It seems like everyone in the tech industry has an opinion on the hostile takeover of Yahoo by our favorite enemy, Microsoft. I certainly am no exception. On the other hand, I am almost as sick of this discussion as I am of the American presidential primaries. A new wrinkle appeared this week though that I think bears some exposition.
Yahoo is in trouble and, despite a PR engine running full steam, they can’t really hide this fact. Despite owning some of my favorite web darlings, like del.icio.us and Flickr, Yahoo has failed to successfully monetize their assets in a way that can fund the whole organization. Suddenly, out of nowhere, Microsoft offers to buy up Yahoo.
Microsoft is in a minor spot of trouble itself, with its new version of Windows, Vista, being widely regarded as a failure. The market share of their main operating system contender, Apple, continues to grow at an alarming rate. Microsoft needs a reliable revenue stream to help offset some of these losses. Yahoo would provide value both in terms of well-known web properties and lots of web advertising dollars.
At this point, Yahoo fired up the internet “bat signal”. “For the love of god, someone rescue us!” Their Batman did not appear, at least not immediately. Google already has competing products for most of Yahoo’s offerings. Murdoch doesn’t have the immediate capital to ante up on Microsoft’s offer. Sun couldn’t care less about this space. Then, out of the dark, a figure appeared. It wasn’t Batman; it was Time Warner. An old conspiracy theory came back to life with real weight behind it.
What about an AOL and Yahoo merger?
To anyone in the industry, this immediately sounds like the stupidest idea possible. AOL is failing. Yahoo is failing. Putting the two companies together sounds like something a venture capitalist in 1998 would have thought up. Underneath the failures, though, are some real benefits, as noted by Silicon Alley Insider. I specifically want to focus on one aspect of Insider’s next-to-last item on their list of benefits: being able to cut operating costs.
(Full disclosure: I worked for AOL’s Network Operations division from Feb 2003 until my layoff in Oct 2007. Due to some annoying paperwork, I can’t let you know my thoughts on sections of AOL quite yet, at least not here. Buy me a drink sometime and I’ll rant for hours. Well, I do that anyway. I just want free booze.)
The Network Is The Value
Every major internet corporation has their own datacenter(s) and runs their own servers. They have hordes of network and systems staff keeping their often delicate software and products running. Both companies are well stocked with operational teams. Yahoo doesn’t need more IT people. To make matter worse, operations teams do things their own way and can be pretty territorial. Trying to force major operations teams together can be a major fight. I’ve seen this drama carried out from the level of 5 man teams merging to teams of hundreds of people merging. It is never pretty.
In the case of a full merger of AOL and Yahoo’s operational teams, many people will probably have to go. Expect a big layoff in which hundreds or more are lost. It will be people from both sides of the merger. Suddenly, there will be two divisions dedicated to network monitoring, two infrastructure teams, two hr structures. To quote Mad Max: Beyond Thunderdome, “two men enter; one man leaves.”
The Silicon Alley Insider article notes this as one way to cut costs for the merged Yahoo/AOL entity. It certainly is an excellent opportunity to realign resources and chuck needless divisions and potentially even datacenters. I think the value proposition goes deeper though. Underneath all the drama and swirl around AOL lies a relatively unknown entity called ATDN, the AOL Transit Data Network. This network is a Tier 1 provider with pipes spanning the globe. This is the real value, in my opinion, behind the AOL/Yahoo merger proposition.
A Lesson In Internet Topology
Almost every individual consumer of the Internet pays for access to the internet. Some of us pay a flat rate for “unlimited” data amounts. (Yes, Comcast, you are the reason that’s in quotes.) Others pay by the megabyte, usually over a certain cap. Some people are lucky enough to have their network access provided for free.
This concept carries over into the large globe-spanning networks as well. Most major corporations like Microsoft or Google pay a lot of money to have their data carried around the world. Every time you go to google.com or microsoft.com or yahoo.com, you cost them money. Individually, your request costs them very little. With products on the scale of Google’s search engine or Yahoo’s Flickr, though, those costs begin to stack up.
Conducting a search on Google or downloading a patch from Microsoft is a bit like shipping packages. You have a bundle that needs to go from point A to point B, preferably as fast and cheap as possible. If the destination is local, you can carry the package yourself effectively for free. If you want to send that package across the country, however, you need to pay a carrier like FedEx or DHL a fee. The company takes your package, carries it across the country, and delivers it to the intended recipient.
With Internet transit, this is a lot more complicated. FedEx or DHL can reach most of the world on their own and don’t need the services of other providers to ship packages. This is not the case on the Internet. No single provider can reach everyone on the Internet on their own. The Internet is actually comprised of lots of network providers who talk to each other either directly or through some other provider. Networks that want to talk to each other have two options. They can pay each other for the traffic or they can be “peers”.
Peers are physically interconnected networks which share routing information. These connections can be free or cost money. Generally, paid peerings cost significantly less than normal transit does. This is why peering is so attractive. (As Wikipedia points out, common usage differs from the technical meaning.)
As a network grows, it gains more and more opportunities to peer with other networks. It gains more cheap access to the world, particularly as it gains smaller customers like local colocation facilities or smaller broadband providers. Eventually, the provider is ready to join a group known as the Tier 1 networks. Every Tier 1 provider peers with all the other Tier 1 providers. They get very significant discounts on any paid peerings and their operations teams work very closely together, particularly during Internet-wide drama like network attacks or incidents like the accidental shutdown of YouTube by Pakistan’s network staff.
The highest and most prestigious class of network is the Settlement-Free Tier 1 (also known as an SFP Tier 1, for Settlement Free Peering). They do not pay for connectivity or access to anyone else’s network. Their traffic goes around the globe, from your desktop to your favorite web site or email server for free. Obviously, this is the best position that any network can be in.
ATDN: Strategies and Value
I can’t tell you a lot of what’s going on behind the walls of ATDN. I can’t tell you if ATDN is a settlement free peer or not. I can’t tell you what discounts they provide or what their transit costs are. What I can say is that ATDN is a Tier 1 network that spans the globe. Their network has termination points in most major cities in the continental US and in many major cities in Europe. A host on ATDN in Southern California can send and receive data to a host in Germany effectively for free. A peer of ATDN can thus access these areas at very little cost, compared to other providers.
For some reason, Time Warner is willing to give up ATDN. They’re willing to give it up to Yahoo, of all companies. Time Warner’s side of this venture is pretty clear. If the venture succeeds, their stake in the company will make a lot of money without a lot of initial risk. Even if the venture fails, they’ve managed to offload a division they don’t particularly care about or value.
For Yahoo, knowing that ATDN is Tier 1 provider, the value proposition becomes clear as well. Yahoo is being presented the opportunity to slash their transit costs and be able to beam content, like the new Flickr videos, around the globe at low cost. What if Yahoo took advantage of the network and put content caches at strategic locations around the globe? Search results, video and picture content can land at in browsers without crossing the planet. It becomes possible to build many well-placed small datacenters instead of one or two gigantic datacenters. In fact, these small datacenters can be small enough to live in colocation space, rather than a bought-and-owned datacenter. Yahoo could pay a fairly small amount for colocation rather than build their own buildings in Germany or Dallas or wherever. There are a lot of excellent possibilities for intelligent network managers and architects. Users also will enjoy the benefits with a more responsive, more redundant experience for the Yahoo/AOL properties and products.
The Sad Reality
I’m excited about this idea. I think it presents both Yahoo and AOL with a great opportunity to be true players in the Internet market again. I am also sad because this merger will probably never happen.
Yahoo is going to undergo a major change soon. One of the proposed mergers or takeovers will happen. Regardless of the players, there are two classes of offers before the Yahoo board of directors. First, Yahoo can take a lot of money, a little risk, and immediately provide value to their stockholders in the form of cold hard cash. Second, Yahoo can take a little money, a lot of risk, and provide value to their stockholder eventually in the form of increased company valuation and stock prices. It will be difficult, as the Wall Street Journal article points out, to convince shareholders that any wait is worth it. They seem increasingly fed up with the drama and want some value soon.