|
If content really is king, Akamai appears to be his favorite royal carriage. The Massachusetts-based company, which gets its name from a Hawaiian word meaning "intelligent" or "clever," has risen in a little more than a year to become the biggest content distribution service on the Internet. Broadband Week associate editor Karen Brown recently talked to Akamai CEO George Conrades about the forces shaping the world of content delivery.
BBW: Akamai has managed to claim a major share in the relatively young content distribution market. How do you explain the company's success, and what has it done better than rivals?
GC: We have achieved significant success in the content delivery market which is a new marketplace, and I think it begins five years ago with the development of the idea and the technology to enable us to deliver content of all forms-be it graphic or streaming or interactive applications from the edge of the Internet thereby finessing many of the bottlenecks that had heretofore made the Internet infeasible for robust electronic commerce and rich media experiences.
So we got a head start and what we did when we opened for business-commercial operations really in April of last year 1999-we already had successful beta experiences with major content delivery companies and so we were able to open for business confident that not only did we have a bigger idea, but that we had a breakthrough technology that works. And with that, then, we hired aggressively both in engineering and in marketing and sales and customer service and aggressively went after customers, and network deployment, because we had confidence in the technology.
Today we have well over 6,000 servers inside 335 networks in 54 countries. Our objective is to have thousands of servers and ultimately over a thousand networks in all of the countries of the world where there are Internet users. That was our number two goal and we have gone after that-growing our customer base and extending the breadth of our server network confident that we have a breakthrough technology that works. Today, on that technology we are handling billions of hits per day for over 2,000 customers, and we are constantly upgrading the robustness of the technology platform and adding new services capabilities. In fact, over the last quarter we introduced over 150 service upgrades to both make the platform ever more robust and to add new functionality on behalf of customers and networks.
Akamai believes that it's important that content providers have better business models behind their streaming business. For example, in addition to producing compelling streaming content, content providers should monetize their streaming business so that they can measure and evaluate return on investment. Akamai is very concerned about its customers' ability to monetize their streaming media and is developing solutions, products and partnerships that support this.
Akamai recently announced a strategic partnership with Engage to provide products, services and applications intended to greatly enhance the targeting and delivery of online content, starting with advertisements and marketing campaigns to prospects and customers across the Web. The two companies are currently working together to integrate their technologies, and Akamai has plans to announce its ad insertion service in the near future.
With Akamai's recent introduction of its Digital Parcel Service, Akamai is providing solutions to the problem of how businesses manage and leverage their digital assets to their fullest potential. This sets the foundation for a comprehensive digital commerce solution that will address content control, storage, and digital asset management, as well as the enablement of new business models such as syndication and pay-per-view.
Akamai's EdgeScape service allows content providers to determine in real-time the geographic location of the network and type of network from which users are accessing their Web site. This offers content providers real-time network knowledge that can be used to increase the relevance of content served to attract and retain site visitors.
BBW: Recently two groups-the Content Bridge sponsored by Inktomi Corp. and AOL and the Content Alliance led by Cisco Systems-have risen to challenge Akamai in the content distribution game. How do you view these rivals, and what steps must Akamai take to counter these moves?
GC: Well, I do think they have a significant challenge ahead of them, and that's because of the point we were just on. First of all, today we offer a significant value proposition to customers in the marketplace in terms of improved speed two to ten times, 100 percent reliability and secure delivery. And we have over 2,000 customers and we continue to add new services. So again, our customers are the content providers, and we are a technology development company that manifests its value in the form of a service to those customers.
Inktomi and Cisco on the other hand are product providers to individual networks, some of whom would like to develop their own content delivery service capability. But what this means is the individual networks have to bring these piece parts together and fill in the white spaces with their engineering capabilities and develop an effective content delivery capability and then they have to acquire customers. And then they have to deal with the issue that they need to cooperate among multiple networks to reach the subscribers on behalf of content providers.
Akamai does all of this today, after five years of development with outstanding services and thousands of servers across hundreds of networks. So not only do the individual networks have to develop the service capability, they then have to cooperate. And that's where Content Bridge and Content Alliance come in and why they are so concerned about establishing standards, because it isn't intuitively obvious that networks want to readily cooperate with one another.
BBW: What is Akamai's position regarding the use of open standards-based systems in content distribution? Should it be the rule, or is there a role for proprietary systems?
GC: Well, we believe that there is a difference between having a proprietary technology and an open standard. We are all for open standards and I think it is up to everyone to come up with ways to meet the requirements of the standards and yet at the same time excel in their product and service offering.
So we've already demonstrated our interest in standards and our support of standards. We have over 150 industry participants who are working with us in the submission of a standard we call ICAP to the IETF. We've already demonstrated that months ahead of these other two initiatives. But what I think will happen is they will all be joined within the IETF which approves these kinds of things. So we say this is good. Standards are good.
BBW: We've recently witnessed the demise of several streaming media content outlets such as Pseudo.com and DEN. Do you think that is a sign of bad times for streaming media sites you serve? What do you think went wrong?
GC: I can't comment on those two. But I think streaming is just at its beginning, and I think if you look at the history of technological innovation what one sees is a great burst of entrepreneurship and innovation often followed by a great deal of hype, then followed by fierce and growing competition where execution is all important, and then ultimately forms of consolidation.
I think we are just seeing some of the fallout from the early stages of that kind of new cycle. So we are very confident about the viability of streaming audio/visual content over the Internet and constantly working to improve the quality of that experience. And while it will take some time, I believe that eventually we will achieve TV-like quality experience over the Internet. Along the way, there are going to be some winners and there are going to be some losers. But there are a lot of reasons why a particular content provider may or may not do well, including the desirability of the content and/or the execution of their firm.
BBW: How is the rise of broadband influencing content delivery on the Web? Is it proving to be a major influence, or is the market still not big enough?
GC: No, we can see that broadband actually drives the need for optimized services, and that's because the user has an expectation about the quality-the performance-of their experience. And so if a content provider has produced their content in anticipation of a 56K connection and is not using Akamai's service, what you have is content that is not as visually stimulating and may travel over many hops and networks to reach the broadband last mile, where the user for sure won't achieve the full potential of their broadband connection. But with Akamai's servers close to the cable of DSL headend/modem, it's possible to have content that is ready for broadband. We can tell now through our service the speed of connection that the requesting user is on at that moment and through content transformation ensure that the broadband ready content is delivered quickly to the requesting user from the server very close to the cable or DSL headend/modem.
BBW: Is broadband forcing you to improve your networks, then?
GC: Well, sure. In fact it spurs innovation. We have developed a form of streaming called Steady Stream that has demonstrated a superior result in much reduced latency and much less packet loss than traditional methods. And what this means is we can achieve, through both satellite and terrestrial linkages-we've already demonstrated in one event over 5,000 simultaneous 1 megabit-per-second streams.
BBW: A lot of Web fans still are complaining that the quality and consistency of streaming content-particularly video-is inconsistent or just plain poor. What do you find are the most common problems and how do you overcome them?
GC: Well, I think there are a variety of providers out there who use their own techniques-not all of them are equal. We think that we have today the leading service in terms of quality and number of concurrent users, and we have gotten pretty good feedback on the quality of events that we have delivered. The expert buyers choose Akamai when the events are important and that's because of the quality of our streaming. So there are a variety of providers out there. There are various levels of quality and we believe we are chosen for the major events when quality is important.
BBW: Just to turn it around, where do you think the danger spots are-what are the things that are most problematic to the content?
GC: Well, there's no question about the quality of encoding can effect the quality of a video stream. That's one component. But I think that last-mile is extremely important as we discussed and I think the ability of the content providers to monetize their streams is important. For example, to have the capability for pay per view and insertion-the additional features and functions that make streaming viable as a business.
In terms of some of the most common problems in terms of quality and consistency, that's because people are trying to deliver streaming through a few number of entry points to the Internet and in a centralized way. For example, if you ever were to try to achieve an audience of let's say 100,000 concurrent users at 300 kilobits per second, which is TV-like quality, that's 30 gigabits per second, which is greater than any Tier 1 network capacity that I know of and a substantial fraction of the Internet backbone capability today.
And yet 100,000 concurrent users is not much by TV standards. And so I think this centralized approach-trying to force it across the Internet-is a major inhibitor to quality. In contrast, Akamai does two things that stand out-that enable us to have reduced latency and far reduced packet loss. One is that we, unlike others, stream multiple streams to the edge of the same encoded stream, and then combine these at the edge, so that we have a chance to look at let's say three streams of the same content coming from different routes so that we can assure we have at least frame one, then we have frame two, frame three, frame four. And we get a chance to pick from three different sources to time that frame. And that is an Akamai approach that is proving very feasible and important to improving the quality of the stream-and in some cases we have achieved zero packet loss by having multiple streams and combining them at the edge.
The second thing that we do is that we do decentralize or distribute from where we stream. In other words we don't try to force the streams across the Internet but in fact send them to a number of points and then deliver from the edge-picking the server that is closest to the requesting user, hopefully at the broadband headend or modem. By busting up the problem, we can deliver many more concurrent streams regionally in the country and around the world.
BBW: What do you think is the next big step for content distribution?
GC: Continue to move the capability for delivery-more capabilities for delivery to the edge. In other words there are a number of services that have yet to be developed, including dynamic assembly of content at the edge of the Internet, that really makes this model of decentralized delivery of content of all forms-graphics, streaming, interactive applications-feasible. There's plenty of opportunity here to make the Internet ever more robust and reliable for rich media experiences and electronic commerce.
|
