As the strategic value of online content assets continues to grow, chaos and confusion are rampant within the content delivery network (CDN) market. Challenging dynamics coupled with increased demand for content delivery solutions means that a diverse array of companies are keeping a close watch on the space and the relevant players within it.

Yankee Group has developed an Anywhere Scorecard for the CDN market that assesses the relative strengths and weaknesses of the key global providers. The scorecard takes an objective look at how a provider’s Vision and Ability to Transform in the market influence the strength of its offering, and how these dynamics inform customer buying decisions.

Earlier today, I hosted a webinar were I explored the transformative trends affecting the CDN market and ranked 10 of its most significant players. The presentation also featured Yankee Group Senior Vice President Zeus Kerravala, who introduced the new Anywhere Scorecard concept. You can read more about the Anywhere Scorecards here.

The webinar runs about an hour: audio (mp3) and slides (pdf).

The role of dynamic content distributed online is growing at an alarming rate. We know this to be true. As enterprises and content owners push more diverse content to internal business audiences and an expanding crowd of demanding consumers, the importance of this content is magnified. And when people talk about online content, the most common question is how to make the business model work. How does Google make money off of YouTube? How does Facebook turn site traffic into real dollars? What’s the direct correlation between enhanced video communications in the enterprise and greater cost savings or productivity?

But these questions are skipping a crucial step. Before asking how to monetize the content delivered, it’s imperative to know how the content is delivered. This new traffic paradigm is increasing traffic levels exponentially, to a degree that the public IP network is no longer a viable conduit for it all. And this is where a dedicated content delivery network (CDN) must come into play. By providing an external outlet to cache and serve high-volume, high-value content, CDNs can assure greater speed, quality and reliability for all manner of online content distribution.

Yet the CDN market is defined by confusion and change. Declining pricing rates fly in the face of a market that is growing in strategic importance. An over-inflated competitive landscape leaves customers wondering if any given CDN will still be around two months from now. Broader service offerings muddy the waters when asking what features and products are actually necessary. And external pressure on the market from advancing communications service providers (AT&T, Level 3, BT, Tata, etc), infrastructure vendors (Alcatel-Lucent, Cisco) and alternative content delivery providers threaten to cannibalize the existing players. There is a notable lack of clarity, both about who will survive the impending market consolidation and who customers should turn to for their content delivery needs.

Yankee Group Anywhere Scorecard: Content Delivery Networks, 2009

In an effort to address this void, Yankee Group is publishing an Anywhere Scorecard for the CDN space this morning. The Anywhere Scorecard is a new type of report from Yankee Group, whereby we apply an extensive series of objective and subjective measurements to the key players in a given market to offer customers a holistic view of what the players’ strengths and weaknesses are and who to target. This is done by measuring both the companies’ core Vision (strategy, products and services) and their Ability to Transform (core business metrics such as customer base, pricing, channel strength and underlying financials). Going forward, Yankee Group will be publishing Anywhere Scorecards on a variety of topics across the communications landscape, focusing on markets and products that are integral to the emergence and evolution of ubiquitous connectivity.

As for the CDN market, the point of the Anywhere Scorecard is not to rank the 10 included players or provide blanket recommendations on who customers should choose. That is too black-and-white in a market of nuance. Akamai and Limelight Networks may appear at the top, but that does not necessarily mean that their strengths align with the needs of a particular customer, just as players at the lower end of the Scorecard may be best suited for certain circumstances. Customers should use the Scorecard data, the commentary in the report, and the companion report on CDN market trends (called “The Fate of the CDNs”, also published this morning) as a guide book through a convoluted market. Too often, key purchasing decisions of all stripes are made on the basis of incomplete or inaccurate information, and Anywhere Scorecards are designed as tools to combat this.

With an event as transformative and historically significant as the recent electoral unrest in Iran, it is only natural that many disparate elements of life and business are affected. The telecommunications industry is no exception. Yesterday, the Wall Street Journal published a provocative piece examining the control and censorship of the Internet by the Iranian government in an attempt to curtail protests. The crux of the piece is the government’s alleged use of deep packet inspection (DPI) technology acquired from Nokia Siemens Networks in this effort. It is a noble attempt to shed light on censorship and the impediments to free expression in countries such as Iran and China. And it is an unfortunate reminder that in the world of the Internet, true anonymity in personal correspondence or activity is largely a myth. The only problem with the piece is that it is largely inaccurate and misleading.

There are a few qualms that I’d like to get out of the way right off the bat:

• DPI is not what Nokia Siemens Networks actually provides.
• Nokia Siemens provided Iran with Lawful Intercept capabilities designed for voice communications, not DPI technology that is being used for censoring Internet traffic.
• Operators and governments worldwide engage in Lawful Intercept (often by regulatory mandate), which may or may not have DPI as a contributing technology.  In the case of the technology that NSN provides, it is not.
• DPI does not allow for the altering of packet content, as the article suggests, to create disinformation campaigns.
• DPI does not have to be inserted directly into the data flow if it is just engaged in monitoring activities, and even when it is, it does not create a slowdown in network traffic that would be perceptible to a large public audience.

The list could go on.

The main point I would like to make though is that DPI is mischaracterized as a “practice” or an “activity”. DPI is a technology. What the article is describing is one potential, particularly malicious, usage of the technology. Yet many things that are potentially benign can also be potentially dangerous. Just because someone can use binoculars to invade another’s privacy does not mean that they cannot also be used for bird watching. To take a more extreme analogy, just because you can get behind the wheel of a car and run someone over does not mean that cars are inherently violent. The onus is on the user of the technology, rather than on the technology itself, and this is what the article misses. The culprits are not Nokia Siemens (leaving aside that NSN does not actually provide DPI technology) or their peers. The culprits are those that would use the technology for malicious purposes. In this case, the Iranian government.

What DPI does is what it says it does: packet inspection. It is a technology used to gain Layer 3 through Layer 7 packet visibility (from the network layer through the application layer) to determine the source, destination, application type, etc of network traffic. It does not read emails. It does not alter Internet content. It does not slow down the Internet. And, directly to the point about Iran, it does not intercept or block traffic. The Iranian government can choose to take action in these regards based on the intelligence about traffic flows that DPI can provide, but again, that is independent of the technology.

This is actually a more extreme (and more politically charged) example of what got Comcast into hot water last year with how it chose to use network information. For those unfamiliar with the case, Comcast raised the ire of public interest groups and the FCC for blocking BitTorrent traffic that it deemed to be overly burdensome on its network. The FCC deemed this inappropriate not because of how Comcast obtained traffic information (using DPI), but because of what it chose to do with that information. In a similar fashion, a number of operators in the US were chastized by the House of Representatives Subcommittee on Telecommunications and the Internet for using DPI-acquired intelligence to do behavioral-based ad-targeting. This ultimately led to behavioral targeting vendor NebuAd getting dragged before Congress for a tongue-lashing, and ultimately folding in the face of public and legislative pressure. Again though, the issue was how the technology was used by NebuAd and its operator customers, not the technology itself.

The technology itself has a number of legitimate uses that are in line with the public good (and operator profits, for that matter), including security threat detection, threat mitigation, enhanced network management, enhanced quality of user experience, the ability to introduce new services, etc. Yankee Group has written a number of pieces on this issue in the past that examine how operators are using the technology today and what the potential opportunites are in the future. These often go unmentioned though when DPI is reported on, because they don’t arouse public debate the way that “Iran’s Web Spying Aided by Western Technology” does.

My goal here is not just to argue with the Wall Street Journal though (an argument I’m sure I would lose) or point out inaccuracies in the article. It is to underscore the consequences of misrepresenting something like this. In the past two years, DPI providers have run afoul of issues around privacy, net neutrality and now censorship, due to the ways in which customers have chosen to use the technology. These issues have attached a scarlet letter to a technology and companies that can provide legitimate value to operators and consumers, when used properly. Instead though, competitors have been forced to retreat from the market, the maturation of the technology has stalled, and operators have turned towards potentially less efficient solutions for network visibility, security and traffic management for fear of igniting a firestorm amongst those that would misconstrue their intentions.

It is a fine line to walk in regards to what is and is not acceptable in the world of traffic inspection, to be sure. But that’s all the more reason why the technology must be accurately understood and represented, rather than demonized off-hand.

Last month, I wrote a piece discussing the growing trend (and growing importance) of carriers entering and reshaping the CDN market. Well, here’s an update: The march continues, unabated.

Just within the last month, the following things have happened:
•    Global Crossing officially expanded its CDN reseller program to include both EdgeCast and Limelight Networks.

•    BT indicated that it will enter the CDN market by the end of 2009 with an internally developed offering.

•    TeliaSonera is expected to announce next month that it will enter the market in some fashion.

Full credit for both the BT and TeliaSonera news goes to Dan Rayburn at streamingmedia.com.

I cover in the previous piece what has happened thus far and how the carrier offerings will differ, both from each other and from pure-play CDNs. What I want to make note of here is the velocity at which this is happening. This makes four major global carriers (Deutsche Telekom, BT, TeliaSonera and Global Crossing) that have thrown their hat in the ring since the beginning of the year, in addition to further announcements from Verizon and AT&T that relate to the CDN space. Yankee Group has spoken to multiple other carriers that have yet to announce anything publicly, but are likely to produce some manner of CDN offering (most likely resale or whitelabel at first) within the next six months.

Also of note is where this activity is occurring. There was a time at which many international operators dismissed this trend as a concern primarily of their US colleagues. The argument was that CDN was of most strategic importance in the US, as that is where the majority of high-demand online video content is created. Therefore, it would be the US carriers that bear the greatest burden and have the most to potentially gain (in terms of cost and bandwidth reduction, as well as potential revenue) from a CDN offering. This supposition was buoyed by the fact that Level 3 and AT&T were the two notable first-movers.

Yet we know this argument to be false. We’ve now seen major incumbents in EMEA and APAC join the fray as well, and there are ongoing discussions from multiple carriers in Latin America and other parts of APAC about how they might sell CDN services. The reason? Even if high-volume content is still often produced in the US (and that is becoming decreasingly true, as a matter of fact), it is consumed globally. Therefore, any carriers with significant long haul assets charged with carrying a large amount of global traffic are gravitating towards this market.

The following graph shows the top 13 providers worldwide in terms of total IP traffic carried in 2008. It reads like a who’s who of providers that have entered the CDN market in the past two years. Level 3? #1. Global Crossing? #3. NTT? #5. AT&T? #7. TeliaSonera? #8. And so on.

Credit for this goes to Earl Zmijewski at Renesys (the original article that accompanies these rankings is fascinating as well, and can be found here).

This is certainly not an exhaustive list, but the salient point is that there is both rhyme and reason to why this is happening, and it’s not just about carriers mimicking the actions of their peers or seeing a potential revenue stream. Those providers with the greatest amount of responsibility for IP transit are looking at the changing nature of content delivery and asking how they can streamline their own processes. Augmenting a large IP network with a dedicated CDN to cache and serve high-demand content locally has significant benefits in terms of capex savings, network efficiency and (some) revenue generation. Take a look at who are the other operators with large IP backbones worldwide, and you’ll get a pretty good sense of who might be kicking the tires on the CDN market in 2009. Or at least who should be.

Following yesterday’s announcement by AT&T that they are cutting 2009 capex, we have seen a new round of stories and queries popping up wondering whether the sky is falling for the infrastructure vendor community. Particularly for mid-tier vendors such as Tellabs and Ciena, a significant cut in US capex could spell real trouble, this much is true. But how significant a cut are we talking about?  And how should the vendors react to it? Glad you asked! Let’s examine it, shall we.

AT&T announced a capex cut of 10-15% off of 2008 levels. 2008 capex was $20.3 billion, meaning that we are talking about a decrease of between two and three billion dollars. To be sure, that’s not chump change. But what’s driving the level of the cut? The answer lies in AT&T’s revenue outlook. 2008 revenue was $124 billion, which was a gain of roughly 4% on 2007. For 2009, the carrier is expecting “continued consolidated revenue growth in the low single-digit range”. Reading between the lines, this means that they’re still going to see growth, but it’s likely to be below that already-modest 4%. I’m going to take the liberty of saying that AT&T is looking at a revenue guidance of roughly $127 billion. Relatively modest growth to be sure, and it may be even overly-optimistic in light of a highly uncertain economic climate. AT&T knows this is a possibility, and therefore caution must be the order of the day.

Yankee Group has said all along that the key metric the vendor community should pay heed to is capex as a percentage of revenue. This ratio is what we believe to be informing new capital investments, and we expect it to remain fluctuating in the mid-teens, even in this difficult climate. In 2008, AT&T’s capex as a percentage of revenue was 16.4%. In 2009, we expect that to drop to 13.5-14.5%. A significant drop, but one that is absolutely understandable in today’s economy. And it is not as if capex is falling off the cliff. We would be worried if this ratio dropped into the single digits, but we are still a long way from that. We expect Verizon will experience a similar decline, with capex comprising roughly 16% of revenue (off of 17.7% in 2008). Significant, but again, not cause for vendors to start Googling “Chapter 11″.

So what does this really mean going forward? It obviously means a drop in overall US capex in 2009, as AT&T and Verizon comprise roughly two-thirds of US telco capital spend. Myself and Brian Partridge will be doing a 2009 capex overview in the coming months as a follow-on to our 2008 projections (please remember, these projections were made before the bottom fell out of the economy). In the immediate term though, what can the vendors expect? Expect existing capital intensive projects to remain fully funded. Namely, U-verse and FiOS ain’t cheap, and the planned roll outs for them aren’t going anywhere (the two carriers have said as much). And long-term plans that are central to core strategy, such as Verizon’s migration to LTE, are still on track.

What’s going to be hard is getting any new dollars out of the carriers. If you don’t have pledged budget for equipment sold to these carriers already, you are facing an uphill climb. Any new initiatives that are being pitched to the carriers must be done so with a quick return on investment in mind. Five year ROI models won’t fly. Remember, the goal has to be normalizing that ratio of capex as a percentage of revenue. So if a carrier is investing in 2009, they’ll want to see a revenue benefit by the end of 2010 at the latest. The long-term money is already pledged, so vendors with new products are fighting it out for short-term investments, and short-term investments require a short-term return. The CTO can love your value proposition with all of his or her heart, but if the CFO’s eyes don’t turn green upon seeing it, don’t count on squeezing a lot of new money out of these carriers. Or at least until the economy finds the bottom.

I spent the past few days at a conference for Procera Networks, a provider of deep packet inspection-based solutions for network monitoring and service control. There was talk of the direction of the DPI market, the overall direction of service provider investment in the current economic climate, and how network intelligence can work in conjunction with other distinct network elements to improve efficiency. But the biggest question on everyone’s minds (mine included) was when will operators finally wake up to the need to change their subscription service models?

For years now, there has been discussion from the vendor, service provider and analyst community about movement towards tiered services and premium priority-based charging. It’s a tried and true story. Operators must avoid relegation to the dreaded DUMB PIPE OF DOOM (apologies for the emphasis, but this expression has become so over-saturated that it has taken on the life of a catch phrase). The opposite of a dumb pipe is a “smart pipe”. Ergo, operators must invest in technologies that provide greater network intelligence, and this will allow them to provide differentiated service to subscribers based on subscriber priority, application, time of day, location, etc, as well as implement convergent charging models that incorporate both pre-paid and post-paid. Easy, right?

Actually, no, it hasn’t been. Operators have been strikingly slow to implement priority-based or tiered subscriber plans. This despite the fact that in just about all discussions with the operators, they say this is what they ultimately want. And Yankee Group’s survey data bolsters this as well. Operators know that flat-rate business models aren’t sustainable, and they want to move beyond them (our resident network software systems guru, Ari Banerjee, just published a report detailing these operator preferences and how to address them).

But when you search out models that are already out there, they are few and far between. The one that always jumps to mind belongs to PlusNet in the UK (now part of BT). You can view some of the details here, but this actually doesn’t show all of the depth in the tiers that PlusNet has created. Subscriber plans are tiered by price, bandwidth, particular application allowance and expected quality of experience (QoE). And PlusNet has been remarkably successful with this model since its introduction. They’ve received positive customer feedback for it. They’ve won industry awards for it. Most importantly, they’ve made money with it. But other than a few limited deployments and endless trials and rumors, we’ve seen few operators introduce something similar. And this was something that was first constructed four years ago! What is beyond the bleeding edge? The hemorrhaging edge?

So do we see this changing? Ultimately, yes. Operators will eventually move away from flat-rate and embrace more personalized service models, but due to necessity rather than foresight. There is a tipping point at which bandwidth constraints and the economic burden of over-provisioning the network will reach a level where operators will be compelled to act (where exactly it is, no one knows). We are already seeing the beginning stages of this as operators such as AT&T, Comcast, Time Warner and T-Mobile play around with bandwidth caps. There is talk that mobile operators are being more progressive in constructing service plans than their fixed-line brethren, and that this community of operators will evolve before they absolutely have to. But again, this is mostly talk with little concrete evidence of commercial deployments to date. So we are left waiting for the point at which a good idea becomes a business imperative. And when we ultimately cross this threshold, operators will be forced to evolve and they’ll be pushed into a more profitable and efficient business model, kicking and screaming all the way.

On this Yankee Group Podcast, David Vorhaus, Vince Vittore, Josh Martin and Jeffrey Breen discuss all manner of free content and services, ranging from freemium services, advertsing-supported business models, free muni WiFi and more.

Free Free Free (mp3 / 6.57 MB / 07:10)

On this Yankee Group Podcast, David Vorhaus, Josh Martin, Vince Vittore and Jeffrey Breen discuss the changing role of piracy in the consumption of digital content, and how operators and content providers can monetize their assets without causing a consumer mutiny.

Piracy…Part 1 (mp3 / 3.58 MB / 03:54)

The role of net neutrality in shaping the telecommunications market continues to grow, both in the US and abroad, and the FCC ruling on Comcast’s network management practices has only amplified its importance. In this Yankee Group podcast, David Vorhaus, Vince Vittore and Josh Martin discuss the issue and how operators are responding.

Net Neutrality podcast (mp3 / 11.3 MB / 12:22)

I’m currently drafting a piece of research on the role of network operators in providing content delivery network (CDN) services. Traditionally, CDN has been the purview of third-party players such as Akamai or Limelight, who cache internal enterprise or consumer-directed content at local servers to avoid the latency, speed and quality issues that can arise with content traversing the network backbone. These players in turn lease long-haul capacity from carriers, and these leased capacity costs typically comprise roughly 30% of a CDN’s cost structure.

Increasing, operators are looking at the CDN market and asking why they can’t do this themselves. In an effort to control bandwidth on their own networks, operators such as AT&T, Verizon, Level 3, Tata, Reliance and others have already made movements towards this market, with countless others kicking the tires on CDN in one form or another. It’s mostly a cost and capacity saving proposition for these guys, though there are certainly a few bucks to be made here as well.

The reason I bring it up today is that there was a notable announcement by CDN-provider Velocix to address this trend. Velocix finally debuted its “Metro” offering, which has been known about in some circles (and highly anticipated) for a while but not publicly announced. Essentially, it’s a direct pitch to the ISP community, offering up Velocix as either a provider of CDN services to them or as a partner to use in offering the ISP’s own CDN offering. It’s a fascinating play, though not an entirely unique one, as other CDNs are looking at ways in which they can develop a reseller or white-label model to position themselves as a partner to the encroaching service providers…as opposed to a competitor that is about to see their market turned upside down. I just spoke to a Malaysian-based CDN that is getting ready to launch in January with a business model entirely predicated on attracting mid-tier ISPs.

There is one particular aspect of the Velocix offering that caught my eye though, and that is the different way it is manifesting itself in the US and in the UK. In the US, Velocix has already announced a notable partnership with Verizon, whereby Verizon will resell Velocix’s CDN service to content providers. Essentially, the pitch is, “Dear content provider X. We are using this CDN to handle our own FiOS content. We would be happy to handle your content in this preferred fashion as well…for a nominal fee, of course”. So from Velocix’s point of view, the ISP is the carrot to attract content provider customers.

In the UK, the situation is inverted. Velocix has a strategic relationship with the BBC, whose iPlayer P2P service represents a huge glut of traffic for the operators. Velocix is going out to ISPs like The Carphone Warehouse and Virgin Media with its Metro offering, and the pitch is, “Dear ISP X. We have a relationship with the BBC whereby we can locally cache iPlayer content. We would be happy to do this specifically for your customers that are consuming this BBC content and weighing down your network…for a nominal fee, of course”. So from Velocix’s point of view, the content provider is the carrot to attract ISP customers.

The reasons for this dichotomy have to do with the relative power of ISPs in the US versus the UK due to local-loop unbundling, as well as the sheer volume of traffic that BBC’s iPlayer generates. But the model of some sort of three-pronged relationship between content provider, network operator and CDN is one that is consistent across the regions, and a viable direction for the overall content delivery market moving forward.

Now the question is what does this mean when measured against the specter of net neutrality? Is it legal for a service provider to ask a content provider to pay extra for delivery, and then give that content preferred status over those that won’t pay-to-play? I’ll leave that can of worms for another time…