Thursday at 2pm ET, I’ll be moderating a StreamingMedia.com webinar on the topic of “Engineering an Enterprise Video Network“. Whether you’re starting from scratch, building on an existing internal platform, or looking to an outside vendor, the challenges facing video engineers and IT professionals in the enterprise are more complicated than ever before in an increasingly bring-your-own-device environment. The roundtable will offer a highly technical deep dive into the nuts and bolts of an effective enterprise video network, from support for legacy formats and protocols to security and SharePoint integration, as well as the pros and cons of on-prem, cloud-based, and hybrid solutions. You can register and join this webinar for free.
- Thursday Webinar: Engineering an Enterprise Video Network
- OTT Services By Apple & Verizon Won't Have A Big Impact On Akamai, Here's The Numbers
- New Whitepaper: The Performance Challenges Unique To Mobile
- Cisco Announces New Royalty Free Video Codec Project To Rival HEVC
- Companies Should Reject Licensing Terms From HEVC Advance Patent Pool
- Streaming Media West Program Published: Now Placing Speakers
- Telstra Plans To Spin Out Ooyala In Public IPO, Setting Unrealistic Expectations
Last week I saw two reports published by Wall Street analysts suggesting that Apple’s still non-existent live OTT service will mean a lot of revenue for CDN provider Akamai. Of course neither of the reports gave out any numbers, provided any estimates of revenue or even showed the formula anyone can use to get a rough idea of what the business would be worth. These reports need to stop being so vague and provide real numbers considering it’s easy to estimate the value of the business using very simple math, which I detail below.
When it comes to the topic of Apple’s own CDN, I still see quite a few who says that Apple is “reportedly building out its own CDN network”, and they speak about it as if it’s speculation. Apple is not “reportedly” building a CDN, they already built it and continue to expand it. Don’t take my word for it, simply do a traceroute and look at how Apple is delivering their content. Apple has a CDN, with massive scale, which is not debatable. Suggesting otherwise, or implying that they “might” have one is simply ignoring the facts.
In one report an analyst said, “we believe that Apple has used Akamai’s CDN capabilities in the past.” You believe it? Why don’t you confirm it? Doing a traceroute shows Apple is still using Akamai for a portion of their content delivery. They also say that Apple “requires significant scale that its own CDN may not yet be able to handle,” but don’t say what that statement is based on. Stop using generic and vague phrases to make it sound like you know what you are talking about when you have no idea at all as to the size and scale of Apple’s CDN. And while we are on the topic of OTT services, the launch of Verizon’s new OTT service will not be a “big deal” for Akamai or provide any “upside” to the company as some suggest. Verizon’s new service will run over their own CDN EdgeCast and not on Akamai.
Whenever Apple rolls out their live OTT service, there are many ways they can handle the distribution of the video. They could use multiple third-party CDN providers like Akamai and Level 3 to deliver everything, or they could use them in combination with their own CDN, splitting traffic amongst multiple providers. Apple could also use their own CDN as primary, and then use third-party CDNs for overflow and extra capacity. Apple could also simply not be involved with the delivery at all and let the content owners in the OTT service deliver the streams themselves, via contracts they already have with third-party CDNs. There are a lot of ways the video delivery can be managed and based on Apple’s history, it’s a safe bet that multiple providers would be involved.
But for revenue calculations, let’s just assume Apple uses Akamai to deliver 100% of the streams. The average HD stream delivered to a TV is about 3Mbps. To tablets and phones, it’s about 700Kbps. We don’t know how many users Apple would sign up for their service, so I’ll use a number of 3M subs in the first month. Then we need to calculate how many hours per month each sub would use and for this example I’ll use 30 hours a month. Then we have to decide what percentage of playback would be on the TV versus a mobile device. If we assume an even split of the traffic, it means one user, in one month, would consume 24.95GB.
[3000Kbps x 60 minutes = 1.35GB per hour X 15 hours = 20.25GB. For the 50% that comes from mobile, 700Kbps x 60 minutes = 315MB per hour, x 15 hours = 4.7GB]
If the traffic was being priced on per GB delivered, at a penny per GB, the value of one customer would be worth $0.25 to Akamai. Multiply that times 3M users and the value of the contract would be worth $750,000 a month. But, Apple doesn’t pay the CDNs based on a per GB delivered model, they pay per Mb/s (95/5). So if you assume that at any given time only 1M of the 3M users will be watching a stream concurrently, with 50% at 3Mbps and 50% at 700Kbps, it means that Apple would be using about 1.5Tb/s. Even at $1 per Mb/s, which is a high price for Apple to pay, the value to Akamai would be $1.5M a month. The numbers get bigger if you calculate them with more than 3M users a month and/or more than 30 hours a month consumed per viewer. But Apple won’t get 3M subs in the first month and there is almost no chance that Apple would give Akamai 100% of the delivery.
Part of the speculation of Akamai’s involvement with Apple on a new OTT service is due to Akamai’s recent earnings call, where the company said that their CAPEX spend would be higher than normal as they added more capacity for OTT video. But Akamai does not break out the percentage of money spent per service and at no time did they say they were adding capacity for a “new” OTT service. All CDNs, including Akamai, don’t keep a ton of extra capacity available just sitting around unused. They operate their networks with a certain level of overflow capacity, but not more than necessary so as not to incur costs for capacity they aren’t using. Akamai could simply be looking at their customers traffic growth and spending money now to put in place the capacity they need for the overall growth of the OTT business for the next 12 months.
Apple’s OTT service will not be a big revenue driver for Akamai or any other CDN right out of the gate. If Apple signs up tens of millions of subscribers and gives a large percentage of the delivery to third-party CDNs like Akamai, then the numbers could start to add up. But Apple would be lucky to get a few million subs, within a few quarters, with what is expected to be such a pared down selection of content available. Based on my calculations, Apple is already spending between $75M-$100M with Akamai this year. So adding a few hundred thousands dollars more a month to that for a new OTT service, or even $1M more per month, it really doesn’t have a big impact on Akamai’s overall numbers.
Frost & Sullivan has just released a new white paper entitled “A Billion Reasons for Inconsistent Mobile Performance and How to Solve for Them“, which is sponsored by and includes data from Twin Prime. Mobile is now the first screen of choice. It’s the most personal device we have had and is rarely a few feet away from us at any given time. Consequently, the gratification amount and user intent is highest on smartphones than any other computing device. However, many times our mobile experience is poor and can be as bad as content not even loading.
The key question about mobile performance is why does it continue to suffer? What is it that makes mobile so uniquely different from the desktop/PC world? Mobile networks face many challenges that make it hard to deliver a consistent and rich user experience. These challenges are unique to mobile and include:
- Last-mile latency inordinately affects mobile performance
- Volatility in the wireless last mile trips transport protocols
- Multitude of last-mile networking technologies hinder performance
- Diversity in last mile creates infinite performance scenarios
- Mobile content severely constrains the efficacy of current acceleration solutions
- Current acceleration strategies were built for Web, not mobile apps
- Server-based solutions are insufficient and suboptimal
- It’s not about a new protocol
In every benchmarking study, mobile performance is at least three times as slow as the Web. For example, Keynote Systems reports that today, on average, an m-commerce app/site loads in about 8 seconds on our mobile devices. This compares poorly with 2-second load times for websites. Read more about mobile performance challenges by downloading the white paper for free from Twin Prime’s website.
With two HEVC patent pools now in the market and a third one being formed, the adoption of HEVC as the successor to H.264 is going to start getting very expensive. With that problem in mind, last week Cisco announced a new video codec project they call Thor. Cisco’s goal is to work with others to provide an alternative to HEVC, with the same or better quality, with no royalty required. While Thor is in the very early stages and is not an alternative to HEVC today, Cisco’s vision is to rally others in the industry to contribute to the project and help make higher-quality video cheaper to deploy, without the need for expensive licensing. As Cisco points out, the total costs to license H.265 from the two current pools in the market today (MPEG LA and HEVC Advance) is up to sixteen times more expensive than H.264, per unit. In addition, Sony, Panasonic, Qualcomm, Nokia & Broadcom all have extensive patents around HEVC and aren’t in either pool. Rumors have it some of them are working to form what would be a third patent pool in the market to license their HEVC technology.
Cisco has hired patent lawyers and consultants familiar with video codecs and created a development process which they say allows them to work through the long list of patents in the space, and continually evolve their codec to work around or avoid other HEVC related patents. Two weeks ago Cisco open-sourced the code and contributed it to the Internet Engineering Task Force (IETF), which already has a standards activity to develop a next-generation royalty free video codec in its NetVC workgroup. This is a group Mozilla has been active in and has been working on technology they call Daala, which they want to be the successor to HEVC.
Cisco is no stranger to video codecs. The company faced a unique challenge with H.264 as they and other players in the video conferencing industry wanted H.264 included in the webRTC standards in order to facilitate interoperability with their install base of conferencing equipment. However, Mozilla simply couldn’t ship H.264 due to both the licensing fees as well as incompatibilities between MPEG-LA terms and its open source nature. To fix this, Cisco offered to help with an interesting solution.
Cisco open sourced its H.264 implementation (www.openh264.org), but more importantly, it also agreed to make a binary build of its implementation available as a module. Mozilla produced a version of Firefox, which upon installation, fetches this module from Cisco and links it into Firefox. In this way, Cisco is the distributor of the module and has to carry responsibility for the license fee. Cisco agreed to foot the bill and paid the full cap to MPEG-LA. This eliminated the need to track downloads of the module since tracking of downloads would be in violation of the privacy considerations that Firefox and the community had. So in the end, thanks to Cisco, everyone won. Cisco and the videoconferencing community got H.264 into the Firefox (and later into the webRTC standards), and Mozilla got H.264 support without needing to actually ship it in Firefox.
Thor is a project; it’s not an actual video codec yet and Cisco and others have a lot of work to do, probably years, before it’s a viable alternative to HEVC. Cisco wants to start getting the word out about project Thor with the hope that others will contribute intellectual property and that technical experts with video codec knowledge will get involved. Hopefully Google takes note of this and contributes VP9 into the NetVC workgroup, which would be great for everyone. The video codec standards development process benefits heavily from having multiple contributions into the process and then evolving them to take the best of all of them for the best overall result. As Cisco points to as an example, the Opus audio codec got to where it is today in exactly the same kind of way; combining two pretty different codecs, Skype’s SILK codec and Xiph.Org’s CELT codec.
It should be noted that both Thor and Daala are far from acting as alternatives to HEVC today, but it shows that when faced with greedy patent pools, companies like Google, Cisco, Mozilla and others will seek out or create alternatives. HEVC patent pools should take note. If they push too hard and get too greedy, they could be outmaneuvered.
Last month newly formed patent pool HEVC Advance announced their licensing terms for patents pertaining to HEVC video. [See: New Patent Pool Wants 0.5% Of Every Content Owner/Distributor’s Gross Revenue For Higher Quality Video] While the pool has yet to detail any of the actual patents in question, or make it clear exactly whom they feel is required to pay for licensing, they still plan to open up for licensing deals on October 1st. Each company has to do their own due diligence with regards to HEVC Advance, but my recommendation is not to agree to HEVC Advance’s licensing terms. There is no rational business case to agree to their terms, especially when there are no details on which patents will be included or what technologies they cover. And considering the patents are only just starting to be independently evaluated this month, there is no way to know the validity of any of the patents in the pool.
If the industry collectively refuses to pay the licensing terms HEVC Advance has put forth, the licensing group will be forced to change terms to make them fair and reasonable, disband, or the patent holders themselves will have to individually file suit against every broadcaster, OTT provider, CE manufacturer and everyone else they feel infringes. That’s going to be nearly impossible to do considering there are tens of thousands of companies that HEVC Advance wants royalties from. Patents holders “expected” to be in the pool including GE, Technicolor, Dolby, Philips, and Mitsubishi Electric would have to take companies to court and it would be a long, expensive and drawn-out process. The average patent case takes five years to resolve. Note that HEVC Advance won’t say who for certain is in the pool and keeps using the term “expected” because even they admitted to me that not all the companies mentioned have completed the paperwork to be in the pool. What a circus.
While some might think there is no way HEVC Advance would change terms or simply give up, content owners and broadcasters have forced this exact result in the past. If the approach HEVC Advance is taking sounds familiar, it’s because we’ve seen this before, with some of the same players. Some might remember a company called Via Licensing which including patent holders Dolby, Philips, and Technicolor. They came to the market with patents for DVB-MHP, targeting the broadcasters and set-top-box makers. The also had the same style of a secret essential patent list and vagueness and got nowhere with their patent pool. Via Licensing forced many to abandon further development of DVB-MHP, but the broadcasters and operators sought an alternative, and the zero royalty HbbTV technology was born. A reader also reminded me that the notion of going after content owners and distributors was also tried with MPEG-4 Part 2, the original MPEG-4 video compression which pre-dated AVC. It failed in large part because of this licensing tactic, which is why the H.264/AVC licenses were more sensible.
The companies in HEVC Advance should have learned from history. Broadcasters, content owners, and distributors have shown they will seek out alternatives when these kind of greedy patent pools come to the market. Companies could bypass HEVC all together and start using Google’s VP9. Another interesting development is that Cisco just announced a royalty free video codec called Thor, specifically to combat all the licensing terms around HEVC. So for the companies in HEVC Advance, you’re being put on notice that alternatives do exist. Your greed will be your downfall if you don’t listen to the market.
If the industry sticks together, I believe that HEVC Advance’s unfair and unreasonable royalty rates can be defeated. Companies like Dolby can’t afford to upset Hollywood and content owners, which is what they are starting to do. I’ve already heard from multiple major studios who said if Dolby want’s to try to enforce unfair royalty terms around HEVC, they will hit back and do everything they can not to licensing Dolby’s audio technology across the board. There are alternatives. Not to mention, Dolby is also telling these same studios that they will soon want royalty payments pertaining to HDR patents as well. Dolby is not making any friends right now and if they then try to sue content owners and CE manufactures that they currently have as partners, Dolby will face some serious trouble.
It should be noted that all of my inquires to Dolby, Technology and Philips to learn more about their involvement in HEVC Advance have gone ignored and HEVC Advance hasn’t made any patent holder available for the media to talk to, even though it’s been requested. The companies in the pool are making a big mistake by allowing GE Licensing to speak on their behalf. Patent attorneys don’t know technology, which is evident by the multiple calls I have had with HEVC Advance. When they can’t answer basic questions around the technology, or answer most questions by saying “we’ll work something out”, the pool loses the ability to create any creditability. If you feel your patents are strong and your terms fair, then communicate that. Because right now all you are doing is communicating that you’re in a disorganized pool that isn’t educated on the market or how the technologies even work. And no I’m not being harsh, it’s the truth.
To defeat HEVC Advance’s terms, all it will take is some major companies publicly coming out to say they don’t plan to pay, based on the current rates. If a few major companies do that everyone else will follow suit. Another alternative is for an industry association to have all their members agree to speak out against these unfair and unreasonable terms and that would stop HEVC Advance in their tracks. I am calling on the industry to do just that. Put aside your competitive offerings and be united – stand together. Let me know how I can help. I take it personally when any patent pool comes into the industry with unfair practices and licensing terms, which has the serious potential to stunt the growth of the market. Not to mention, a third patent pool around HEVC is already forming.
To the companies in the HEVC Advance patent pool, I challenge any of you to a debate on your patents, your rates, and your approach to the market. I’ll give you the stage at our next Streaming Media West show in November to state your case as to why anyone should license from the pool. The ball is in your court.
Updated 8:43pm: Technicolor replied to me to say that all communications requests are answered by HEVC Advance directly.