 48% Internet penetration rate with 21M users with an average broadband speed of 4.32Mbps.
• Mexico: In 2010 the country had a 27% Internet penetration rate with 30.6M users with an average broadband speed of 3.54Mbps.
• Peru: As of June 2010 the country had a 27% Internet penetration rate with 8.8M users with an average broadband speed of 4.62Mbps.

In 2010, South America had a estimated population of just under 400M with 156M Internet users. Central America had a estimated population of 154M with 38M Internet users. Depending on which territories exactly Netflix launches in, the company has the potential opportunity of expanding into a new market with a combined population of more than 500M users, with just under 200M of them online, with a combined average broadband speed of 3.2Mbps.

From what I am hearing, the new Netflix service is already being tested in certain regions and is expected to launch shortly. Last week, on Akamai's earnings call, the company commented that they had signed a "new" deal with Netflix saying that they, "expect to work closely with them to leverage our globally distributed network to their market expansion." While Akamai has been doing caching of small objects for Netflix for some time, to date, Netflix has not had a reason to deliver content outside the U.S. or Canada. Netflix's new contract with Akamai, which I don't believe to be exclusive, is just another sign that Netflix is about to launch out into International markets very soon.

There are a lot of questions about what type of content Netflix will offer and while I don't know those details, I do know that some content will be in-country and other content will come from the U.S. and be streamed to the new territories. From what I am hearing, it sounds like we should see the new service launch Internationally before the end of March.

* Data on broadband penetration, population and download speeds were compiled from Wikipedia, the US Census Bureau, AMIPCI, ITU, eMarketer and SpeedTest.net

Note: I didn't contact Akamai asking for a comment as I knew they would not be able to give out any information and while Netflix did return my inquiry, they were not willing to confirm or deny anything.

Earlier today, signs went up at my local Blockbuster store saying the location was closing and that everything was now on sale. I decided I would pick up a few games and DVDs on the cheap, but boy was I wrong. Of course I should of known that this is Blockbuster and they don't do anything that makes sense, so I probably should not have been so surprised to see that Blockbuster is selling most of their pre-owned DVDs for 2x the price that Amazon sells them for new.

As an example, Blockbuster is selling the Deluxe Unrated Edition of Salt for $16.99 for a pre-owned copy. Amazon sells it new for $8.99 and sells the Blu-ray version for $13.99. Blockbuster's store shelves were full of movies priced at $14.99 and $16.99, all pre-owned, and all 2x more expensive than buying them new from Amazon. Apparently Blockbuster doesn't know what movies rent OR sell for these days. I don't like seeing a company go under, it puts a lot of people out of jobs, but in this case, as a consumer, one can't help but be happy to see Blockbuster closing down. Any company that treats consumers as if they are idiots and thinks they don't already know of other options in the market for getting movies cheaper and in better quality, does not deserve to stay in business.

I didn't get to listen to Level 3's fourth quarter earnings call live last week (transcript) and initially missed hearing them disclose, for the first time, actual revenue numbers for their CDN business. Level 3's CDN revenue grew 27% sequentially, compared to 11% in Q3 and the company said that their CDN services now represent about 2% of their total CNS revenue. That puts their total CDN revenue for 2010 at about $60M with another approximately $100M coming from their Vyvx broadcast services.

While we constantly hear about how much growth there is in video and the impact video traffic is having on last mile networks, the fact is that from a revenue stand point, the market for outsourced video delivery is still small. In a report that Frost & Sullivan will publish shortly, we estimate that the total worldwide size of the video delivery market was $545.5M last year. We expect to see a CAGR of about 38% this year, but even with that, the market is still just getting started.

That number does not include all of the different products in the video delivery ecosystem like video servers, hardware, P2P solutions or platforms sold to companies to build out their own CDN network. The estimates are revenue from third party CDNs who essentially rent out their networks for content owners.

While some might think Level 3's CDN number is small, it's easily large enough to make them one of the top five CDNs in the industry based on revenue. Potentially even the top three, but we don't know how much revenue Amazon's CloudFront CDN services has or how much revenue AT&T does. It's a safe bet to say that Level 3 is doing far more CDN revenue right now than AT&T, but Amazon CloudFront is the real question mark.

Level 3 is making video a long-term play for their business and when the industry turns into a billion dollar market, which I estimate to be in 2013, that's when this whole video delivery industry is going to get really interesting.

Added: I forgot to mention that in an era where most companies don't want to go into too many details on their revenue or break out numbers based on product lines, it's nice to see that Level 3 was willing to do it publicly. They didn't have to break out their CDN revenue and they did. Doing so really helps those who are trying to track and project growth in the CDN market and it helps the industry as a whole since it now serves as another data point we can use as a barometer for growth.

On Monday I mentioned that I knew of one company selling some P2P assets and now I can officially talk about the deal. Ignite Technologies has acquired the live peer-to-peer technology from Abacast along with two patents (6,970,937 and 12,265,581). Terms of the deal were not disclosed but I can confirm that the sale price was under $500,000 dollars.

Ignite, which was already working with Abacast for a few years, will leverage the peer-to-peer technology to enhance the peering capabilities of Ignite's client technology for their enterprise customers. Abacast has been out of the peer-to-peer video space for a few years now and is primarily focused on providing streaming, ad insertion and management and analytics for radio stations. With their business so focused on the radio vertical, it's not surprising to see Abacast sell off their peer-to-peer technology. The sale to Ignite is specific to Abacast's live peer-to-peer platform, but I expect they wil also sell off their on-demand peer-to-peer platform as well.

Adaptive streaming over HTTP is gradually being adopted by content owners as it offers significant advantages in terms of both user quality and resource utilization for content and network service providers. To understand whether today's existing commercial players perform well, especially under dynamic network conditions, Cisco and The Georgia Institute of Technology just released a technical white paper on the subject. Their experiments covered three important operating conditions:

• First, how does an adaptive video player react to either persistent or short-term changes in the underlying network available bandwidth? Can the player quickly converge to the maximum sustainable bitrate?
• Second, what happens when two adaptive video players compete for available bandwidth in the bottleneck link? Can they share the resources in a stable and fair manner?
• And third, how does adaptive streaming perform with live content? Is the player able to sustain a short playback delay?

The paper identifies major differences between Microsoft's Smooth Streaming, the player used by Netflix, and one open source player (Adobe OSMF). Their findings report that the Smooth Streaming player is quite effective under unrestricted available bandwidth as well as under persistent available bandwidth variations. It quickly converges to the highest sustainable bitrate, while it accumulates at the same time a large playback buffer requesting new fragments (sequentially) at the highest possible bitrate.

On the negative side, the paper says that the Smooth Streaming player reacts to short-term available bandwidth spikes too late and for too long, causing either sudden drops in the playback buffer or unnecessary bitrate reductions. Further, the experiments with two competing Smooth Streaming players indicate that the rate-adaptation logic is not able to avoid oscillations, and it does not aim to reduce unfairness in bandwidth sharing.

The Netflix player is similar to Smooth Streaming as they both use Silverlight for the media representation. However, the paper reports that the former showed some important differences in its rate-adaptation behavior, becoming more aggressive than  the latter and aiming to provide the highest possible video quality, even at the expense of additional bitrate changes. Specifically, the Netflix player accumulates a very large buffer (up to few minutes), it downloads large chunks of audio in advance
of the video stream, and it occasionally switches to higher bitrates than the available bandwidth as long as the playback buffer is almost full. It shares, however, the previous shortcomings of Smooth Streaming.

The paper says that the OSMF player often fails to converge to an appropriate bitrate even after the available bandwidth has stabilized. This player has been made available so that developers will customize the code including the rate-adaptation algorithm for HTTP Dynamic Streaming for their use case.

There is a lot of technical details in the paper and it makes for a good read if you are interested in the adaptive streaming topic. If you have a question on the paper, leave it in the comments section and I'll ask one of the authors to answer them.

Last Friday, in a letter sent to the FCC, Global Crossing urged the commission to regulate the pricing for connections between last mile providers and network service providers saying that “broadband ISPs are distorting the economics of the historical peering relationships that existed between carriers“. Global Crossing also disagrees with a letter Verizon sent to the FCC in January saying that, “the Internet that Verizon describes is at least five years out of date.”

Global Crossing argues that today, the Internet is no longer comprised of networks isolated to one purpose and that many carriers networks are now integrated with last-mile access, content, backbone networks and CDNs. They go on to say that since carriers are no longer of equal size and now compete much more directly with each other, “the historical cooperative practices that formed the Internet are breaking down and being replaced with more commercially driven tactics such as charging for access to last-mile facilities.”

Global Crossings argument, which I think is a good one, is that this was the same problem that happened in the telephony market when the competitive local exchange carriers (CLECs) upset the peering relationships that existed with local exchange carriers. And while Global Crossing says they recognize that Comcast has not attempted to impose tariff-based access charges in the traditional sense of how telephone companies have imposed access charges, “the assessment of a fee for the delivery of traffic over last-mile facilities is the functional equivalent of an access charge“.

If this were to take place, Global Crossing says, “the ultimate result would be to skew the development of broadband competition while undermining the interests of consumers in obtaining the Internet content that they want, when they want it.”

In the Implementation of the Local Competition Provisions in the Telecommunications Act of 1996; Intercarrier Compensation for ISP-Bound Traffic, Global Crossing points out that the FCC specifically “sought to guard against carriers’ improperly shifting all of their costs to other carriers by charging them excessive rates for the use of their networks“. The Commission invoked that general principle in addressing the terminating access rates charged by CLECs, which Global Crossing argues is a scenario that is, “substantially similar to the one presented here“.

Global Crossing’s real argument is that the predicament faced by parties that send traffic to a broadband ISP’s subscribers today is substantially similar to those which the Commission has not hesitated to remedy in the past. Global Crossing says that, “the Commission has refused to tolerate such practices in the past” and that, “the application of access charges without any regulatory oversight would thoroughly undermine all of the Commission’s efforts in the instant docket to preserve an open Internet.”

While many have been quick to argue that they don’t want the government involved in this disputer, Global Crossing ends the letter by saying that, “the Commission can ameliorate the problems associated with terminating monopolies in today’s Internet and preserve an open Internet without complex and extensive regulation. The Commission need only apply its existing, long-held principles of cost-recovery and safeguards for terminating monopolies in the context of the instant dispute.”

This topic is a hot debate in the industry right now and each time it comes up, many networking and technical people want to weigh in on how the Internet and peering works or want to explain the complexity of the Internet. But the bottom line is we need more business people weighing in on the topic as the heart of this debate is not about technology or network architecture but rather the impact this could have on business models and consumers.

Similar Posts:

– The Real Issue In Comcast’s Dispute With Level 3 Is About Power, Not Money

– Comcast Says Their Dispute With Level 3 Is “Not About Online Video” – Yeah Right

– Level 3’s Lower Cost Comes From Owning The Network, Not Free Peering