Telcos and the State of Content Delivery
Ultimately, both arrive with the same proposition: we will get your stuff where you want it. However, if the customer wants to ship a lot of stuff to a place near a good rail link, then it is likely that shipping will be cheaper and less affected by weather than it may be with the air freight model.
Say the customer wants to ship to a location where there is little or no local rail linkage. Then, despite the effect of weather introducing more complexity, this is mitigated on the basis that waiting for the rail operator to build out a railway network will mean that the cargo could not arrive in a timely manner.
Now, to continue this analogy a little further (I’m sure you are not sick of it yet), it is interesting to note that in the well-organised world of logistics, an air freight operator may also use rail freight on occasion. Equally, a rail freight operator may well use air freight for some legs of the delivery. So everything becomes very convoluted. However, should the air freight business begin to cannibalise the rail operator’s own business (or vice versa), it will have to manage that very carefully, or else it will cause damage to its own existing clients and cost bases. While an air freight operator may elect to land at a different airport and adjust its local road courier relationship quite easily, a rail operator has to consider that its railways need to have a good volume of traffic (and thus revenue) to be maintained in order to be available in the first place. If this volume of traffic falls away too quickly, the rails may suffer, and there may simply be no ability to offer a service.
To complete this analogy, and to switch back into streaming media lingo, this means that while TDNs and CDNs will inevitably interoperate, using each other to ensure that content gets delivered, the CDN is ultimately very nimble since it can move its physical interconnectivity as it wishes. It can benefit from competitive forces and can shop around to ensure it can offer a good price. With volume purchasing power provided by the demands of its customers, this purchasing power can become very significant. Many of the largest CDNs are included in Arbor Networks’ 2009 research, the “Internet Observatory Report,” which noted that only 30 companies produce more than 30% of internet traffic (www.merit.edu/news/newsarchive/ article.php?article=20091013_arbor). Given that Akamai and Limelight carry a significant amount of the other 28 players’ content on occasion too, the CDNs clearly have significant buying power.
In contrast, the TDNs don’t really have to worry about buying prices. Level 3 Communications, the one truly significant TDN player, really does use its essentially zero internal-cost prowess to simply win deals on price. However, like all TDNs, it is limited to where the extremities of its network can get it to. From these extremities, Level 3 has to then get to the end user’s ISP. This is hardly a challenge for Level 3, since it is one of the eight Tier 1 networks—the cerebral cortex that holds the internet together. It is almost utterly pervasive in the core networks. So if you have huge amounts of big content and want to get it out globally, Level 3 would be worth exploring. However, if you don’t have that kind of traffic profile, Level 3 may require too much configuration and setup to make it worthwhile to deliver a relatively small amount of content.
The Regional Advantage
With the number of truly global Tier 1 networks being limited, though, most TDNs will be regional telcos, with a physical network covering an area generally the size of a country or, in some cases, the size of a continent.
So if the customer wants to reach out within only that area, a TDN may be ideal. At this point, we need to look a little deeper into the services telcos offer and why establishing a “telecontent delivery network” is a good idea for telcos.
Although internally a telco will have a great deal of control over its network utilisation costs—a factor that gives telcos an advantage over their customers, CDNs, and ISPs included—these costs are not zero. If the costs were zero (returning to our freight analogy), then there would be no reinvestment into the maintenance of the network itself (be it fibre or rail). And over time, faults and decay would gradually reduce the effectiveness of the service, and this would lead to customer churn.
Actually, these costs are often far from zero, even if they are significantly lower than for their customers. At the end of the day, the lower the utilisation of the network, the less the maintenance cost is and the more capacity is available to sell to other clients over and above the TDN service’s clients.
With our own Content Delivery Summit coming up in October, we took a look at one of the other CDN events, the CDN World Forum in London.
Europe's leading "home-grown" CDN spins off technologies into new products, including the "CDN in a Box"