All of the presentations and sessions from the Content Delivery Summit are now online and available for viewing on the show website at www.ContentDeliverySummit.com.
You can also download all of the presentations from the event on the agenda page. You may use use or embed the videos any way you like. The Twitter hashtag for the event is #cdnsummit
I was hoping I wasn't going to have to write this post today, but it's clear that far too many people on Wall Street either aren't reading the details of what Netflix announced, or are simply trying to create a panic for their own benefit. I've see quite a few reports this morning from those who put out coverage on Akamai and while some of them are on point, many of them are full of inaccurate information. I've also gotten a lot of questions this morning about Netflix's CDN news, so I will try and answer those questions in this post as well. [updated: Level 3 has just posted their thoughts on the Netflix news]
At the time of this post, Akamai shares are down almost 6.5% on the Netflix news. While I'm not one who speculates on share prices or gives out any kind of share guidance or recommendations, the real impact on Akamai by Netflix building their own CDN is minimal. Far too many on Wall Street want to imply this has a big impact on Akamai's business or that this is the start of a new trend of content owners taking their video delivery in-house. This could not be further from the truth. We have public data in the market to prove it, yet many want overlook that data to support their own agenda to investors.
From what I know, the vast majority of Netflix's video traffic in North America is delivered by Limelight and Level 3. While Akamai does have some of the U.S based traffic, it's not a big percentage. I estimate that Limelight and Level 3 deliver more than 80% of Netflix's U.S. based video traffic. While Akamai delivers a larger portion of Netflix's video outside of North America, Netflix still does not have many subscribers outside the U.S. so the overall traffic volume isn't huge. In addition, these caches being deployed by Netflix are for video traffic only. Netflix also uses Akamai for other services like small object delivery and these caches only support the delivery of video content. So none of Akamai's small object delivery business, or Amazon's AWS business with Netflix is impacted in any way, which was something Netflix confirmed for me during my call with them yesterday.
I've also seen some suggest that Akamai lost an opportunity to sell Netflix their managed or licensed CDN platform. It's clear that anyone who implies that does not have an understanding of what a licensed and managed CDN platform is for or how it works. Licensed and managed CDN platforms can only be deployed inside a carriers network. They are licensed to network operators who own their own network. Since Netflix does not own a network and is placing their caches inside third party networks, there is no way a licensed or managed CDN platform could work for what Netflix is building. So Akamai didn't "miss out" on any kind of licensed of managed CDN sale with Netflix.
Next up is the absurd notion I have seen from far too many people implying that over time, Netflix will be in the CDN business competing against vendors like Akamai and will deliver content for services outside of Netflix. Statements like this show a complete lack of knowledge of how caching works and what Netflix's caches do. Not only are Netflix's caches setup to only proactively cache Netflix's content, but they are not demand driven caches. Netflix caches use a pre-population method that works for a set library with a known set of subscribers, but would not be appropriate for a general-purpose CDN or caching system. So Netflix's caching platform, as it stands now, would not even work to cache content the way ISPs do it today with transparent caching platforms.
I've seen some suggest this is a new trend in the market, content owners delivering video themselves, and then they mention Facebook or Apple as those who may follow Netflix's route. Why that is possible, and likely, that's not a "trend". How many companies are there the size of Google, Apple, Facebook, Yahoo!, Netflix or Microsoft? Very, very few. Not to mention, most of Facebook's traffic is small object delivery, not video. There are maybe half-a-dozen content owners who are delivering enough volume of bits, have the technical expertise and have the money to build out their own CDN. If half a dozen companies end up doing this themselves, out of thousands of CDN customers, that is not a "trend".
Many have stated that in addition to Akamai, Limelight Networks is also at risk since Netflix accounted for about 11% of Limelight's total revenue in 2011.While that is accurate, Limelight has had this happen before with both YouTube and Microsoft taking their traffic off of Limelight's network over an extended period of time. I'm not implying that Limelight will automatically be able to replace that revenue over the next 18 months time, they will have to work to do that, but it's not something they are unfamiliar with and all of the CDNs have known of Netflix's plans for some time. While I saw one report say that Netflix's news "could deplete revenue for third-party CDNs", that's a completely overblown statement.
I've also seen members of the media get into the act by saying third-party CDNs could see, "sizable amounts of revenue depleted". If Akamai loses 1% of their revenue, over an 18 month period of time, from one customer, is that a "sizable amount"? Of course not. But the people writing these pieces have no clue how the business works and what the real numbers are, or else they wouldn't speak in such general terms.
This reaction to Netflix's news by many on Wall Street reminds me of a similar reaction Wall Street had when in December of 2007, AT&T announced they would spend close to $70M dollars to grow their streaming and caching services. Akamai's share price fell more than $2 dollars that day, and tons of Wall Street analysts downgraded the stock saying that AT&T would now move into the market and take share from Akamai. Six months later, AT&T made another CDN announcement that they would build out their CDN to have 400Gbps of global capacity by the end of the year. Rather than look at real numbers, like the fact that 400Gbps of capacity would only be 20% of the network capacity Limelight had at the time, Wall Street once again implied it was doom and gloom for CDN competitors. Of course that didn't happen and five years later, AT&T isn't any closer to impacting any of the major players in the CDN space.
Far too many Wall Street analysts want to create uncertainty, panic and confusion when news like this comes out. I wish they would focus more of their time and efforts on learning more about how these technologies actually work, how they are deployed, what they cost and better understand the competitive landscape. If you cover CDN vendors in this space and you don't know what DSA or FEO stands for, don't know how transparent caching works, why or how a licensed or managed CDN is deployed, or what services the CDNs actually offer, then you really can't expect to be taken seriously when you publish coverage that is heavy on fiction instead of facts.
Note: Not all Wall Street analysts do a poor job in their coverage of the CDN space. Some know the space well. But ever since about a year ago when Akamai's share price dropped, many new analysts have jumped in to cover CDN vendors and they have no expertise or insight into the CDN market at all. As a result, the quality of the coverage put out on CDN vendors as a whole has really suffered in the past 18 months.
Related Post:
The content delivery market is rapidly changing with new products and services coming to the market including licensed and managed CDN, dynamic site acceleration, front-end optimization, mobile content acceleration and transparent caching. If you're a bit confused by all of these different technologies, I've provided definitions of some of the various technologies below.
Today, Netflix announced that they have been actively working to build out their own network of caches inside ISP networks and have officially launched the "Netflix Open Connect Content Delivery Network". With this offering, Netflix aims to lower their CDN costs, rely less on third party CDNs, provide higher quality streaming and most importantly, give network operators more control over the video that flows through their pipes. While their blog post doesn't give out too many details, I had the chance to speak with Netflix earlier today to get some more information on their announcement.
To support the launch of their CDN, Netflix has a new website at openconnect.netflix.com which gives ISPs more details on the hardware and software design of Netflix's caches, details on how ISPs can peer with Netflix and access to a deployment guide. While many might compare what Netflix is doing with their cache deployment to what Google has been doing for years, the big difference is that Netflix is giving ISPs control over the caches, allowing them to adjust the volume of traffic flowing through their network. This is a smart move on Netflix's part as many ISPs don't view Google caches as being friendly because once placed inside the operators network, they can only be controlled by Google.
Under a typical scenario, one Netflix cache can handle about 8.5Gbps of throughput and can support about 25,000 subs. While Netflix is going after the major ISPs for partners, they did say that for any network with more than 100,000 subs it would make sense for them to deploy caches. Netflix said the caching technology they are using was all built in-house and they haven't licensed any content delivery technology from any third party. Netflix said that based on the caches they already have deployed, about 5% of their global video delivery is already being delivered via their Open Connect Content Delivery Network, with some locations like the UK, seeing as much of 50% of their traffic coming from their new CDN platform. Netflix plans to serve nearly all of their video globally via this new CDN, but estimate it will take a few years for them to build it out. I would compare this approach to a similar one Microsoft had, which saw them bringing the majority of their CDN in-house over a few short years.
While some might suggest this is a new trend in the market, content owners delivering video themselves, it's not. There are maybe half-a-dozen content owners who are delivering enough volume of bits, have the technical expertise and have the money to build out their own CDN. Only companies the size of Google, Apple, Microsoft, Netflix and Facebook can take on such a task. Netflix would not say exactly how much money they would save from using their new CDN over third party CDNs, but even a small fraction of savings per Mbps would be substantial to Netflix since they peak at multiple Tbps of video every day.
Something that's pretty unique to Netflix's cache deployments is that they are not demand driven caches. Unlike most content owners who see maybe 80% of their traffic come from only 20% of their content, Netflix gets traffic from all of their content. So instead of someone having to request the content before it is fetched and placed on the caching server, which is how most CDNs work today, Netflix is actually pre-populating the caches inside the ISP's networks. During peak hours, Netflix's caches are in read only mode and during off peak hours, they get populated with content.
Naturally, this news is going to make a lot of people question what impact this will have on Akamai, Limelight, Level 3 and Amazon, now that Netflix plans to do video delivery themselves. For starters, this news does not impact Amazon in any way as Netflix has built a bunch of their non-video platforms on top of Amazon Web Services (AWS), which won't be impacted. For Limelight and Level 3 who have the majority of Netflix's video traffic, they will lose most of this traffic over the next eighteen months, if Netflix deploys their caches as expected. About eighteen months ago, Netflix signed three year contracts with Limelight and Level 3 and Netflix said their don't plan to break any contractual commitments with the CDNs.
While some might take this as negative news for the CDNs, it's actually the opposite. In particular for Limelight and Level 3, they have spent a lot of money to build out their networks to support Netflix, and most of the revenue they get in return simply offsets their costs. Netflix's video traffic isn't really profitable to the CDNs and it's one of the reasons you have seen Limelight announce they would cut back their CAPEX spend by nearly $15M this year, when compared to 2011. With Netflix doing their own delivery, the CDNs can now spend money to build out services where they can actually see good margins and don't have to compete so hard on price.
Of course vendors like Limelight and Level 3 are going to have a lot of revenue to replace in the next 18 months as Netflix contributed to just over 10% of Limelight's total revenue in 2011, but they have been through this shift before with YouTube and Microsoft. If the CDNs don't have to spend the money to deploy services that are about break even, it means they can spend their time and effort towards more profitable business, which is good for their bottom line. So for anyone who thinks that by Netflix bringing their video delivery in-house means doom and gloom for the CDNs, it doesn't.
Building out their own content delivery network is a smart move for Netflix as over time it will reduce their costs, enable them to help ISPs have more control of Netflix's content going over their pipes and also improve their streaming quality to consumers. This is good news for everyone involved.
Updated: I've gotten a few emails already asking me which ISPs are participating in Netflix's Open Connect Content Delivery Network. While I know of a bunch, I can't mention them by name since so far, none of the ISPs have given me permission to make that info public.
Related Netflix Traffic/Pricing Posts:
- Netflix's Streaming Cost Per Movie Drops 50% From 2009, Expected To Spend $50M In 2011
- Netflix Viewers Consume Almost 10 Hrs Of Video A Month, Do Last-Mile Providers Have A Strategy?
- Detailing Netflix's Streaming Costs: Average Movie Costs Five Cents To Deliver
- Bandwidth Pricing Trends: Cost To Stream A Movie Today, Five Cents: In 1998, $270
For more than a few years now, many have been proclaiming doom and gloom to CDNs like Akamai with the reasoning that AT&T and Verizon would come in and disrupt the market, forcing pricing down and taking away market share from the industry leaders. While owning the network does have major advantages, just look at what Level 3 has done, owning the network by itself does not make one a player in the content delivery market.
Two weeks ago at my Content Delivery Summit in NYC, the most common question I got asked by media customers and industry executives was why AT&T and Verizon were absent from the show. While we had just about every other CDN vendor presenting, many with their customers, and had numerous telcos and carriers discussing their deployment strategy, no one from AT&T or Verizon presented or spoke in any capacity.
I offered Verizon multiple speaking opportunities, but they declined all requests saying they weren't yet ready to talk about what they are doing with regards to their external digital media offerings. While it's natural for a company not to want to speak if they aren't yet ready to talk about their technology, it's now been 14 months since Verizon first announced some of their Verizon Digital Media Services (VDMS) at NAB. Since that time, the company still hasn't offered any insight into their products, has no technical information on their website, has not done a press release with a single customer and has refused all briefing requests. And a few months ago, the number they had listed on their website for customers to call to get more information about their Verizon Digital Media Services, didn't work and wasn't a valid number.
While I know Verizon is in fact working on their external digital media services, one really has to wonder if they are getting any traction at all. I hear of them having meetings with large media and entertainment companies, including the big boys like Netflix, but to date, I have yet to hear of a single large customer using them. Amongst many in the industry, Verizon is quickly being compared to AT&T, another vendor who to date, hasn't gotten any real traction with their CDN or digital media services. Unlike Verizon, AT&T has put out a lot of releases about their CDN offering and talked a big game, but has little to show for it. From multiple people I have spoken to who have an insight into AT&T's CDN traffic, it looks as if AT&T will do under $20M this year in CDN services. That's not disruption in the market.
During my pricing presentation at the Content Delivery Summit, I gave out details of a recently completed survey I did of 725 customers who use CDNs for the delivery of video. Of the 408 qualified responses from companies who spent more than $100K a year on video delivery, not a single customer was using AT&T or Verizon. Of the 307 qualified responses from companies who spent less than $100K per year, only 6 customers were using AT&T and 7 were using Verizon. Here's a link to the slides from my presentation and the video will be up shortly.
The reason for me highlighting this data is not to make fun of any vendor, but rather to use it as an example to many who are still saying that vendors like AT&T are going to hurt Akamai, Limelight or Level 3's CDN business. Even as recent as last week, I saw a report from a Wall Street analyst saying AT&T and other carriers were putting pressure on Akamai's CDN pricing. While Level 3, EdgeCast and Amazon are being aggressive on pricing, none of the major carriers, outside of Level 3, have yet to impact the CDN market in any way.
For more than five years now, Wall Street analysts have implied that AT&T will be a big threat to Akamai. In 2007 I wrote a post saying "Analysts Covering Akamai Should Not Be Worried About AT&T" and in 2008, followed up with two more posts entitled "AT&T's CDN Offering Not Displacing Akamai or Limelight Anytime Soon" and "Goldman Has It Wrong: Akamai Not Affected By Network Operators." Yet even with all the evidence we have in the market to show what's really taking place, and the data to show it takes a few hundred million dollars to really enter the CDN market and do any substantial revenue, many still think carriers are going to somehow disrupt the CDN market. That notion could not be further from the truth and the biggest new disruptor's in the market will come from the likes of Amazon and Dell, not carriers.
As for Verizon, I'm not counting them out just yet. I think they still have a shot at doing something disruptive, but to date, they have been working on these services for two years now and haven't gotten to scale or signed up many customers. I think they could still do something positive, but they have a lot to prove in the market and their window of opportunity won't be open forever. But in reference to AT&T, I do think it's about time people stop saying they will "disrupt" anything having to do with the content delivery business.
Note: I notified Verizon as soon as I realized they had a non-working sales number listed on their website. Also, this post is in reference to Verizon's external digital media and CDN services, not their internal CDN, which is quite advanced.
In February, Boston based Yottaa (pronounced "Yo-ta") announced a new dynamic site acceleration service aimed at the small and medium sized business (SMB) markets with the goal of trying to fill the void left in the market for DSA services, due to Akamai's recent purchase of Cotendo. This week, the company raised $9M in a Series B round and has now raised $13M to date.
For those not familiar with dynamic site acceleration, it is a suite of technologies and products that deals with optimizing dynamically served content across the network. Traditional DSA services often include TCP optimization, route optimization, connection management, on-the-fly compression, SSL offload and pre-fetching technologies.
In addition to offering DSA services, Yottaa also offers mobile content acceleration and content delivery for small objects, not video. The company also announced that they have been granted two patents in the areas of web performance optimization (WPO), Front-end Optimization (FEO), web security and cloud routing. While Yottaa has a long way to go to become the next Cotendo, the company in on track to do under $10M this year, (Cotendo did $24M in 2011) Yottaa did say that in the last month, 5% of the Internet population has visited sites accelerated by the their network.
Yottaa was one of the presenters at last week's Content Delivery Summit and you can see their presentation here that they did with Mocospace, which is one of the largest entertainment destinations on the mobile web with 25 million registered users, spending over 1 million hours per day playing games and making friends.
The company now employees 40 people and current investors General Catalyst Partners, Stata Venture Partners and Cambridge West Ventures combined for the second round of funding, along with additional undisclosed investors.
At last week's Content Delivery Summit, I presented my findings from a recently completed CDN pricing survey, with data collected from over 800 customers that use content delivery networks for the delivery of video. Later this week, I will post all of my slides from the event along with details on how pricing has changed as well as expected growth forecasts for traffic volumes.
I'll be breaking down the pricing based on small customers who spend under $100K-$250K per year, medium sized customers spending $250-500K per year, and large customers who spend more than $1M per year. I also have data from those who spend less than $100K per year and use resellers and regional service providers. Here's a video interview I did with Andy Plesser from Beet.tv talking about what the pricing data showed.
Thanks to everyone who helped make last week's Content Delivery Summit such a success. The rooms were packed, the presentations and discussions on-point and I heard a lot of very positive feedback from the attendees. Lots of great information was shared and various CDN technologies showcased. The summit will be back for it's 5th year in 2013, with the event scheduled to take place on May 20th - so mark your calendar now.
We had attendees from all over the world, carriers and telcos from many different countries and presenters gave out a lot of really detailed info on current video traffic trends, federated CDN models, transit and CDN pricing, DSA and transparent caching technologies and we had a lot of great customer presentations from those deploying these web acceleration technologies. Companies also announced some new services at the event and I'll have a list of all the news from the Content Delivery Summit and the Streaming Media East show in another post.
All of the presentations from the summit can be downloaded from the CDN Summit website and we're working on getting all of the sessions up for video on demand viewing. I'll have a bunch of posts over the next two weeks about things I saw and heard at the CDN Summit that I found interesting including Amazon's new DSA service and details on my CDN pricing data. If there is any kind of post-show help you need or something you'd like me to follow up on with you, please don't hesitate to call or email me at anytime.
When it comes to video on the web, delivering a quality experience is what drives customers and brings them back. At the Content Delivery Summit, taking place Monday, May 14th in NYC, this presentation will discuss some of the ways content owners can maximize their monetization strategy by ensuring a high quality video experience. Learn about the impact of switching from RTMP to HTTP, offering multi-platform adaptive bitrate streaming, multi-device targeting and how to drive a quality user experience by leveraging real-time analytics with measurable performance data. Presenters include:
Readers of my blog can take advantage of special pricing by registering using the promo code of DRF01 to attend the CDN Summit for only $295. Taking a look at the pre-register attendee list we have carriers, telcos, ISPs, MSOs, content owners and vendors coming from all over the world. This is the event to be at if you are in the content delivery business. Twitter hash tag #cdnsummit
With its TV Connect initiative for delivering live programs and On Demand titles via Apple devices, Cox needed a content delivery platform architected to meet its performance, scalability, manageability, and TCO needs. At the Content Delivery Summit, taking place Monday, May 14th in NYC, this presentation will focus on the key attributes Cox looked for in CDN technology to meet their needs for cost, quality and scale. Confirmed speakers include:
Readers of my blog can take advantage of special pricing by registering using the promo code of DRF01 to attend the CDN Summit for only $295. Taking a look at the pre-register attendee list we have carriers, telcos, ISPs, MSOs, content owners and vendors coming from all over the world. This is the event to be at if you are in the content delivery business. Twitter hash tag #cdnsummit
At the Content Delivery Summit, taking place Monday, May 14th in NYC, one of the presentations will cover the Time Warner Network vision and the critical role CDN plays in expanding their video delivery capabilities of Time Warner Cable beyond the set-top-box to the multi-screen, multi-device world. Hear about the deployment and operation of a CDN in a Cable Network from the person responsible for content and video operations and hear why CDN was key to Time Warner Cable's network vision.
It's not too late to get a pass for the event and you can use the special registration code of DRF01 and get access to the keynotes, all of the sessions, presentations and networking events for only $295. Register here and put the code DRF01 in the box at the bottom. Twitter hash tag #cdnsummit
Recent Comments