Here's How Akamai Wins

Bravo for BitGravity! On Monday, the would-be Akamai (Nasdaq: AKAM  ) killer announced an expansion of its existing partnership with Tata Communications (NYSE: TCL  ) . Tata also pledged $11.5 million in additional funding for the upstart.

Is this how Akamai dies?
It's news like this that seems to be spooking Akamai investors. They're wondering whether this Rule Breakers recommendation is permanently broken. (Shares of Akamai are down more than 45% year to date.)

I don't think so. Here's why:

Over the last couple of years we saw a tremendous amount of growth related to broadband adoption. This year we've seen that moderate a little. I think we'll see a second wave as we go from [1 terabit per second] to 500 [terabits per second]. We think in order to scale, you can't do that from a centralized architecture, you have to do that from the edge where the capacity is the most abundant and the least expensive. I think what we'll see is it will be a little bit slower to come but bigger than most people expect.

Those are comments from Akamai Chief Financial Officer J.D. Sherman at a recent investor conference sponsored by Citi. Notice the language -- Sherman is conceding that Akamai's growth story faces shorter-term challenges.

What I find interesting, and comforting as an investor, is his take on competitive advantage. Translated, he's saying that Akamai is best positioned to handle the explosive growth of Web content.

A line in the Web
Is that really a fair claim? I can't be 100% sure. I've long been an Akamai bull because of its algorithm. But that's only one part of the story, according to company Senior Vice President of Networks and Operations Bobby Blumofe, whom I recently interviewed.

Think of a line, he says. On the left side of the line is the Internet, and on the right side are users. Akamai's servers operate on the right side. BitGravity, Internap (Nasdaq: INAP  ) , Limelight Networks, and others operate on the left side. So long as the bridge -- the interconnection -- between the Web and users is reasonably fast, this isn't too large of a problem to overcome. Live video can be delivered straight to users a la BitGravity, for example.

The trouble is that the interconnection has limits. As Blumofe explains:

If the servers are on the wrong side of the link, then every viewer causes an additional copy to be streamed over that link. For example, consider a 1 [megabit-per-second] stream, maybe TV quality, with an ISP having 10,000 viewers. In that case, that ISP is going to have 10 [gigabits] of traffic going over their uplink. That's a lot of load, and they likely don't have that much spare capacity on that link.

Thus, according to Akamai, the line between the Web and ISPs such as AT&T (NYSE: T  ) and Verizon (NYSE: VZ  ) becomes a brick wall. Crossing at full speed is impossible, and what should have been a live feed dies.

Contrast that with how Akamai works. By modeling itself in the same decentralized style as the Internet itself, and by using an algorithm that allows data to be broken into bite-size chunks and then reassembled very close to its destination, Akamai doesn't need the delivery system -- the interconnect between users and the Internet -- to scale. Rather, Akamai's algorithm and architecture scale the Web.

Or at least that's the theory. It could take years for Akamai to be proved right, if ever. BitGravity points to Akamai's storage problem. And Google (Nasdaq: GOOG  ) and Microsoft (Nasdaq: MSFT  ) are among those spending billions to beef up the connections between the cloudy confines of the Web and the ISPs that deliver data.

But even here Blumofe is skeptical. He asserts that unless these connections become the digital equivalent of a 100-lane multi-terabit highway -- versus the five-lane gigabit highway we have today -- no CDN positioned on the wrong side of the line will scale to the degree that Akamai can. Patient investors will thus be rewarded.

I think he's right. I've bet my hard-earned dollars on it. Do you agree? Disagree? Use the comments box below to express your point of view.

Get your clicks with related Foolishness:

Fool contributor Tim Beyers had positions in Akamai's and Google's shares and Google's LEAP options at the time of publication. He also hunts for the best of tech as a contributor to Motley Fool Rule Breakers, which counts Akamai and Google among its holdings. Here's how to try this market-beating service free for 30 days. Get access to all of Tim's Foolish writings.

Microsoft is an Inside Value pick. The Motley Fool's disclosure policy is brick-solid when it comes to being timely.


Read/Post Comments (9) | Recommend This Article (7)

Comments from our Foolish Readers

Help us keep this a respectfully Foolish area! This is a place for our readers to discuss, debate, and learn more about the Foolish investing topic you read about above. Help us keep it clean and safe. If you believe a comment is abusive or otherwise violates our Fool's Rules, please report it via the Report this Comment Report this Comment icon found on every comment.

  • Report this Comment On September 09, 2008, at 5:09 PM, lightman66 wrote:

    Your articles should address the fact that Akamai management pays itself too handsomely. The investment community does NOT want to buy a stock where management is self-serving. If management would concentrate on shareholder value instead of making themselves mega-millionaires then the stock would be more likely to appreciate instead of steadily declining. A little more than a year ago Akamai hit $59 share. Granted the entire stock market itself has been visiting the toilet thanks to our policy decision makers but Akamai needs to immediately change its philosophy if it desires to exist in the future. Management needs to concentrate on profits and expending the services the company offers instead of making themselves richer. This company has an all too common corporate disease.

  • Report this Comment On September 09, 2008, at 5:21 PM, rg807 wrote:

    Akamai runs a lean mean machine and doesn't come close to overpaying. I know, I used to work there.

    The company, which didn't miss it's forecast is getting clobbered unfairly and it is coming at a time when the market overall is getting crushed.

  • Report this Comment On September 09, 2008, at 11:24 PM, foolfanuno wrote:

    I disagree with the analysis. Watch the fastcompany.tv interview of BitGravity CEO Perry Wu (http://tinyurl.com/6opuhw) and CTO Barrett Lyon (http://tinyurl.com/6s5y3b) and tell me what you think. I would like your opinion. I think they are possibly closer to the user than Akamai. As they grow with this partnership with TATA, they will be that close to more users, without massive server farms, shareholders or tremendous VC to answer to.

  • Report this Comment On September 10, 2008, at 8:08 AM, TMFMileHigh wrote:

    Hello foolfanuno. Neat video and interesting interview with CTO Lyon. Excellent ideas; not closer to the user than Akamai. BitGravity may, however, be comparable in instances where Tata is the ISP for the user in question. The problem of the Internet is that the ISP is always an intermediary. The "line" is the interconnection between the ISP and the Web itself.

    Remember: Akamai doesn't suck at video. It never has. MySpace and ITunes are big customers and the network has been tuned for HD: http://www.akamai.com/html/about/press/releases/2007/press_0...

    I really think BtGravity is interesting; the point-of-presence storage arrays, in particular. I continue to like AKAM because (1) it is constantly upgrading its network and (2) for as good a partner as Tata is IBM is at least as good. AKAM and IBM have been partners since its earliest days. Please don't assume that BitGravity or anyone else has a hammer lock on advanced hardware and storage techniques.

    None of this is to disparage BitGravity. Among the upstarts, it is my favorite. Bluster with backing, if you will. I just don't think it's going to kill Akamai.

  • Report this Comment On September 10, 2008, at 10:34 AM, rg807 wrote:

    foolfanuno - Bitgravitys revenues are less then $5M per year and the Tata deal is probably a lot closer to a buyout then a partnership.

    More importantly, most internet media is consumed in Japan, Korea, parts of Europe, and overwhelmingly in the US. There is little to no consumption outside of these areas so Tata's value is questionable or at best probably not a threat to any of Akamai's existing revenue.

    Finally, lost in this perhaps is that ISP's like Akamai. Akamai cuts their upstream costs significantly so I fail to see a big advantage for BitGravity there.

  • Report this Comment On September 10, 2008, at 11:59 AM, jb77068 wrote:

    Tim - you try to mix the products to make it seem like secret sauce that AKAM is doing - this is not the case. They stream the same way the other vendors do, their is no "WAA" in the streaming aspect.

    "WAA" is the accelerator product you are trying to describe mixed with the streaming product.... 2 different solutions.

    It is a crazy market in the CDN arena - there are way to many players & pricing is extremely competitive. The WAA product is unique to AKAM, Stream OS (CMS) - unique, but all the other aspects of the business are within most other players hands. They do have their stuff together, but have to get past the pride to really become a viable discussion again.

    I do disagree with the Tata evaluation above - this person needs to realize with the world getting flat - most of these people are now getting computers and starting to enjoy other aspects of life that we have had in the US for years (there is a middle class to consume the videos & ever hear of "Bollywood?)".

  • Report this Comment On September 10, 2008, at 2:02 PM, bkinn wrote:

    You say: ""The trouble is that the interconnection has limits. As Blumofe explains:

    If the servers are on the wrong side of the link, then every viewer causes an additional copy to be streamed over that link. For example, consider a 1 [megabit-per-second] stream, maybe TV quality, with an ISP having 10,000 viewers. In that case, that ISP is going to have 10 [gigabits] of traffic going over their uplink. That's a lot of load, and they likely don't have that much spare capacity on that link.

    Thus, according to Akamai, the line between the Web and ISPs such as AT&T (NYSE: T) and Verizon (NYSE: VZ) becomes a brick wall. Crossing at full speed is impossible, and what should have been a live feed dies.""

    Are you serious? I can understand you talking about a local ISP etc... but using VZ and T as your example of a carrier that won't have enough uplink capacity looks pretty silly ... doesn't it? For crying out loud .. many ISP's uplink to VZ in our area. VZ has more edge capacity for local uplinks in the areas they will offer internet service via FiOS than local MSO's (since most uplink via VZ truncs), and though cable co's like comcast seem to be well equipped ... eventually they will need to upgrade if they want to provide the plethora of BW offered by carrier turned ISP's like T and VZ.

    Yes broad band use has been growing, though not at the levels Akamai would have everyone believe. The bottom line however is that core and edge infrastructure has been upgrading as well (and you can check with others who work in telecom). Core AND edge capacity is and has been building to meet the needs of an architecture that will eventually erode the need for Akamai's "decentralized content delivery model.

    Given the sweet deals that management recieves, it appears to me that they may be aware of these challenges are are attempting to maximize their "time in the sun". In my opinion, Akamai's model will fill needs in today's market but will eventually begin to lose ground to the trend that has been identified by even carriers such as AT&T, Level 3, etc. These carriers will not only take share from Akamai due to next generation changes, but will actually out price them because they own the networks.

  • Report this Comment On September 11, 2008, at 7:24 AM, TMFMileHigh wrote:

    Hello bkinn,

    >>using VZ and T as your example of a carrier that won't have enough uplink capacity looks pretty silly ... doesn't it?

    I wouldn't have written it if it did. The math problem still applies. 1 mbit/sec is an extremely small and low-quality stream. Once you start talking about HD and more concurrent users, you're well past 1 tbit/sec. There is no common infrastructure available that pushes that much data that fast. Even BitGravity knows this; it's why they're building a lot of their own hardware.

    Foolish best, Tim

  • Report this Comment On September 18, 2008, at 10:49 PM, Internettech wrote:

    Sorry Tim,

    With all due respect, you mean well. But it's contradicted by the evidence. Like I said in previous posts Akamai's architecture was designed for dialup users and the patented techniques of splitting up the html webpage code from the embedded graphic objects and CSS style sheets is not required anymore. Microsoft and NBC and ATT and others have realized this. It's just better and faster to send the whole webpage from multiple points on the Internet with far less problems with nested DNS, overloaded edge servers, and missed cache hits. This is what the new CDNs are doing and what Microsoft, Amazon, and Google are doing.

    So the king is naked and isn't wearing any clothes. The king should buy some new clothes with a new design instead of trying to convince us that hardware router speed doesn't have an impact and lower the need for splitting up the html code and the graphic elements in the Akamai model.

    Carrier CapEX is real for a reason. If math software could do it all then Cisco wouldn't be in business. Here's the real math:

    The actual maximum capacity of Cisco's top of the line router in use at ATT and Verizon's ISP backbones is the Cisco CRS-1

    See:

    http://www.cisco.com/en/US/products/ps5763/index.html

    Notice in the first paragraph where it says "it is designed for always-on operation while scaling system capacity up to 92 Tbps."

    Translated that means that the router will route up to 92 Terabits per sec which equals 92,000 gigabits or 92, 000,000 megabits.

    Which is well past the 1 terabit/sec you mention.

    Also don't overlook when ISPs stream they can use a hardware based multicasting to stream a HD file so 10 megabit stream will support 10,000 viewer-users at the edge or premise and a 10 gig connection would support 1000 times that. The edge router creates the last mile duplicate copies and the core routers only send the original stream. Since some streams (like youtube videos) only have one user this isn't really 1000 times that, but you get the idea.

    100 gigabit network interface connections will be in ISP router networks late this year from Cisco, Juniper, Foundry and others. The 1 terabit/sec=1tbit/sec number you refer to is an aggregate number equal to 10 100 gigabit network interfaces. The bigger routers can hold 100's of network interfaces.

    I have real problem with the left/right analogy; it not even close to that simple. That analogy just doesn't work. ISP's manage networks and connect their networks to an upstream peer like ATT, Level 3 or Verizon or connect at peering points with other equal sized ISPs. As the new version of IP (IP version 6) becomes more available here in the US, the ISP connectivity and cyber defense issues will become more complex and will impact Akamai's older edge model.

    I say that Akamai is not even a billion market cap company based on revenue and net, so it still has some room to fall to be priced right. I figure 6 or 7 dollars a share. They really need to do what Microsoft did when they made an about face in 1995 and focused on the Internet. Akamai needs to get away from the original math method/split html/objects from the dialup days and start working on something new.

    I don't think they will and one day they will miss their earnings and the stock will fall to 2. I said it was over priced at 32 and I say it's still very over priced.

    And please get them to fix their DNS servers so that their inverse addresses work properly. This is a real issue in proper national CYBERDEFENSE and they need to fix it ASAP.

Add your comment.

DocumentId: 723774, ~/Articles/ArticleHandler.aspx, 4/24/2014 11:37:37 AM

Report This Comment

Use this area to report a comment that you believe is in violation of the community guidelines. Our team will review the entry and take any appropriate action.

Sending report...


Advertisement